author's personal copy - paginas.fe.up.ptlepabe/publicacoes_restritas/papers/lmm/... ·...

This article was published in an Elsevier journal. The attached copyis furnished to the author for non-commercial research and

education use, including for instruction at the author’s institution,sharing with colleagues and providing to institution administration.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Chemical Engineering Science 62 (2007) 6534–6547www.elsevier.com/locate/ces

Characterization of potassium glycinate for carbon dioxideabsorption purposes

A.F. Portugala, P.W.J. Derksb, G.F. Versteegb, F.D. Magalhãesa, A. Mendesa,∗aLEPAE–Departamento de Engenharia Química, Faculdade de Engenharia, Universidade do Porto, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal

bProcede Group BV, P.O. Box 328, 7500 AH Enschede, The Netherlands

Received 8 May 2007; received in revised form 26 July 2007; accepted 28 July 2007Available online 6 August 2007

Abstract

Aqueous solutions of potassium glycinate were characterized for carbon dioxide absorption purposes. Density and viscosity of these solutions,with concentrations ranging from 0.1 to 3 M, were determined at temperatures from 293 to 313 K. Diffusivity of CO2 in solution was estimatedapplying the modified Stokes–Einstein relation. Solubilities of N2O at the same temperatures and concentrations were measured and the ionspecific parameter based on Schumpe’s model was determined for the glycinate anion; the solubilities of CO2 in these solutions were thencomputed.

The reaction kinetics of CO2 in the aqueous solution of potassium glycinate was determined at 293, 298 and 303 K using a stirred cellreactor. The results were interpreted using the DeCoursey equation for the calculation of the enhancement factor. The rate of absorption as afunction of the temperature and solution concentration for the conditions studied was found to be given by the following expression:

−rCO2 = 2.42 × 1016 exp

(−8544

T

)exp(0.44CS)CSCCO2 .

� 2007 Elsevier Ltd. All rights reserved.

Keywords: Carbon dioxide; Absorption; Glycine; Mass transfer; Kinetics; N2O solubility

1. Introduction

The carbon dioxide removal from closed anesthetic loopsis currently achieved using soda lime canisters (a mixture ofcalcium, potassium and sodium hydroxides) which is an un-safe technique (Mendes, 2000). The use of dehydrated sodalime is associated to explosions due to the hydrogen formationand excessive heating during the reaction with carbon dioxide.Soda lime can also originate toxic compounds resulting fromthe reaction with some halogenated anesthetics (Whalen et al.,2005). Because of that and because exhausted soda lime can-isters are a hospital solid waste, this outdated system needs tobe replaced by a safer technology. A possible candidate is theuse of absorption membrane contactors. This strategy presentsvarious advantages. The use of a dense highly permeable

∗ Corresponding author. Tel.: +351 1 22 508 1695; fax: +351 22 508 1449.E-mail address: [email protected] (A. Mendes).

0009-2509/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.doi:10.1016/j.ces.2007.07.068

membrane isolates the absorption system from the anestheticloop and the absorption solution can be regenerated after con-tacting with carbon dioxide. However, the absorbent shouldhave a suitable carbon dioxide absorption kinetics and capacity,negligible vapor pressure, high chemical and thermal stabilityand should be harmless to the patient.

Mostly as a consequence of the Kyoto protocol, the new strin-gent environmental regulations towards the emission of acidicgases raised concern about carbon dioxide capture and storage.In the last decades, hollow fiber membrane contactors havebeen studied using absorbent aqueous solutions such as alka-nolamines or blends of alkanolamines for the selective removalof acid gases like H2S and CO2 from a variety of industrialand natural gas streams (Al-Juaied and Rochelle, 2006; Kumaret al., 2003a). However, for applications in highly oxygenatedenvironments, such as flue gas treatment, life support systemsor anesthetic gas circuits, alkanolamines might not be of interestsince they undergo oxidative degradation (Goff and Rochelle,


A.F. Portugal et al. / Chemical Engineering Science 62 (2007) 6534–6547 6535

2006; Kumar et al., 2003c; Supap et al., 2006). Amino acids arenow being studied as a possible alternative for alkanolamines(Feron and Jansen, 2002). Although being more expensive, afew advantages make amino acids attractive like being gener-ally more stable to oxidative degradation and presenting lowervolatilities while showing similar absorption kinetics and ca-pacities in comparison to alkanolamine solutions (Kumar et al.,2003a). Moreover, amino acids aqueous solutions have highersurface tensions and the viscosities are very similar to wa-ter’s. If membrane contactors are to be used, it is importantto consider that the membrane materials should be compatiblewith the absorption liquid. A liquid with higher surface tensionand lower corrosiveness will make possible the efficient use ofcheaper and commercially available membranes, economicallyimproving the process (Kumar et al., 2003a).

Despite of rising interest, few studies have been performedso far on amino acids as carbon dioxide absorbents. TNO En-vironment Energy and Process Innovation has been developinga process for carbon dioxide removal from flue gas processbased on the use of amino acids and salts (Feron and Jansen,2002). Kumar and co-workers studied in detail the absorptionof carbon dioxide in potassium salts of taurine and briefly ana-lyzed glycine (Kumar et al., 2001, 2002, 2003a–c). Holst et al.(2006) compared the apparent absorption rate constants of CO2with different amino acid salt solutions and concluded that theywere comparable with alkanolamines. Recently Lee et al. stud-ied the physical properties and the absorption kinetics of sodiumglycinate as an absorbent of carbon dioxide (Lee et al., 2005,2006, 2007; Song et al., 2006). However, the data available inliterature are still too limited to permit a suitable design andoptimization of processes using amino acid absorbents.

After a pre-screening of a set of different amino acid salts,potassium glycinate presented several interesting properties,such as very good thermal stability and fast apparent reactionrate towards carbon dioxide. Besides, it is commercially avail-able and relatively cheap. For these reasons, it was selected forcharacterization as a carbon dioxide absorbent in the presentwork. This includes the determination of the densities and vis-cosities of aqueous solutions with concentrations between 0.1and 3 M and temperatures from 293 to 313 K. The solubility ofN2O in potassium glycinate solutions was also measured andthe absorption kinetics of carbon dioxide in potassium glyci-nate solutions obtained.

2. Zwitterion reaction mechanism

The zwitterion mechanism, originally proposed by Caplow(1968) is generally applied to model the carbon dioxide ab-sorption in amino acid aqueous solutions. According to thezwitterion mechanism, CO2 reacts with the amino acid salt(potassium glycinate in the present case) forming a zwitterionthat is subsequently deprotonated by a base present in solution.

Formation of the potassium glycinate zwitterion:

(1)

Removal of a proton by a base:

−OOC+H2N– CH2– COO−K+ + Bi

kBi−→−OOCHN– CH2– COO−K+ + BiH

+, (2)

where Bi are the bases present in solution able to deprotonatethe zwitterion. In amino acid salt solutions, these bases are H2O,OH− and the amino acid salt H2NCH2COO−K+ (Blauwhoffet al., 1984).

Assuming quasi-steady-state condition for the zwitterionconcentration and since the second proton transfer step can beconsidered irreversible, the overall reaction rate, −rCO2 , canthen be obtained:

−rCO2 = k2

1 + k−1∑ikBi

CBi

CSCCO2 , (3)

where CS is the concentration of the amino acid salt and CCO2

is the concentration of carbon dioxide in the liquid. Limitingconditions lead to simplified reaction rate expressions that arewell described in literature (Derks et al., 2006; Kumar et al.,2003a).

During the absorption, carbon dioxide reacts also with thehydroxide ions present in solution:

CO2 + OH− kOH−→ HCO−3 . (4)

Taking reaction (4) into account, the overall reaction rate (3)becomes

−rCO2 =

⎛⎜⎜⎝ k2

1 + k−1∑ikBi

CBi

CS + kOH−COH−

⎞⎟⎟⎠ CCO2 . (5)

However, as potassium glycinate is a weak base, the contribu-tion of reaction (4) to the overall reaction kinetics can be con-sidered negligible as well as the contribution of OH− to thedeprotonation of the zwitterion (Kumar et al., 2003c).

The overall rate of reaction of CO2 with potassium glycinatetherefore becomes

−rCO2 = k2CSCCO2

1 + 1(kH2O/k−1

)CH2O + (kAmA/k−1) CS

. (6)

Primary amines such as monoethanolamine (MEA) usually re-act with CO2 following a second order reaction kinetics, whichmeans that the deprotonation of the zwitterion is relatively fastwhen compared to the reversion rate of CO2 and the amine(k−1/

∑ikBi

CBi>1). Eq. (6) is then reduced to

−rCO2 = k2CSCCO2 . (7)

A thermodynamically sound model for the calculation of thekinetic constant should be expressed in terms of activities ratherthan concentrations (Haubrock et al., 2005). However, such amodel would require the knowledge of a number of parameters


6536 A.F. Portugal et al. / Chemical Engineering Science 62 (2007) 6534 – 6547

including equilibrium data that are not available. To account forthe solution non-idealities it is common to use a semi-empiricalmodel which relates the kinetic constant to the solution ionicstrength (Cullinane and Rochelle, 2006):

keff = k exp(bI ), (8)

where keff is the effective kinetic constant, corrected for the so-lution ionic strength, b is a constant and I is the ionic strengthgiven by I = 1

2

∑Ciz

2i , where Ci and zi are, respectively, the

molar concentration and the charge of ion i in solution. Thismodel is not thermodynamically sound and cannot be extrapo-lated for different ions present in solution since all the solutionnon-idealities are lumped in the effective kinetic constant, keff ;however, it is generally sufficient to represent the experimentaldata of a single absorption system (Haubrock et al., 2005).

3. Mass transfer

The absorption of a pure gas (carbon dioxide in the presentwork) into a lean reactive liquid (potassium glycinate solution)is described by the following equation (Danckwerts, 1970):

NCO2 = E · kL

PCO2

HCO2

A, (9)

where NCO2 is the molar flow of CO2 entering the liquid, kL

is the physical mass transfer coefficient, PCO2 is the CO2 par-tial pressure in the gas phase, HCO2 is the Henry constantof CO2 in solution, A is the interfacial area between the gasand the liquid phases and E is the enhancement factor. Theenhancement factor represents the ratio between the rate ofabsorption in the presence of the chemical reaction and thephysical rate of absorption. When the reaction rate is suffi-ciently high, the reaction occurs entirely in the liquid film andnot in the liquid bulk and the absorption rate can be dividedinto three main regimes depending on the dimensionless Hattanumber:

Ha =√

kovDCO2

kL

, (10)

where kov is the overall reaction kinetic constant (kov =−rCO2/CCO2 ) and DCO2 is the diffusion coefficient of CO2 insolution.

Fast pseudo-first order (PFO) reaction regime can be assumedif the following criterion is fulfilled (Danckwerts, 1970):

3 < Ha>E∞. (11)

In this case, the processes of diffusion and reaction occur inparallel in the liquid film. The enhancement factor can be con-sidered equal to the Hatta number and the gas absorption ratebecomes, therefore, independent of the physical mass transfercoefficient. The infinite enhancement factor, E∞, correspondsto a situation of instantaneous reaction and can be estimated,according to the penetration theory, by the following equation

(Danckwerts, 1970; Higbie, 1935):

E∞ =√

DCO2

DS

+ CS

�CO2

PCO2

HCO2

√DS

DCO2

, (12)

where DS is the amino acid salt diffusion coefficient and �CO2

is the stoichiometric coefficient. The instantaneous reactionregime can be considered when E∞>Ha.

Between the limiting situations of fast PFO and instantaneousreaction regime, there is the intermediate regime. According toDeCoursey, the enhancement factor in the intermediate regimecan be approximated as a function of the Hatta number and theinfinite enhancement factor (DeCoursey, 1974; Van Swaaij andVersteeg, 1992):

E = − Ha2

2(E∞ − 1)+

√Ha4

4(E∞ − 1)2+ E∞ · Ha2

E∞ − 1+ 1. (13)

Since carbon dioxide reacts with the amino acid salt solution,the physical properties such as physical solubility and diffu-sivity cannot be directly measured and need to be estimatedindirectly by analogy with a non-reactive gas with similar prop-erties. Typically, N2O is the gas used for this purpose becauseit has a very similar molecular configuration, volume and elec-tronic structure and it does not react with the absorbent solution(Laddha et al., 1981).

Since the amino acid salt solutions are ionic, the so-calledN2O analogy cannot be directly applied to estimate the solu-bility of CO2 in these solutions. Schumpe proposed a model todescribe the solubility of gases in ionic solutions, which takesinto account the salting out effect observed in electrolyte solu-tions (Schumpe, 1993; Weisenberger and Schumpe, 1996). Thismodel enables a reliable estimation of the solubility of CO2 inelectrolyte solutions.

The diffusion coefficient is usually difficult to accuratelydetermine and requires time consuming experiments. Many au-thors studied the dependence of the diffusion coefficient on thetemperature and on the concentration of the absorbent solutionand concluded that it can be related to the solution viscosity, �,through a modified Stokes–Einstein equation (Brilman et al.,2001; Joosten and Danckwerts, 1972; Kumar et al., 2001;Versteeg and Van Swaaij, 1988).

D�� = constant, (14)

where � is a constant that depends on the pair gas/solution.

4. Experimental

4.1. Chemicals

Since the amino acid exists in solution with the amine groupprotonated, it is necessary to make it react with a hydroxidesalt to deprotonate the amine group enabling it to react withcarbon dioxide. The potassium glycinate aqueous solutionswere prepared by adding to the amino acid an equimolaramount of potassium hydroxide in a volumetric flask with dis-tilled and deionized water. The concentrations of the solutions



were verified by a standard potentiometric titration with 1 NHCl solution.

4.2. Density and viscosity

The density of the solutions was measured using a commer-cial density meter (DMA 58, anton Paar GmbH).

Viscosities of potassium glycinate solutions were determinedexperimentally using a standard Cannon–Fenske viscosimeter.

4.3. N2O solubility

The procedure adopted to measure the solubility of N2O inthe amino acid salt solutions is described in detail by Derkset al. (2005) and will only be briefly summarized here. The set-up used is composed of two vessels with calibrated volumes;one for storing the nitrous oxide and the other for the absorbentsolution which is magnetically stirred. A known volume ofsolution is transferred to the absorbent vessel and degassedby applying vacuum. The vapor equilibrium is allowed to bereached at a given temperature; the vapor pressure, Pvapor, isthen recorded. The gas vessel is filled with N2O. A certainamount of N2O is allowed to enter the absorbent vessel and theinitial pressure, Pinit , is recorded. The stirrer is then switchedon and the solution equilibrium is allowed to be established.The final pressure, Peq, is recorded as well as the temperature,Tinit . The temperature is then set to a different value, T , withthe help of the thermostatic bath and a new equilibrium state isestablished. The amount of absorbed gas is calculated applyingthe ideal gas law. The Henry coefficient for N2O, HN2O, is thencomputed from the following equation:

HN2O(T ) = [Peq(T ) − Pvapor(T )][Pinit−Pvapor(Tinit)]

Tinit− [Peq(T )−Pvapor(T )]

T

(VL

RV g

),

(15)

where Vg and VL are, respectively, the volume of gas and liquidin the absorbent vessel and R is the universal gas constant.

The solution vapor pressure at each temperature is estimatedby the following relation:

Pvapor(T ) = xH2OPpureH2O(T ), (16)

where xH2O is the molar fraction of water in solution. The vaporpressure as a function of the absolute temperature, P

pureH2O(T ),

is obtained from the Antoine equation (Poling et al., 2001).The experimental solubility of N2O as a function of the tem-

perature is hence obtained using the same sample. The volumeof liquid as a function of temperature and the amino acid molarfraction are obtained using the density and the mass of solution.

4.4. Kinetic measurements

The experiments were performed in a stirred cell reactorwith a smooth gas–liquid interface, with an interfacial area of6.490×10−3 m2, operating batchwise with respect to the liquidphase and semi-continuously with respect to the gas phase. The

Fig. 1. Simplified scheme of the experimental set-up.

set-up and procedure are described in detail by Derks et al.(2006) and will be only briefly summarized here. The stirredcell reactor is connected to a calibrated gas vessel filled withpure carbon dioxide by means of a pressure controller (Brooks,model 5866, 0–500 mbar, 0.5 FS precision). A fresh potassiumglycinate solution, previously degassed by applying vacuum,is transferred into the stirred reactor. Subsequently, after thevapor–liquid equilibrium is attained at a given temperature, thegas phase pressure inside the stirred reactor is recorded, Pvapor.One begins the experiment by letting the carbon dioxide toflow from the gas vessel into the stirred cell reactor. Duringthe experiment the pressure inside the stirred cell reactor iskept constant, Psc, using the pressure controller, and the flowof absorbed carbon dioxide is computed following the pressuredecrease inside the gas vessel. A sketch of the unit is presentedin Fig. 1.

The flow of absorbed carbon dioxide in the stirred reactor,NCO2,sc, is given by Eq. (9) and PCO2 is the carbon dioxidepartial pressure in the stirred cell (PCO2 = Psc − Pvapor). In thegas vessel, by simply applying the ideal gas law, the flow ofCO2 is given by

NCO2,gv = Vgv

RT

dPgv

dt, (17)

where dPgv/dt is the pressure decrease rate in the gas vesseland Vgv is the volume of the gas vessel:

NCO2, gv = NCO2,sc. (18)

If fast PFO regime is fulfilled, it is possible to determine ex-perimentally the overall reaction kinetic constant, kov, knowingHCO2 and DCO2 . When PFO is considered, the carbon dioxideflow into the reactor tank is given by

Vgv

RT

dPgv

dt= √

kovDCO2

PCO2,sc

HCO2

A. (19)

However, to decide the operating conditions that lead to fastPFO reaction regime, it is necessary to calculate Ha, whichimplies the previous knowledge of kov. For experiments per-formed at a given temperature, absorbent concentration andstirring speed, the Hatta number is constant. Changing the par-tial pressure of carbon dioxide inside the reactor, one changesthe infinite enhancement factor and, consequently, the ratio be-tween Ha and E∞, which means that the absorption regimechanges. By lowering the carbon dioxide partial pressure at



Fig. 2. NCO2 as a function of PCO2 at 298 K for a potassium glycinateconcentration of 0.587 M.

constant Ha, the ratio between the flow and the partial pres-sure of carbon dioxide becomes eventually constant, that isthe value of NCO2/PCO2 becomes independent of PCO2 . Un-der these conditions it is very probable that the PFO reactionregime is attained. Plotting NCO2 as a function of PCO2 at thefast PFO regime, the slope of this curve is related with kov at agiven temperature and amino acid concentration. Fig. 2 showsexperimental values of NCO2 as a function of the partial pres-sure of carbon dioxide. The slope of the fitted line is relatedwith kov through Eq. (19).

For higher carbon dioxide partial pressures, still at constantHa, the flow of carbon dioxide into the liquid (absorption) de-pends not only on the overall kinetic constant but also on thediffusivity ratio of carbon dioxide and absorbent, DCO2/DS .For sufficiently high partial pressures, the instantaneous reac-tion regime is reached when the enhancement factor becomesindependent on the overall kinetic constant.

5. Results and discussion

Densities of potassium glycinate aqueous solutions with con-centrations from 0.1 to 3 M and temperatures from 273 to 313 Kwere determined and are presented in Table 1.

The experimental solubility of N2O in potassium glycinatesolutions is given in Table 2 and Fig. 3.

The same experimental method was also used to obtain thesolubility of N2O in water and results compared with the onesby Versteeg and Van Swaaij (1988). It was verified that theyagree within 2% relative error.

The solubility data of N2O in potassium glycinate was fittedusing the Sechenov relation:

log

(HN2O

HN2O,w

)= K · Cs , (20)

where HN2O and HN2O,w are, respectively, the Henry constantsof N2O in the amino acid salt solution and in water. For eachconcentration and temperature, averaged values of HN2O were

Table 1Densities of potassium glycinate solutions

T (K) � (kg m−3)

293 298 303 313CS (M)

0 998.29 997.13 995.71 992.250.102 1004.37 1003.06 1001.59 997.280.296 1015.97 1014.56 1013.00 1001.770.594 1033.28 1031.69 1030.02 1025.151.003 1056.57 1054.81 1052.98 1047.941.992 1112.29 1110.13 1108.01 1102.342.984 1163.85 1161.44 1159.07 1150.37

used. For each temperature, the computed Sechenov constants,K , failing the t-test were rejected based on a 5% confidencelimit.

For a single salt, Weisenberger and Schumpe (1996)proposed the following model to predict the Sechenov’sconstant, K:

K =∑

(hi + hG)ni , (21)

where hi and hG are the ion and gas specific parameters and ni

is the valency number of the ion. In the present work, nGly− =1and nK+ = 1 and the Sechenov constant becomes

KN2O = (hGly− + hN2O) + (hK+ + hN2O). (22)

Values of parameters hK+ and hN2O for the cation and thegas, respectively, are reported in literature (Weisenberger andSchumpe, 1996). These values, together with the experimentalSechenov’s constant, KN2O, were used to calculate the anionspecific parameter, hGly− . The values of Sechenov’s constantas well as the specific parameters of gas and cation and thecalculated value of the anion parameter are given in Table 3.

Parameters hK+ and hGly− are expected to be essentially con-stant with the temperature (Weisenberger and Schumpe, 1996).The values obtained for hGly− show, however, slight tempera-ture dependence. The average hGly− over the temperature rangeis 0.0397. Taking the values of solubility in water reported byVersteeg and Van Swaaij (1988), the anion parameter hGly−obtained by Kumar et al. (2001) is 0.0413 at 295 K which isin agreement with the present work. The value of the anionspecific parameter hGly− along with the CO2 specific param-eter, hCO2 , determined by Weisenberger and Schumpe (1996)enables to predict Sechenov’s constant of CO2 in potassiumglycinate solutions, KCO2 , presented in Table 4.

The computed physical solubility of carbon dioxide in potas-sium glycinate solutions is given in Table 5.

Viscosities of potassium glycinate solutions were determinedexperimentally and the Stokes–Einstein relation was used toestimate the diffusion coefficient of N2O. Versteeg and VanSwaaij (1988) obtained parameter � of the Stokes–Einstein



Table 2Experimental Henry constants of N2O in potassium glycinate solutions

CS (M) T (K) HN2O (Pa m3 mol−1) CS (M) T (K) HN2O (Pa m3 mol−1)

0.102 293.2 3640 1.003 293.9 47180.102 297.4 4086 1.003 298.3 53680.102 298.5 4196 1.003 298.7 53010.102 303.1 4719 1.003 303.0 59910.296 293.1 3908 1.003 303.4 58540.296 293.3 3861 1.003 312.4 72520.296 293.4 3876 1.992 293.1 60170.296 293.5 3830 1.992 293.2 61400.296 298.0 4369 1.992 293.6 61120.296 298.2 4455 1.992 293.8 60240.296 303.3 4995 1.992 298.1 67110.296 311.9 6094 1.992 298.4 69740.594 293.4 4241 1.992 302.9 77530.594 293.7 4235 1.992 303.0 75190.594 298.3 4856 1.992 312.0 93150.594 302.2 5375 2.984 293.2 76940.594 312.0 6782 2.984 293.3 76211.003 293.0 4683 2.984 298.5 85451.003 293.1 4539 2.984 303.1 93051.003 293.3 4626

Fig. 3. Experimental Henry constants of N2O in water and in potassium gly-cinate solutions as a function of temperature. Comparison with the solubilityin water determined by Versteeg and Van Swaaij (1988).

relation from experimental values of the diffusion coefficientof N2O in several alkanolamines aqueous solutions. These au-thors proposed � = 0.8, while Brilman et al. (2001) concludedthat the ionic strength of the salt solutions does not influencethe diffusion coefficient. For these reasons, in the present workit is assumed � = 0.8 for the estimation of the diffusion coef-ficient of N2O in potassium glycinate solutions. Kumar et al.(2001) studied the diffusivity of N2O in amino acid salts aque-ous solutions and found � = 0.74 for potassium taurate. Thedifferences in the calculated diffusivities using one or the othervalue for � are lower than 5%, which is within the general ex-perimental error for the determination of diffusion coefficients.

The diffusion coefficient of CO2 in solution is determinedusing the so-called N2O : CO2 analogy (Gubbins et al., 1966):

DN2O

DN2O,w

= DCO2

DCO2,w

. (23)

The values of the diffusivity of nitrous oxide and carbon dioxidein water were obtained from the literature (Versteeg and VanSwaaij, 1988). The results of the experimentally determinedviscosities are given in Table 6 along with calculated diffusioncoefficients of N2O and CO2.

5.1. Overall kinetic constants

The overall kinetic constants of the carbon dioxide absorp-tion in potassium glycinate aqueous solutions were calculatedusing the described methodology for a potassium glycinateconcentration from 0.1 to 3 M and at 293, 298 and 303 K.Table 7 shows the overall kinetic constants of carbon dioxideabsorption obtained at the potassium glycinate concentrationsand temperatures employed. Only the experimental values ofNCO2 as a function of PCO2 considered to be in the fast pseudo-first order reaction regime were used for that calculation. Thecomplete set of kinetic results is shown in Appendix.

One must now verify if inequality (11), corresponding to thePFO reaction criterion, is fulfilled. The Hatta numbers and theinfinite enhancement factors were then calculated for each ex-perimental condition. However, to calculate the Hatta numberone needs to determine the physical mass transfer coefficient,kL—see Eq. (10), and to calculate the infinite enhancementfactors one needs to determine the diffusion coefficient of potas-sium glycinate in solution, DS—see Eq. (12).



Table 3Sechenov’s constants for solubility of N2O in aqueous potassium glycinate solutions

T (K) KN2O (dm3 mol−1) hN2O (dm3 mol−1) hK+ (dm3 mol−1) hGly− (dm3 mol−1)

293 0.115 −0.0061 0.0922 0.0352298 0.116 −0.0085 0.0922 0.0408303 0.112 −0.0109 0.0922 0.0417313 0.102 −0.0157 0.0922 0.0409

Table 4Sechenov’s constants for solubility of CO2 in aqueous potassium glycinatesolutions

T (K) hCO2 (dm3 mol−1) KCO2 (dm3 mol−1)

293 −0.0155 0.101298 −0.0172 0.097303 −0.0189 0.094313 −0.0223 0.087

Table 5Henry constants of CO2 in potassium glycinate solutions computed based onSechenov’s model

T (K) HCO2 (Pa m3 mol−1)

293 298 303 313CS (M)

0.10 2710 3044 3405 42170.30 2839 3183 3556 43900.59 3036 3397 3787 46531.0 3340 3725 4138 50532.0 4212 4662 5139 61793.0 5313 5835 6382 7554

The physical mass transfer coefficient, kL, was calculated us-ing the empirical expression referred by Versteeg et al. (1987):

Sh = c2Rec3Scc4 , (24)

where Sh, Re and Sc are respectively the Sherwood, Reynoldsand Schmidt dimensionless numbers defined as

Sh = kL · dS

DCO2

, (25)

Re = (dS)2N�

�, (26)

Sc = �

� · DCO2

, (27)

where dS and N are, respectively, the characteristic dimensionand the speed of the stirrer which are, in the present case,dS = 9.09 × 10−2 m and N = 108 min−1. The constants c2, c3and c4 were determined experimentally performing absorptionexperiments of CO2 in water at different temperatures and theyshow to be within the usual values for this kind of systems(Versteeg et al., 1987).

Sh = 0.1018 · Re0.7279Sc0.4076. (28)

It was verified that all computed Ha were much higher than 3and therefore the first part of the inequality (11) is confirmed.

The diffusion coefficient of potassium glycinate in solu-tion, DS , was computed assuming that it follows the modifiedStokes–Einstein relation (14) with � = 0.6 (Versteeg and VanSwaaij, 1988). To estimate the diffusion coefficient of the saltat infinite dilution, D0

S , the Nernst equation for the diffusion inelectrolyte solutions was applied (Poling et al., 2001):

D0S = RT [(1/z+) + (1/z−)]

F 2[(1/�0+) + (1/�0−)] , (29)

where F is the Faraday constant, z+ and z− are the valenciesof the cation and anion, respectively, and �0+ and �0− are theionic conductances of the cation and anion, respectively, at in-finite dilution. Values of �0+ at each temperature was calculatedbased on the work of Fell and Hutchiso (1971). �0− at 298 Kwas obtained from Miyamoto and Schmidt (1933) and it wasassumed that it depends on the temperature in the same way as�0+. The computed diffusion coefficient of potassium glycinatein solutions, DS , are shown in Table 8.

The Hatta number, Ha, along with the minimum value ofE∞ (corresponding to the higher pressure used for computingkov assuming the PFO) are given in Table 9 for the absorbentconcentrations and temperatures studied.

Fast PFO regime can only be ensured for ratios between E∞and Ha close to 10. For this reason, the DeCoursey relation wasapplied and new values for kov were calculated by minimizingthe sum of the square residues between the experimental carbondioxide absorption flow and the flow calculated by applying theDeCoursey equation. These values are given in Table 10.

The deviation of the enhancement factor determined ex-perimentally and the one calculated based on the DeCourseyequation is presented in Fig. 4.

Since all the experiments were performed at very low load-ings, the only ions contributing to the ionic strength of thesolutions are potassium cation and glycinate anion, both mono-valent species (z2 = 1) and thus I = CS . Combining Eq. (7)with (8), the absorption rate of CO2 becomes

−rCO2 = k2CS exp(bCS)CCO2 . (30)

In addition, assuming that the kinetic constant obeys theArrhenius equation:

k2 = k2,0 exp

(A

T

). (31)



Table 6Viscosity and diffusivity of N2O and CO2 in potassium glycinate solutions

CS (M) T (K) � × 103

(kg m−1 s−1)

DN2O × 109

(m2 s−1)

DCO2 × 109

(m2 s−1)

0.102 293 1.030 1.52 1.67298 0.914 1.75 1.89303 0.819 1.99 2.12313 0.666 2.57 2.66

0.296 293 1.075 1.47 1.61298 0.962 1.68 1.81303 0.851 1.93 2.06313 0.693 2.49 2.58

0.594 293 1.148 1.40 1.53298 1.020 1.60 1.73303 0.909 1.8 1.95313 0.746 2.35 2.43

1.003 293 1.263 1.30 1.42298 1.136 1.47 1.58303 1.008 1.69 1.80313 0.826 2.16 2.24

1.992 293 1.620 1.06 1.16298 1.449 1.21 1.30303 1.287 1.39 1.48313 1.070 1.76 1.82

2.984 293 2.109 0.86 0.94298 1.861 0.99 1.07303 1.677 1.12 1.20313 1.363 1.45 1.50

Table 7Experimental values of the overall kinetic constant assuming pseudo-first order behavior

T (K) Slope × 107 (mol mbar−1 s−1) kov (s−1)

293 298 303 293 298 303CS (M)

0.0994 – 2.51 – – 732 –0.299 4.44 4.38 5.19 2340 2540 39300.587 4.30 5.94 7.37 2640 5590 94900.999 7.09 8.83 9.40 9390 16 200 20 0001.984 7.92 9.55 12.7 22 800 36 100 68 0003.005 – 10.7 – – 86 300 –

Table 8Computed values of DS used to calculate E∞

T (K) DS × 1010 (m2 s−1)

293 298 303CS (M)

0.0994 – 5.59 –0.299 4.28 5.42 6.640.587 4.12 5.23 6.380.999 3.89 4.90 6.001.984 3.35 4.24 5.183.005 – 3.65 –

Table 9Ha and minimum values of E∞ used for computing kov assuming PFO

T (K) Ha E∞

293 298 303 293 298 303CS (M)

0.0994 – 38.5 – – 182 –0.299 73.1 72.9 86.2 316 477 7130.587 79.4 110 137 437 500 7760.999 154 194 205 748 937 14601.984 261 314 410 1930 2430 37103.005 – 528 – – 4400 –



Table 10Experimental values of the overall kinetic constants of potassium glycinatecalculated using the DeCoursey equation

T (K) kov (s−1)

293 298 303CS (M)

0.0994 – 881 –0.299 2860 2710 43600.587 3130 6420 11 0000.999 11 500 19 600 22 0001.984 24 200 38 600 69 8003.005 – 93 900 –

Fig. 4. Parity plot of experimental enhancement factor and the DeCourseyapproximation.

Plotting kov as a function of CS it is possible to perform a globalfitting to the experimental results for all temperatures and con-centrations considered, in order to obtain the kinetic parametersk2,0, A and b. The resulting expression for computation of theoverall kinetic constant as a function of temperature and aminoacid salt concentration, obtained by minimizing the sum of therelative residues squared is (with CS expressed in mol dm−3):

kov = 2.42 × 1016 exp

(−8544

T

)exp(0.44CS)CS s−1, (32)

where a coefficient of determination of 0.991 was obtained.The zwitterion mechanism constants, k2, (kH2O/k−1)

and (kAmA/k−1), were also fitted using the same proce-dure but not accounting to the solution ionic strength:k2 = 2.81 × 1010 exp(−5800/T ) m3 mol−1 s−1; (kH2O/k−1) =1.05 × 10−4 exp(−1265/T ) m3 mol−1 and (kAmA/k−1) =4.89 × 103 exp(−5307/T ) m3 mol−1, where a coefficient ofdetermination of 0.956 was obtained. Both fittings are shownin Figs. 5 and 6. Although the first model fits better the exper-imental results (smaller sum of squared residues), generallyboth models are in agreement with the experimental resultsfor concentrations up to 3 M. It is, however, noticeable thatabove this concentration it is no longer possible to neglect thenon-idealities of the solution. The second model has six fitting

Fig. 5. Overall absorption kinetic constant as a function of potassium glycinateconcentration and for different temperatures: experimental values and modellines. Solid lines correspond to the model that takes into account the ionicstrength and dashed lines to the zwitterion model.

Fig. 6. Apparent absorption kinetic constants as a function of potassiumglycinate concentration and at different temperatures: experimental valuesand model lines. Solid lines correspond to the model that takes into accountthe ionic strength and dashed lines to the zwitterion model.

parameters while the first has just 3. The second model has alarge number of fitting parameters and over fitting can easilyoccur. It is very difficult in such circumstances to identify ifthe non-idealities play or not a significant role. The simplerfirst model then becomes more attractive in the present work.

The values of kov obtained are in good agreement with thework of Kumar et al. (2003c) for low concentrations but devi-ate for higher concentrations. This is possibly due to the effectof the ionic strength not being taken into account in Kumar’swork. On the other hand, the results of the present work arequite different from the ones by Lee et al. (2007). Those authorsmention apparent kinetic constants for carbon dioxide absorp-tion in aqueous sodium glycinate solutions about two ordersof magnitude lower. However, a careful analysis of that workshows that the kinetic measurements were performed far from



Fig. 7. BrZnsted plot of Penny and Ritter (1983) at 293, 298 and303 K—comparison with the present work.

the fast PFO reaction regime, since in some cases E∞ was evenlower than Ha.

It is common to relate the kinetic constant of reaction, k2,with the pKa of the amino acid salt by means of a BrZnstedplot. Penny and Ritter (1983) determined the kinetic constantand the pKa of several amino acids (including glycinate anion)and alkanolamines. The values of k2 determined in the presentwork deviate less than 30% (relative error) from the BrZnstedplot based on the work from Penny and Ritter (1983) in theentire temperature range. Fig. 7 presents the BrZnsted plot ofPenny and Ritter (1983) along with the results from this work.

6. Conclusions

Density and viscosity of potassium glycinate aqueous solu-tions ranging from 0.1 to 3 M and at temperatures between 273and 313 K were obtained. The diffusion coefficient of N2O insolution was estimated using a modified Stokes–Einstein rela-tion and the CO2 diffusion coefficient in solution was estimatedusing the so-called N2O : CO2 analogy (Gubbins et al., 1966).

The solubility of N2O in the potassium glycinate solutionswas experimentally determined. The salting out effect of the saltconcentration in the solubility showed to be well described bythe Sechenov equation. The specific parameter of the glycinateanion, based on the Schumpe model (Schumpe, 1993), wascalculated (hGly− =0.0397 mol dm−3) and the solubility of CO2in solution was then estimated.

The rate of reaction of CO2 with potassium glycinate was de-termined in a stirred cell reactor operating in semi-continuousmode. Two approaches were used to obtain the relevant param-eters of the model. Since the conditions for fast PFO reactionregime were apparently not fulfilled, the DeCoursey equationwas employed to calculate the enhancement factor. The resultsindicate that the reaction kinetics significantly depend on theionic strength of the solution. The apparent rate of reaction isin line with the BrZnsted plot based on the work from Pennyand Ritter (1983). The obtained overall kinetic constants point

out that potassium glycinate is a fast absorbent when comparedwith other absorbents, namely with MEA, which shows an over-all kinetic constant at 298 K for a 1 M solution of 5920 s−1

(Glasscock et al., 1991) against the value of 13 400 s−1 ob-tained in the present work for potassium glycinate at the sameconcentration and temperature conditions. In the future, theseresults will be applied in the design and optimization of a mem-brane contactor to be used for carbon dioxide removal fromanesthetic gas streams.

Notation

A gas–liquid interfacial area, m2

C concentration, M or mol m−3

dS Stirrer diameter, mD diffusion coefficient, m2 s−1

E enhancement factor, dimensionlessE∞ infinite enhancement factor, dimensionlessF Faraday constant, 96 500 ◦C mol−1

h ion and gas specific constants in the Shumpeequation, dm3 mol−1

H Henry coefficient, Pa mol−1 m3

Ha Hatta number, dimensionlessI Ionic strength of the solution, mol dm−3

JCO2 carbon dioxide absorption flow, mol m−2 s−1

k−1 zwitterion kinetic constant of the reverse re-action, s−1

k2 zwitterion kinetic constant of the reaction,m3 mol−1s−1

kAmA zwitterion deprotonation rate constant foramino acid, m3 mol−1s−1

kapp apparent rate constant defined as: kapp =kov/CS , m3 mol−1s−1

kBizwitterion mechanism deprotonation rateconstant by base, m3 mol−1s−1

kH2O zwitterion mechanism deprotonation rateconstant for water, m3 mol−1s−1

kL liquid phase physical mass transfer coeffi-cient, m s−1

kOH− reaction rate constant with hydroxide ionm3 mol−1s−1

kov overall kinetic constant, s−1

K Sechenov constant, dm3 mol−1

N stirrer speed, rpsNCO2 carbon dioxide absorption flow, mol s−1

PCO2 carbon dioxide partial pressure, mbar−rCO2 rate of reaction, mol s−1

R universal gas constant, J mol−1K−1

Re Reynolds number, dimensionlessSh Sherwood number, dimensionlessSc Schmidt number, dimensionlessT temperature, KV volume, m3

x molar fraction, mol mol−1

z+, z− valencies of the cation and anion



Greek letters

� constant from the modifiedStokes–Einstein equation

� solution viscosity, kg m−1 s−1

�0+, �0− ionic conductances of the cation and anionat infinite dilution, cm2 �−1

�CO2 stoichiometric coefficient� solution density, kg m−3

Subscripts

CO2 Carbon dioxideeff effective (after correcting for the ionic

strength)eq equilibriumfinal finalg gas phasegv gas vesselGly− glycinate anioninit initialK+ potassium cationL liquid phaseMEA monoethanolamineN2O nitrous oxidesc stirred cellS amino acid saltw water

Acknowledgements

The work of Ana F. Portugal was supported by FCT GrantBD/1662/2004. The research was also supported by funds fromFCT project POCTI/EQU/45182/2002 and project GrowthGRD1-2001-40257. The authors would also like to thankBenno Knaken for his technical support.

Appendix A. Experimental kinetic data

Values of the carbon dioxide flux, JCO2 , as a function ofthe carbon dioxide partial pressure for all temperatures andpotassium glycinate concentrations studied are presented inTables A1–A14.

Table A1Kinetic data of the reaction of CO2 with potassium glycinate at 0.0994 Mand 298 K

PCO2 JCO2 × 104 �max

(mbar) (mol s−1 m−2) (molCO2 mol−1AmA)

3.42PFO/DeCo 1.34 0.0105.05PFO/DeCo 2.05 0.0156.43PFO/DeCo 2.40 0.017

11.44DeCo 3.94 0.02916.50DeCo 4.78 0.03721.65DeCo 5.31 0.04231.54DeCo 6.52 0.05452.11DeCo 7.65 0.068

Table A2Kinetic data of the reaction of CO2 with potassium glycinate at 0.299 M and293 K



4.71PFO/DeCo 3.36 0.0085.66PFO/DeCo 4.27 0.0106.27PFO/DeCo 4.57 0.0116.76PFO/DeCo 4.92 0.0127.28PFO/DeCo 5.58 0.0138.28PFO/DeCo 5.34 0.0138.85PFO/DeCo 5.48 0.0139.84PFO/DeCo 6.38 0.015

12.57DeCo 8.35 0.01914.60DeCo 8.49 0.02019.84 10.0 0.02324.81 12.4 0.01934.73 15.2 0.02340.05 16.2 0.03844.87 16.1 0.03954.97 16.8 0.023




2.71PFO/DeCo 1.78 0.0043.52PFO/DeCo 2.50 0.0024.77PFO/DeCo 3.09 0.0075.89PFO/DeCo 4.14 0.0106.71PFO/DeCo 4.33 0.0107.18PFO/DeCo 4.99 0.0127.74PFO/DeCo 5.16 0.0128.82DeCo 5.55 0.0139.76DeCo 5.80 0.0139.82DeCo 6.27 0.015

10.56DeCo 6.44 0.01511.66DeCo 6.79 0.01616.71 9.02 0.02126.66 13.3 0.03141.96 16.3 0.040




2.10PFO/DeCo 1.64 0.0043.94PFO/DeCo 3.12 0.0075.22PFO/DeCo 4.36 0.0106.00PFO/DeCo 4.66 0.0108.17DeCo 6.62 0.015

11.34DeCo 8.62 0.02116.37 11.6 0.02731.74 17.3 0.04056.62 21.2 0.051






2.16PFO/DeCo 1.54 0.0022.22PFO/DeCo 1.50 0.0024.43PFO/DeCo 3.43 0.0044.46PFO/DeCo 3.29 0.0047.62PFO/DeCo 5.28 0.0069.47PFO/DeCo 6.26 0.0069.51PFO/DeCo 6.38 0.009

10.12PFO/DeCo 6.62 0.00912.06PFO/DeCo 8.10 0.01012.19PFO/DeCo 8.11 0.01015.00PFO/DeCo 9.67 0.00715.00PFO/DeCo 9.73 0.00823.75 13.5 0.01324.79 13.5 0.01124.85 13.5 0.01335.42 19.4 0.01545.50 22.7 0.01862.55 26.7 0.020




5.08PFO/DeCo 5.01 0.0035.22PFO/DeCo 4.49 0.0057.91PFO/DeCo 7.26 0.0087.94PFO/DeCo 7.32 0.008

10.48PFO/DeCo 9.47 0.00710.51PFO/DeCo 9.50 0.00615.54PFO/DeCo 13.9 0.00815.55PFO/DeCo 14.7 0.00820.41 15.9 0.00920.53 16.4 0.01020.57 17.4 0.01120.75 17.8 0.01030.88 22.2 0.01641.00 28.2 0.01641.11 26.2 0.01749.65 31.3 0.01867.11 33.7 0.03467.20 31.7 0.02268.41 37.4 0.02668.44 35.0 0.021





11.61PFO/DeCo 12.8 0.01416.95 16.5 0.02131.56 26.0 0.02556.87 36.7 0.022




5.42PFO/DeCo 6.26 0.0047.06PFO/DeCo 8.82 0.006

10.37PFO/DeCo 11.3 0.00810.43PFO/DeCo 11.0 0.00712.27PFO/DeCo 13.9 0.00913.61PFO/DeCo 13.8 0.01013.86PFO/DeCo 16.3 0.01015.23PFO/DeCo 17.2 0.01115.42PFO/DeCo 15.8 0.01015.46PFO/DeCo 17.4 0.01216.54PFO/DeCo 19.0 0.01320.42 18.8 0.01220.48 18.9 0.01125.59 25.6 0.01735.30 32.3 0.01845.53 37.7 0.01755.70 42.1 0.017





10.62PFO/DeCo 15.2 0.01011.08PFO/DeCo 15.2 0.01012.68PFO/DeCo 17.6 0.01113.78PFO/DeCo 18.1 0.01115.26PFO/DeCo 20.9 0.01415.64PFO/DeCo 20.6 0.01317.59 22.5 0.01517.62 23.2 0.01022.64 25.2 0.01327.52 28.0 0.01732.98 32.7 0.01755.02 43.4 0.01667.09 48.8 0.015




3.42PFO/DeCo 4.99 0.0035.23PFO/DeCo 7.68 0.0056.94PFO/DeCo 10.0 0.0078.59PFO/DeCo 11.9 0.0079.23PFO/DeCo 13.4 0.009

10.54PFO/DeCo 15.5 0.01011.59PFO/DeCo 16.9 0.01117.70 24.7 0.01627.91 34.7 0.01657.65 54.7 0.020






2.77PFO/DeCo 3.35 0.0014.76PFO/DeCo 6.11 0.0026.74PFO/DeCo 8.33 0.0037.71PFO/DeCo 9.33 0.0038.31PFO/DeCo 9.23 0.0048.61PFO/DeCo 11.6 0.0048.74PFO/DeCo 9.67 0.003

10.49PFO/DeCo 13.9 0.00410.78PFO/DeCo 12.4 0.00410.81PFO/DeCo 12.1 0.00412.56PFO/DeCo 15.7 0.00513.07PFO/DeCo 18.2 0.00613.26PFO/DeCo 13.9 0.00515.16PFO/DeCo 19.6 0.00615.51PFO/DeCo 18.6 0.00616.07PFO/DeCo 18.4 0.00616.47PFO/DeCo 21.2 0.00621.75 20.4 0.00721.78 19.8 0.00724.81 24.2 0.00825.00 24.1 0.00826.84 27.4 0.00936.88 35.0 0.00947.01 40.2 0.00757.00 47.9 0.00877.29 55.9 0.010




3.34PFO/DeCo 5.12 0.0015.20PFO/DeCo 7.51 0.0025.28PFO/DeCo 7.01 0.0028.37PFO/DeCo 13.3 0.0049.31PFO/DeCo 15.5 0.005

10.06PFO/DeCo 13.8 0.00510.24PFO/DeCo 14.4 0.00311.09PFO/DeCo 16.6 0.00511.30PFO/DeCo 15.8 0.00511.32PFO/DeCo 15.3 0.00512.52PFO/DeCo 17.1 0.00512.54PFO/DeCo 18.5 0.00513.42PFO/DeCo 19.6 0.00613.53PFO/DeCo 21.9 0.00614.33PFO/DeCo 21.5 0.00715.36PFO/DeCo 22.9 0.00717.34 27.3 0.00717.50 29.4 0.00819.42 29.4 0.00819.47 28.0 0.00921.52 26.8 0.00824.52 30.3 0.00729.47 33.9 0.00749.46 52.5 0.00869.86 66.3 0.007




5.44PFO/DeCo 10.6 0.0037.44PFO/DeCo 14.8 0.0059.51PFO/DeCo 19.1 0.006

11.52PFO/DeCo 21.9 0.00714.34DeCo 23.8 0.008




3.43PFO/DeCo 5.45 0.0016.37PFO/DeCo 11.4 0.003

11.44PFO/DeCo 19.0 0.00416.46PFO/DeCo 26.7 0.00521.61 35.6 0.00726.49 38.4 0.00831.32 44.1 0.00651.45 60.3 0.005

All the experiments began with fresh solution. The maximumloading reached at the end of each experiment, �max, is alsoshown.

The values used to calculate kov considering PFO reactionregime and using the DeCoursey approach are marked, respec-tively, with PFO and DeCo. Since at high carbon dioxide partialpressures, the overall kinetic constant plays a minor role on theenhancement factor and the values of DCO2 and DS are esti-mated, only experiments at low partial pressures were used tocalculate kov even when the DeCoursey approach was used.

References

Al-Juaied, M., Rochelle, G.T., 2006. Absorption of CO2 in aqueousdiglycolamine. Industrial & Engineering Chemistry Research 45 (8),2473–2482.

Blauwhoff, P.M.M., Versteeg, G.F., et al., 1984. A study on thereaction between CO2 and alkanolamines in aqueous-solutions. ChemicalEngineering Science 39 (2), 207–225.

Brilman, D.W.F., Van Swaaij, W.P.M., et al., 2001. Diffusion coefficientand solubility of isobutene and trans-2-butene in aqueous sulfuric acidsolutions. Journal of Chemical and Engineering Data 46 (5), 1130–1135.

Caplow, M., 1968. Kinetics of carbamate formation and breakdown. Journalof the American Chemical Society 90 (2), 6795–6803.

Cullinane, J.T., Rochelle, G.T., 2006. Kinetics of carbon dioxide absorptioninto aqueous potassium carbonate and piperazine. Industrial & EngineeringChemistry Research 45 (8), 2531–2545.

Danckwerts, P., 1970. Gas–Liquid Reactions. McGraw-Hill Book Company,New York.

DeCoursey, W.J., 1974. Absorption with chemical-reaction—development ofa new relation for Danckwerts model. Chemical Engineering Science 29(9), 1867–1872.



Derks, P., Hogendoorn, K., et al., 2005. Solubility of N2O in and density,viscosity, and surface tension of aqueous piperazine solutions. Journal ofChemical and Engineering Data 50 (6), 1947–1950.

Derks, P.W.J., Kleingeld, T., et al., 2006. Kinetics of absorption of carbondioxide in aqueous piperazine solutions. Chemical Engineering Science61 (20), 6837–6854.

Fell, C.J.D., Hutchiso, H.P., 1971. Diffusion coefficients for sodium andpotassium chlorides in water at elevated temperatures. Journal of Chemicaland Engineering Data 16 (4), 427–429.

Feron, P., Jansen, A., 2002. CO2 separation with polyolefin membranecontactors and dedicated absorption liquids: performances and prospects.Separation and Purification Technology 27 (3), 231–242.

Glasscock, D.A., Critchfield, J.E., et al., 1991. CO2 absorption desorptionin mixtures of methyldiethanolamine with monoethanolamine ordiethanolamine. Chemical Engineering Science 46 (11), 2829–2845.

Goff, G.S., Rochelle, G.T., 2006. Oxidation inhibitors for copper and ironcatalyzed degradation of monoethanolamine in CO2 capture processes.Industrial & Engineering Chemistry Research 45 (8), 2513–2521.

Gubbins, K.E., Bhatia, K.K., et al., 1966. Diffusion of Gases in ElectrolyticSolutions. A.I.Ch.E. Journal 12 (3), 548–552.

Haubrock, J., Hogendoorn, J.A., et al., 2005. The applicability of activities inkinetic expressions: a more fundamental approach to represent the kineticsof the system CO2– OH− in terms of activities. International Journal ofChemical Reactor Engineering 3.

Higbie, R., 1935. The rate of absorption of a pure gas into a still liquidduring a short time of exposure. Transactions of the American Instituteof Chemical Engineers 31, 365–389.

Holst, J., Politiek, P., et al., 2006. CO2 capture from flue gas using aminoacid salt solutions. GHGT-8 Proceedings.

Joosten, G.E.H., Danckwerts, P.V., 1972. Solubility and diffusivity of nitrous-oxide in equimolar potassium carbonate—potassium bicarbonate solutionsat 25 ◦C and 1 atm. Journal of Chemical and Engineering Data 17 (4),452–454.

Kumar, P., Hogendoorn, J., et al., 2001. Density, viscosity, solubility, anddiffusivity of N2O in aqueous amino acid salt solutions. Journal ofChemical and Engineering Data 46 (6), 1357–1361.

Kumar, P., Hogendoorn, J., et al., 2002. New absorption liquids for the removalof CO2 from dilute gas streams using membrane contactors. ChemicalEngineering Science 57 (9), 1639–1651.

Kumar, P., Hogendoorn, J., et al., 2003a. Equilibrium solubility of CO2in aqueous potassium taurate solutions: Part 1. Crystallization in carbondioxide loaded aqueous salt solutions of amino acids. Industrial &Engineering Chemistry Research 42 (12), 2832–2840.

Kumar, P., Hogendoorn, J., et al., 2003b. Equilibrium solubility of CO2 inaqueous potassium taurate solutions: Part 2. Experimental VLE data andmodel. Industrial & Engineering Chemistry Research 42 (12), 2841–2852.

Kumar, P., Hogendoorn, J., et al., 2003c. Kinetics of the reaction of CO2with aqueous potassium salt of taurine and glycine. A.I.Ch.E. Journal 49(1), 203–213.

Laddha, S.S., Diaz, J.M., et al., 1981. The N2O analogy—the solubilitiesof CO2 and N2O in aqueous-solutions of organic-compounds. ChemicalEngineering Science 36 (1), 228–229.

Lee, S., Choi, S., et al., 2005. Physical properties of aqueous sodium glycinatesolution as an absorbent for carbon dioxide removal. Journal of Chemicaland Engineering Data 50 (5), 1773–1776.

Lee, S., Song, H.J., et al., 2006. Physical solubility and diffusivity of N2Oand CO2 in aqueous sodium glycinate solutions. Journal of Chemical andEngineering Data 51 (2), 504–509.

Lee, S., Song, H.J., et al., 2007. Kinetics of CO2 absorption in aqueoussodium glycinate solutions. Industrial & Engineering Chemistry Research46 (5), 1578–1583.

Mendes, A.M.M., 2000. Development of an adsorption/membrane basedsystem for carbon dioxide, nitrogen and spur gases removal from a nitrousoxide and xenon anaesthetic closed loop. ACP-Applied CardiopulmonaryPathophysiology 9 (2), 156–163.

Miyamoto, S., Schmidt, C.L.A., 1933. Transference and conductivity studieson solutions of certain proteins and amino acids with special reference tothe formation of complex ions between the alkaline earth elements andcertain proteins. Journal of Biological Chemistry 99 (2), 335.

Penny, D.E., Ritter, T.J., 1983. Kinetic-study of the reaction between carbon-dioxide and primary amines. Journal of the Chemical Society—FaradayTransactions I 79, 2103–2109.

Poling, B.E., Prausnitz, J.M., et al., 2001. The Properties of Gases andLiquids. McGraw-Hill Book Company, New York.

Schumpe, A., 1993. The estimation of gas solubilities in salt-solutions.Chemical Engineering Science 48 (1), 153–158.

Song, H.J., Lee, S., et al., 2006. Solubilities of carbon dioxide in aqueoussolutions of sodium glycinate. Fluid Phase Equilibria 246 (1–2), 1–5.

Supap, T., Idem, R., et al., 2006. Analysis of monoethanolamine and itsoxidative degradation products during CO2 absorption from flue gases:a comparative study of GC-MS, HPLC-RID, and CE-DAD analyticaltechniques and possible optimum combinations. Industrial & EngineeringChemistry Research 45 (8), 2437–2451.

Van Swaaij, W.P.M., Versteeg, G.F., 1992. Mass-transfer accompanied withcomplex reversible chemical-reactions in gas–liquid systems—an overview.Chemical Engineering Science 47 (13–14), 3181–3195.

Versteeg, G.F., Van Swaaij, W.P.M, 1988. Solubility and diffusivity of acidgases (CO2, N2O) in aqueous alkanolamine solutions. Journal of Chemicaland Engineering Data 33 (1), 29–34.

Versteeg, G.F., Blauwhoff, P.M.M., et al., 1987. The effect of diffusivity ongas–liquid mass-transfer in stirred vessels–experiments at atmospheric andelevated pressures. Chemical Engineering Science 42 (5), 1103–1119.

Weisenberger, S., Schumpe, A., 1996. Estimation of gas solubilities in saltsolutions at temperatures from 273 K to 363 K. A.I.Ch.E. Journal 42 (1),298–300.

Whalen, F.X., Bacon, D.R., et al., 2005. Inhaled anesthetics: an historicaloverview. Best Practice Research Clinical Anaesthesiology 19 (3),323–330.

author's personal copy - paginas.fe.up.ptlepabe/publicacoes_restritas/papers/lmm/... ·...

Documents