6136
Ind. Eng. Chem. Res. 2004, 43, 6136-6141
SEPARATIONS Kinetics of Carbonyl Sulfide (COS) Absorption with Aqueous Solutions of Diethanolamine and Methyldiethanolamine Fatiha Amararene and Chakib Bouallou* Centre d’E Ä nerge´ tique (CENERG), Ecole Nationale Supe´ rieure des Mines de Paris, 60 Boulevard Saint Michel, 75006 Paris, France
A constant gas-liquid interface reactor was used to measure carbonyl sulfide (COS) absorption rates in alkanolamine aqueous solutions. Two alkanolamines were studied: a secondary alkanolamine, diethanolamine (DEA), and a tertiary alkanolamine, N-methyldiethanolamine (MDEA) in the temperature range 313-353 K and for alkanolamine mass fractions going from 0.05 to 0.50. The reactor used is Lewis type, modified with a nonrotating valve which shortens considerably the cell loading time. For COS absorption in DEA aqueous solutions, the limiting step of the reaction is the deprotonation of the zwitterion. The kinetics data for COS absorption into MDEA aqueous solutions is interpreted using the base-catalyzed mechanism for the hydrolysis of COS. The new kinetic data show that previously reported results concerning the COS absorption by MDEA aqueous solutions were overestimated. Introduction Industries and urbanization widely contribute to air pollution. The sulfur compounds produced by refineries and other operations of the energy industrial sector have serious consequences on the environment. Because of its economic and ecological advantages, natural gas becomes more attractive for many countries. However, after extraction, natural gas can neither be transported nor used commercially. First, it must be processed to remove acid gases (CO2, H2S) and impurities such as COS and RSH (mercaptans). Sulfur compounds are harmful toxic and corrosive gases; therefore, the industry has developed strict specifications for their transportation, limiting the total content of sulfur to 35-50 ppm. Separation of the various sulfur compounds is often achieved by absorption by a chemical solvent such as alkanolamines. A wide variety of alkanolamines such as monoethanolamine (MEA), diethanolamine (DEA), N-methyldiethanolamine (MDEA), and diglycolamine (DGA) have been used commercially. Despite abundant literature, only few works1,3-7 address the absorption kinetics for COS/DEA + H2O or COS/MDEA + H2O systems. The data that exist mainly concerns the kinetics in gas-unloaded solutions. The first study devoted to COS by Sharma1 was carried out with several primary and secondary alkanolamines, but only at one temperature (T ) 298 K) and only at one concentration (C ) 1000 mol‚m-3). It suggested that COS reaction mechanisms with primary and secondary alkanolamines were similar to that with CO2. Singh and Bullin2 studied COS absorption, using DGA aqueous solutions in a reactor which simulate a * To whom correspondence should be addressed. E-mail:
[email protected]. Tel.: 331-40519111. Fax: 33143265910.
stage of column absorption. The temperature varied from 307 to 322 K, the DGA concentration from 3220 to 5630 mol‚m-3, and the COS partial pressure from 37 to 394 kPa. From kinetic experiments using a wetted sphere apparatus Al-Ghawas et al.3 provide data for COS Henry’s constant and diffusion coefficient in MDEA aqueous solutions in a temperature range from 298 to 313 K, and with MDEA concentrations from 1259 to 2599 mol‚m-3. The kinetic results are well represented by base-catalyzed COS hydratation. Additional data were obtained by Little et al.4 who used a stirred cell reactor, with primary and secondary alkanolamines, over a temperature range of 283-333 K. These authors used a COS-N2O analogy to determine the COS physicochemicals parameters. They found that the reaction of COS with primary and secondary alkanolamines is well-represented by the zwitterion mechanism, the zwitterion deprotonation being the limiting step. Other studies (Little et al.5,6) were dedicated to COS absorption in aqueous solutions of tertiary alkanolamines and proposed absorption models. The temperature was fixed at 303 K for all experiments except those with MDEA aqueous solutions (293-323 K). The alkanolamine concentration was varied from 153 to 1011 mol‚m-3. Recently, the study by Hinderaker and Sandall7 provided kinetics data of COS absorption by 5 to 25 wt % DEA aqueous solutions and at temperatures between 298 and 348 K. These authors carried out measurements of solubility and diffusivity of COS and N2O in aqueous solutions of poly(ethylene glycol) at 298 K and confirmed the validity of the COS-N2O analogy. They found that the limiting step in the reaction of COS absorption in DEA aqueous solutions is the deprotonation of the zwitterion, resulting in overall third-order kinetics.
10.1021/ie030540f CCC: $27.50 © 2004 American Chemical Society Published on Web 08/12/2004
Ind. Eng. Chem. Res., Vol. 43, No. 19, 2004 6137
Figure 1. Flow diagram of the apparatus.
Experimental Section The apparatus (Figure 1) is a thermostated Lewistype reactor with a constant gas-liquid interface (15.34 ( 0.05) 10-4 m2. The temperature is controlled by circulating a thermostatic fluid through the glass double jacket. The reactor is closed at both ends by two metallic flanges. The liquid and gas phases are agitated respectively by a 4.25 × 10-2 m diameter six-bladed Rushton turbine and 4 × 10-2 m diameter propeller, and they are driven magnetically by a variable speed motor. A DRUCK pressure (0-250 kPa) is mounted on the upper flange and is calibrated at a temperature higher than the experiment temperature to avoid liquid condensation in its measuring chamber. A tube is mounted through the upper flange and allows one either to degas the cell or to connect it to a tank of COS gas. The lower flange is equipped with a temperature probe and a nonrotating stem valve. The standard reactor configuration was modified to expedite the loading of the solutions. In a previous study (Pani et al.8) which used the same apparatus, the solution was introduced by means of a hypodermic needle, and the loading time was significant. Also, four vertical baffles are placed inside the cell in order to avoid the formation of a vortex. The pressure transducer is calibrated within (42 Pa against a pressure calibration device. The temperature in the reactor is known within 0.02 K; it is calibrated against a 25 Ω platinum probe. A microcomputer is used to record both pressure and temperature signals. Water and alkanolamine are degassed independently and aqueous solutions are prepared under vacuum. The masses of water and alkanolamine are known by differential weighing. The flask containing the solution is connected to the reactor to allow the solution to transfer by gravity under vacuum. Accurate weighing of the flask before and after transfer yield the mass of solution
actually introduced in the cell, and the liquid-phase volume was calculated using the density. At a given temperature and vapor pressure PI, pure COS is introduced during a very short time in the upper part of the cell, the volume of which is noted as Vg. Then stirring is started, and the pressure drop resulting from absorption is recorded. The estimated maximum experimental error in the COS absorption rate is 8%. Chemicals. Twice-distilled water and reagent-grade MDEA are used. DEA is from Aldrich, with a certified minimum mass purity of 99%. MDEA is from Fluka, with a certified minimum mass purity of 98%. Carbonyl sulfide is from L’Air Liquide, with a certified volume purity of 99.997%. Results and Discussion On the basis of mass balance of COS for the gas phase, the influence on absorption kinetics of all chemical reactions between dissolved COS and reactants in solution is expressed by an enhancement factor E over physical absorption:
kLECCOS,inta )
-Vg dPCOS RT dt
(1)
The gas phase is assumed ideal and COS is completely consumed by the chemical reaction. At the interface, vapor-liquid equilibrium is assumed. The partial pressure PCOS is related to the concentration of unreacted dissolved COS by Henry’s law:
CCOS,int )
PCOSint HCOS
(2)
6138 Ind. Eng. Chem. Res., Vol. 43, No. 19, 2004
PCOS is obtained from the measured pressures:
PCOS ) PT - PI
(3)
The concentration of COS resulting from absorption does not change much the composition of the solution so kL, HCOS, and E remain constant with time, and integration of eq 1 yields
Ln
(
)
PT - PI ) -β(t - t0) PT,0 - PI
(4)
kLEaRT VgHCOS
(5)
where
β)
For each experiment, the enhancement factor E is obtained from β using eq 5. The density and the viscosity of the DEA and MDEA aqueous solutions are estimated using Hsu and Li9,10 correlations. The DEA diffusion coefficient is estimated from the correlation supplied by Hikita et al.11 The correlation given by Pani et al.8 is used for estimates of the MDEA diffusion coefficient. The COS-N2O analogy provides estimates of the COS diffusion coefficient and Henry’s law constant in alkanolamines aqueous solutions. The COS diffusion coefficient in water is obtained from Al-Ghawas et al.3 correlation. The N2O diffusion coefficient in water and in the aqueous solutions is found using Versteeg and van Swaaij12 correlations. Henry’s law constant of COS in water is calculated from Little et al.’s13 correlation. The model of Wang et al.14 is used for the calculation of N2O Henry’s law constant in alkanolamines solutions with the parameters of binary interactions given by Tsai et al.15 The mass-transfer coefficient kL is calculated using the following correlation established for our apparatus from N2O absorption experiments by MDEA aqueous solutions,
Sh ) 0.352Re0.618Sc0.434
(6)
where the dimensionless Sherwood (Sh), Reynolds (Re), and Schmidt (Sc) numbers are expressed by
Sh )
kLdcell , DN2O
2
Re )
dNdag , µ
Sc )
µ FDN2O
Kinetics for the COS/DEA + Water System. According to the literature, the reaction of COS in DEA aqueous solutions is generally represented by the zwitterion mechanism: k2,k-2
COS + (C2H4OH)2NH 798 (C2H4OH)2NH+COSkb,H
k 2O; b,DEA
(C2H4OH)2NH+COS- + b98 (C2H4OH)2NCOS- + bH+ b is for any base (alkanolamine, H2O). The zwitterion deprotonation is often considered as being the limiting step, implying the participation of a second molecule of DEA. Therefore, it was assumed that the reaction of COS with DEA aqueous solutions is irreversible and of first order with respect to COS and
Figure 2. Rate constants of COS absorption in aqueous DEA. Comparison with previous data.
second order with respect to DEA. In these conditions, the rate of reaction between COS and DEA is given by
rCOS ) kCOS-DEACCOSCDEA2
(7)
and the enhancement factor is equal to the Hatta number:
E ≈ Ha )
x
kCOS-DEACDEA2DCOS kL2
(8)
All the reactions were studied with a large excess of alkanolamine over COS and gave good pseudo-firstorder according to
rCOS ) kobsCCOS
(9)
The determination of the enhancement factor allows the calculation of the rate constants:
E2kL2 kobs ) DCOS
(10)
The rate constant kCOS-DEA is calculated by the relation
kCOS-DEA )
kobs CDEA2
(11)
Only the results for DEA and MDEA concentrations ranging from 5 to 25 wt % are used for the calculation of the rate constants that ensure the validity of the COS-N2O analogy. A comparison between our experimental rate constants, and the data available in the literature, shows that our results (Figure 2) are consistent with the data of Hinderaker and Sandall7 and Little et al.4 at low DEA concentrations. The difference between our kinetic results and those of Hinderaker and Sandall7 becomes more significant from a concentration of 1400 mol‚m-3 (T ) 313 K); beyond they reported higher values. For a temperature of 333 K our results are slightly lower compared with the data of Little et al.4 whereas those of Hinderaker and Sandall7 are highest. The origin of these deviations can be very probably attributed to an underestimation of the COS solubility and an overes-
Ind. Eng. Chem. Res., Vol. 43, No. 19, 2004 6139 Table 2. Comparison with Activation Energies of COS Absorption in Aqueous DEA from Previous Works references Hinderaker and our results
CDEA/mol m-3 Ea/kJ mol-1
T/K
Sandall7
293-348 313-353
476-2440 474-2400
52.35 48.10
taking into account the influence, at the same time, of the temperature and the concentration:
(
kCOS-DEA ) (2121 - 0.4 × CDEA) exp -
Figure 3. Arrhenius law for the COS absorption in aqueous DEA and aqueous MDEA solutions. Table 1. Conditions of COS Absorption Kinetics by DEA Aqueous Solutions T/K
CDEA/mol m-3
Vg/m3
PT,0 - PI/Pa
E
kobs/s-1
313.88 313.82 323.68 333.35 333.45 352.90
474.17 474.18 491.16 498.33 498.30 183.54
181.83 183.60 183.47 181.10 182.20 180.59
115761 115248 111776 108922 101563 74567
3.00 2.86 4.06 4.85 4.85 7.73
4.62 4.26 9.20 14.21 14.15 41.21
313.98 323.75 333.35 333.54 353.11
1433.06 1422.54 1425.50 1425.36 1409.64
179.98 183.06 182.62 181.59 181.09
113706 103780 97232 98818 77670
8.08 10.01 12.01 12.98 19.16
28.30 48.69 77.45 89.27 226.11
313.94 323.73 333.42 333.55 352.91 352.90
2425.56 2388.52 2392.72 2412.32 2359.66 2375.63
183.40 182.98 182.63 182.68 180.31 180.38
101154 108683 99454 99953 59268 66023
13.85 16.92 20.27 20.51 32.21 31.36
70.65 117.57 185.99 189.84 555.67 526.58
timation of the diffusivity values in rather concentrated amine solutions. In addition, the COS-N2O analogy is only valid for relatively diluted aqueous solutions. The results of Hinderaker and Sandall7 are somewhat higher than those found in the literature. On one hand, the values for the COS solubility used by these authors are on average 6% smaller than those used in this study. On the other hand, the average deviation between the various diffusivity values used in the comparative studies here can be estimated at 13%. The combined influence of these two parameters on the determination of β (eq 5) thus results in an uncertainty of 12.5% ((∆β/ β) ) 1/2(∆Dcos/Dcos) + (∆Hcos/Hcos)). Thus, while considering the most unfavorable case at 313 K, the relative deviation between our result and that of Hinderaker and Sandall7 in terms of the rate constants for COS absorption in aqueous DEA, which was 35.0%, becomes 14.5%. The regression of the experimental rate constants (Table 1) obtained by eq 10 according to 1/T for the three studied concentrations is represented in Figure 3. The rate constants depend on the concentration of DEA. This observation does not appear in the literature, which often present kinetic laws for the complete concentration range. Consequently, we developed eq 12 that allows
5785 T
)
(12)
This equation was developed by extrapolating the physicochemicals and thermodynamic parameters. We have observed (Table 2) that the activation energy obtained in this study is in agreement with that determined by Hinderaker and Sandall.7 Assuming the participation of water as a second base in the zwitterion deprotonation, the rate constants become
kobs,cal )
CDEA 1 1 + k2 kH2OCH2O + kDEACDEA
(13)
The rate constants kH2O and kDEA, being defined by
kH2O ) kDEA )
k2kb,H2O k-2 k2kb,DEA k-2
(14)
(15)
kH2O and kDEA represent the zwitterion deprotonation rate constants (m6‚mol-2‚s-1), which were determined by adjusting the experimental constants (kobs) with those given by eq 13. This procedure gave the following equations for zwitterion deprotonation rate constants:
(-5772 T ) -6333 exp( T )
kDEA ) 9.63 × 102 exp
(16)
kH2O ) 4.55 × 102
(17)
The term 1/k2 can be neglected (values of k2 were on the order of 108 m3‚mol-1‚s-1). The rate constants kDEA obtained here agree with activation energy values of Little et al.,4 whereas the rate constants kH2O obtained by these same authors from only two points is lower than ours. The restricted number of points retained in the case of Little et al.4 explains this difference. Moreover, the influence of DEA in the deprotonation of the zwitterions was more important compared with the influence of water (Figure 4). Kinetics for the COS/MDEA + Water System. Most of the previous studies on COS absorption by aqueous solutions of tertiary alkanolamines consider base catalysis for the reaction of COS with water:
COS + C5H13NO2 + H2O T C5H13O2NH+ + HCO2SThe order with regard to MDEA is assumed 1. For the COS/MDEA + water system, conditions of the
6140 Ind. Eng. Chem. Res., Vol. 43, No. 19, 2004
Figure 4. Comparison of the zwitterions deprotonation rate constants with the previous data.
Table 4. Kinetic Laws of COS Reaction According to MDEA Concentration
Table 3. Conditions of COS Absorption Kinetics by MDEA Aqueous Solutions T/K
CMDEA/mol m-3
Vg/m3
PT,0 - PI/Pa
E
kobs/s-1
313.86 313.80 323.66 333.21 352.91 352.73
445.08 445.09 471.53 469.30 464.00 464.05
183.57 181.94 183.57 181.68 180.51 180.36
117484 112456 112206 106711 82879 76586
1.25 1.21 1.53 1.89 3.14 3.02
0.40 1.17 0.98 1.92 6.73 6.22
313.98 323.60 333.57 352.94
1432.82 1426.09 1418.48 1401.85
184.07 182.75 182.01 180.40
133011 114238 108944 79453
1.54 2.09 2.56 4.25
0.80 2.00 3.46 11.06
313.80 313.90 323.60 333.46 333.54 352.68 353.00
2408.22 2408.10 2395.62 2381.99 2381.87 2323.98 2352.16
180.40 183.86 183.35 181.76 182.34 179.44 181.70
136760 118046 110440 112683 105536 83390 81930
1.84 2.15 2.60 3.21 3.36 6.47 6.40
1.09 1.23 2.77 4.63 5.09 22.55 22.46
pseudo-first-order are not satisfied for all experiments, notably those realized at lower concentrations and lower temperatures. We have considered therefore a general form of the enhancement factor:
E)
Ha thHa
(18)
The rate constant kobs is calculated from eq 10. Summary of the results of COS absorption in aqueous MDEA solutions are presented in Table 3. The plot of Ln(kCOS-MDEA) vs (1/T) for the three studied concentrations (0.05, 0.15, and 0.25) is presented in Figure 3. We have not been able to determine in this case a general kinetic law taking into account the influence of MDEA concentration and temperature. We have proposed the following kinetic law for each of the studied concentrations:
kCOS-MDEA ) ACOS-MDEA exp
(
)
BCOS-MDEA T
Figure 5. Observed rate constants of COS absorption in aqueous MDEA.
(19)
ACOS-MDEA and BCOS-MDEA are given in Table 4. We have compared our experimental results with those of Al-Ghawas et al.3 at 313 K. As shown in Figure 5, the rate constants given by Al-Ghawas et al.3 are higher than our experimental data. The study carried out by Little et al.5 also showed that the data of Al-
T/K
CMDEA/mol m-3
ACOS-MDEA/s-1
BCOS-MDEA/K
313.83 323.66 333.21 352.82
445.08 471.53 469.30 464.03
4.17 × 107
7688
313.98 323.60 333.57 352.94
1432.82 1426.09 1418.48 1401.85
9.25 × 106
7358
313.85 323.64 333.54 352.84
2408.16 2395.62 2381.87 2338.07
1.77 × 108
8357
Ghawas et al.3 are high. Theses discrepancies are explained by the presence of contaminant (primary and secondary alkanolamines) in MDEA. Our results are in good agreement in terms of activation energy with Little et al.5 By considering the same operating conditions, we have compared the COS rate absorption in aqueous solutions of DEA and aqueous solutions of MDEA. For a given concentration, COS rate absorption in DEA aqueous solutions is higher compared with that in MDEA aqueous solutions. This is explained by the high reactivity of DEA compared with MDEA. Conclusion Kinetics of COS absorption by DEA and MDEA aqueous solutions have been studied in the extended range of temperatures from 296 to 353 K, and of DEA concentrations from 5 to 40 wt % and of MDEA from 5 to 50 wt %. In the case of the COS/DEA + water system, kinetic results obtained were consistent with the literature data for temperatures between 313 and 333 K. By assuming a partial order of 2 with regard to DEA, we have determined a general kinetic law taking into account the influence of DEA concentration and the temperature. The results of COS absorption in DEA aqueous solutions show that the limiting step in the reaction is the deprotonation of the zwitterion. Rate constants of the zwitterion deprotonation kDEA and kH2O were compared in this study, and we found that the influence of water in the deprotonation of the zwitterions is lower than that of the DEA.
Ind. Eng. Chem. Res., Vol. 43, No. 19, 2004 6141
The limited data available in the open literature regardless of the COS absorption in MDEA aqueous solutions enabled us to confirm that the rate constants given by Al-Ghawas et al.3 are overestimated. Our kinetic results are consistent in terms of activation energy with those of Little et al.,5 but we observed a significant difference in the values of the absolute rate constants. A rigorous transfer model with chemical reactions will certainly make it possible to explain these discrepancies which can have origin in the presence of secondary reactions, impurities in solvent, and the validity of the considered mechanism. List of Symbols a ) interface area (m2) C ) concentration (mol‚m-3) D ) diffusion coefficient (m2.s-1) dag ) stirrer diameter (m) dcell ) internal diameter of the Lewis cell (m) E ) enhancement factor H ) molar scale Henry’s law constant (Pa‚m3‚mol-1) Ha ) Hatta number k ) reaction rate constants kL ) liquid-side mass-transfer coefficient of unreacted COS (m‚s-1) P ) pressure (Pa) PI ) vapor pressure over the aqueous alkanolamine solutions (Pa) rCOS ) rate of the reaction between COS and DEA (mol‚m-3‚s-1) R ) gas constant (8.3143 J‚K-1‚mol-1) Re ) (FNdag2/µ) ) Reynolds number T ) absolute temperature (K) t ) time (s) V ) volume (m3) Sc ) (µ/FDN2O) ) Schmidt number Sh ) (kLdcell/DN2O) ) Sherwood number Greek Letters β ) slope (s-1) µ ) viscosity of the aqueous solutions (Pa‚s) F ) density of the aqueous solutions (g‚cm-3) Subscripts cal ) calculated exp ) experimental g ) gas int ) interface obs ) observed overall rate constants T ) total 0 ) initial
Literature Cited (1) Sharma, M. M. Kinetics of Reactions of Carbonyl Sulphide and Carbon Dioxide with Amines and Catalysis by Bronsted Bases of the Hydrolysis of COS. Trans. Faraday Soc. 1965, 61, 681. (2) Singh, J.; Bullin,A. Determination of Rate Constants for the Reaction between DiGlycolAmine and Carbonyl Sulphide. Gas Sep. Purif. 1988, 2, 131. (3) Al-Ghawas, H. A.; Ruiz-Ibanez, G.; Sandall, O. C. Absorption of Carbonyl Sulphide in Aqueous MethylDiEthanolAmine. Chem. Eng. Sci. 1989, 44, 631. (4) Littel, R. J.; Versteeg, G. F.; van Swaaij, W. P. M. Kinetics of COS with Primary and Secondary Amines in Aqueous Solutions. AIChE J. 1992, 38, 244. (5) Little, R. J.; Versteeg, G. F.; van Swaaij, W. P. M. Kinetics Study of COS with Tertiary Alkanolamine Solutions. 1. Experiments in an Intensely Stirred Batch Reactor. Ind. Eng. Chem. Res. 1992, 31, 1262. (6) Little, R. J.; Versteeg, G. F.; van Swaaij, W. P. M. Kinetics Study of COS with Tertiary Alkanolamine Solutions. 2. Modeling and Experiments in a Stirred Cell Reactor. Ind. Eng. Chem. Res. 1992, 31, 1269. (7) Hinderaker, G.; Sandall, O. C. Absorption of Carbonyl Sulphide in Aqueous Diethanolamine. Chem. Eng. Sci. 2000, 55, 5813. (8) Pani, F.; Gaunand, A.; Cadours, R.; Bouallou, C.; Richon, D. Kinetics of Absorption of CO2 in Concentrated Aqueous MethylDiEthanolAmine Solutions in the Range 296 K to 343 K. J. Chem. Eng. Data. 1997, 42, 353. (9) Hsu, C. H.; Li, M. H. Viscosities of aqueous blended amines. J. Chem. Eng. Data 1997, 42, 714. (10) Hsu, C. H.; Li, M. H. Densities of aqueous blended amines. J. Chem. Eng. Data 1997, 42, 502. (11) Hikita, H.; Ishikawa, H.; Uku, K.; Murakami, T. Diffusivity of Mono-Di and TriEthanolAmines in Aqueous Solutions of Amine. J. Chem. Eng. Data 1980, 25, 324. (12) 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. (13) Littel, R. J.; Versteeg, G. F.; van Swaaij, W. P. M. Solubility and Diffusivity Data for the Absorption of COS, CO2, and N2O in Amine Solutions. J. Chem. Eng. Data 1992, 37, 49. (14) Wang, Y. W.; Xu, S.; Otto, F. D.; Mather, A. E. Solubility of N2O in Alkanolamines and in Mixed Solvents. Chem. Eng. J. 1992, 48, 31. (15) Tsai, T. C.; Ko, J. J.; Wang, H. M.; Lin, C. Y.; Li, M. H. Solubility of Nitrous Oxide in Alkanolamine Aqueous Solutions. J. Chem. Eng. Data 2000, 45, 341. (16) Sandall, O. V. Kinetics of Sulphur Species - Hydrocarbon - Aqueous Amine Systems. Gas Processors Association, Project 962, RR-182, 2002.
Received for review June 30, 2003 Revised manuscript received March 29, 2004 Accepted April 29, 2004 IE030540F