Energy Procedia 4 (2011) Energy Procedia 00 (2010) GHGT-10. Qing Xu, Gary Rochelle 1 *

Transcription

1 Energy rocedia 4 (011) Energy rocedia 00 (010) Energy rocedia GHG-10 otal ressure and CO Solubility at High emperature in Aqueous Amines Qing Xu, Gary Rochelle 1 * Department of Chemical Engineering, he University of exas at Austin, 1 University Station, Austin, X 7871, USA Elsevier use only: Received date here; revised date here; accepted date here Abstract otal pressure was measured in CO loaded aqueous monoethanolamine, piperazine, 1-methyl-piperazine, -methyl-piperazine, and diglycolamine at 8 to 191 C from 115 to 819 ka. CO solubility is estimated from these data. Empirical models have been developed to predict the CO partial pressure of these amine solutions from 40 to 160 o C. he heat of CO absorption derived from these models varies from 66 kj/mol for piperazine and its derivatives to 71 and 73 kj/mol for monoethanolamine and diglycolamine and does not vary significantly with temperature. c ublished Elsevier Ltd. by Elsevier All rights Ltd. reserved Open access under CC BY-NC-ND license. Keywords: CO capture, CO solubility, aqueous amine, high temperature, high pressure. 1. Introduction Amine scrubbing will be an important technology for CO capture form coal-fired power plants. Various aqueous amines have been used for absorption. 7 m (molal, gmol/ kg water, 30 wt %) monoethanolamine (MEA) is considered the baseline solvent. revious studies have shown that concentrated piperazine (Z) is very promising, because of greater CO capacity and absorption rate and greatly reduced thermal and oxidative degradation [1]. Improved energy performance can be achieved by elevating the stripping temperature and pressure []. MEA is not thermally resistant so the stripper temperature cannot exceed 10ºC. However, Z, diglycolamine (DGA) and Z derivatives can be used up to 150 ºC without significant degradation. herefore thermodynamic data at high temperature will be especially useful in the design of stripper, multi-stage flash, reclaimer and other high temperature processes. Mid-temperature thermodynamics can also be interpolated from high and low temperature data. For MEA, CO solubility at high temperature is available from Jou et al. in 30 wt% MEA up to 150 C and 0,000 ka [3]. hese data will be compared to the CO solubility in MEA in this work to validate the experimental methods. here is little data available for Z, DGA and Z derivatives at high temperature and pressure. * Corresponding author. el.: ; fax: address: gtr@che.utexas.edu. doi: /j.egypro

2 118 Q. Author Xu, G. name Rochelle / Energy / Energy rocedia rocedia 00 (010) 4 (011) Experimental Methods.1 Apparatus.1.1 Calorimeter he first measurements of total pressure were performed using a 400 ml stainless steel calorimeter (by arr Instrument) as the equilibrium cell (Figure 1). ressure was measured with a Validyne D15 transducer, calibrated by heating water and correlating the readings with known water vapor pressures from DIR [4]. he voltage of the heating tape was manually controlled by a power controller to maintain selected temperature. An Omega K type thermocouple was installed inside the thermal well and an Omega 4001A was used as the temperature indicator. Air motor for the agitator DAQ system Heating tape and insulation Vap. Liq. ower control Heating jacket Vap. Liq. 400 ml SS Calorimeter 500 ml SS Autoclave Figure 1: otal ressure Measurement Figure : otal ressure Measurement with a Calorimeter with an Autoclave.1. Autoclave As shown in Figure, a 500 ml stainless steel autoclave (ZipperClave, by Autoclave Engineers) was used as the equilibrium cell for most of the measurements. Closure was effected by a resilient spring member inserted through a circumferential groove in the body and cover [5]. A magnetic hollow shaft agitator (MAG075, MagneDrive II Series, by Autoclave Engineers) was used to get equilibrium without leaking to the atmosphere. A compressed air motor (AM-NCC-16, by Gast ) provided agitation from 100 to 500 rpm. he agitator circulates both liquid and vapor phases. emperature was controlled by a Fuji Electric XZ-4 temperature controller, with connection to a K-type thermocouple placed inside the thermal well of the autoclave. A pressure transducer (Druck X 611, 0-30 bar absolute) was connected to a signal converter and a data logger used for to record data. he pressure indicator was calibrated by a dead weight pressure tester (S/N 19189/78, by Budenberg Volumetrics, INC.). he 1 run with MEA and 3 runs with Z were conducted in the calorimeter. he autoclave was used in all the other experiments. he results do not show obvious differences between the two apparatuses.. rocedure Before each run, 300 to 330 ml of CO loaded aqueous amine was prepared and added into the equilibrium cell. o avoid the effects of O, N was used to purge air and then the cell was sealed. he initial pressure of N and temperature were recorded for correction. hen the cell was heated. Recording of both temperatures and pressures started at around 100ºC. After holding at temperature for at least 30 min or until the pressure did not change for 10 min, the system was assumed to be at equilibrium. he solution was heated to about 160ºC (or higher for Z) and then cooled down to 100ºC. Data were taken during both heating and cooling processes to make the interval 10ºC in general. Liquid samples were collected before and after each experiment at room temperature and analyzed by total inorganic carbon analyzer and acid titration.

3 Q. Xu, Author G. Rochelle name / / Energy rocedia00 4 (010) (011) Analytical Methods otal Inorganic Carbon (IC) he concentration of CO in solution was determined by IC analysis. he liquid samples collected before and after each run were diluted by a factor of 100. About L diluted sample was injected into a CO analyzer (Model 55, Horiba IR 000). Details can be found in Appendix B. of Hilliard [6]. Acid itration he total alkalinity of solution was determined by acid titration using a Metrohm-eak 835 itrando equipped with an automatic dispenser, Metrohm-eak 801 stirrer, and 3M KCl ph probe. Details are available in Appendix A.3 of Hilliard [6] and Appendix F of Sexton [7]..4 Chemicals CO is C (Clinical urity) 99.5% by Matheson rigas. MEA is 99% by Acros Organics. Z is 99% by Sigma- Aldrich. 1-methyl-piperazine (1MZ) is 99+% by Acros Organics. -methyl-piperazine (MZ) is 99% by AK Scientific. Diglycolamine (DGA ) is 99% by Huntsman. DDI water was used for solution preparation. 3. Results and Discussion he total pressure of amine-h O-CO system is t = meas - N, where meas is the measured pressure and N is the N partial pressure; N was assumed to be an ideal gas so =, (K) (K), 0 stands for the initial condition. = = x x (1) HO and amine : partial pressure of water and amine. * HO and * amine: pressure of pure water and pure amine at, from DIR [4]. he mole fractions x HO +x amine =1, assuming there is water and total amine but no free CO in the solution. For Z derivatives and DGA, amine was ignored. Liquid analysis gives the total CO loading at room temperature, from which high temperature liquid loadings were corrected by the estimated amount of CO in the vapor Results able 1 lists the measured total pressure and calculated CO partial pressure over aqueous MEA, Z, 1MZ, MZ, Z/MZ and DGA. CO loading is defined as the mol CO /mol alkalinity. able 1: CO Solubility and otal ressure in MEA, Z, 1MZ, MZ, Z/MZ and DGA Amine Loading CO t Amine Loading CO t Amine Loading CO t m ºC ka ka m ºC ka ka m ºC ka ka MEA Z

4 4 10 Q. Author Xu, G. name Rochelle / Energy / Energy rocedia rocedia 00 (010) 4 (011) Amine Loading CO t Amine Loading CO t Amine Loading CO t m ºC ka ka m ºC ka ka m ºC ka ka MZ MZ Z/MZ, m is the total moles of Z and MZ per kg water. Z and MZ have approximately the same concentration.

5 Q. Xu, Author G. Rochelle name / / Energy rocedia00 4 (010) (011) Amine Loading CO t Amine Loading CO t Amine Loading CO t m ºC ka ka m ºC ka ka m ºC ka ka DGA Empirical Models Empirical models were regressed based on the data in this work and low temperature literature data for each amine. Literature data includes: C MEA data by Hilliard [6], Dugas et al. [8] and Jou et al. [3]; C Z data from Hilliard [6], Dugas et al. [8], Ermatchkov et al. [9] and Nguyen et al [10]; C 1MZ, MZ, Z/MZ and DGA data by Chen et al. [11, 1]. able : Empirical Correlation of CO artial pressure ( CO, a) with loading (α, gmol CO /equiv. alkalinity) and (K). () = Amine a b c d e f R MEA 39.3± ± (19.0±3.6) 1105± ± Z 35.5± ±17 0 -(.4±3.1) 470± ± MZ 35.± ±34 -(6.4±3.5) ± MZ 39.8± ±6 -(19.6±3.) -(8.1±.6) 14509± Z/MZ 41.± ±38 -(7.0±.8) 7.3± ± DGA 8.1± ± ±10 -(115±18) -(509±3383) 50113± Figure 3 shows a favorable comparison of the MEA data with data by Jou et al.[3]. 1.0E+7 CO (a) 1.0E+6 1.0E+5 1.0E C 10 C 100 C Open points: Jou et al., 1995 Filled points: this work Lines: empirical model CO Loading (moles/mea) Figure 3: Comparison of CO Solubility in MEA at 100, 10 and 150 C with Jou et al. [3] Figures 4 through 9 show the temperature and loading dependence of CO solubility in MEA, Z, 1MZ, MZ, Z/MZ and DGA, respectively. In Figure 4 the data is compared with the MEA Aspen model prediction by Hilliard [6]. he Hilliard model is good for C within loading. At higher than 1, it over predicts the CO partial pressure. In Figure 5 the data is compared with the Z Aspen model prediction by Frailie et al. [13]. he Frailie model predicts well for 40- with loading. At higher than 0.45 loading, there is no experimental data and the model prediction is higher than the empirical correlation.

6 6 1 Q. Author Xu, G. name Rochelle / Energy / Energy rocedia rocedia 00 (010) 4 (011) CO (a) 1.0E+7 ln 1.0E+6 1.0E+5 1.0E+4 1.0E+3 1.0E+ 1.0E+1 1.0E CO Loading (gmol/mea) Figure 4: CO Solubility in m MEA, revious Work by Hilliard [6], Dugas et al. [8] and Jou et al.[3]. CO (a) 1.0E+7 1.0E+6 1.0E+5 1.0E+4 1.0E+3 1.0E+ 1.0E CO 1 10 C 100 C 60 C ln CO 1 10 C 100 C 60 C Open points: previous work Filled points: this work Solid lines: empirical model Dashed lines: 7 m MEA Hilliard Aspen model [6] Open points: previous work Filled points: this work Solid lines: empirical model Dashed lines: 8 m Z, Frailie Aspen Model [13] CO Loading (gmol/equiv. Z) Figure 5: CO Solubility in m Z, revious: Hilliard [6], Dugas [8] Ermatchkov [9] & Nguyen [10]

7 Q. Xu, Author G. Rochelle name / / Energy rocedia00 4 (010) (011) E+07 1.E+07 1.E+06 1.E+06 1.E C 1.E C CO (a) 1.E+04 1.E+03 CO (a) 1.E+04 1.E+03 1.E+0 1.E CO Loading (gmol/mol Alk.) Figure 6: CO Solubility in 1MZ Filled points: this work; Open pts: [11]; curves: ln CO E+07 1.E+0 1.E CO Loading (gmole/mol Alk.) Figure 7: CO Solubility in MZ Filled points: this work; Open pts: [11]; curves: 1554 ln CO E+7 1.E+06 1E+6 CO (a) 1.E+05 1.E+04 1.E C CO (a) 1E+5 1E+4 1E+3 10 C 1.E+0 1.E CO Loading (gmol/mol Alk.) Figure 8: CO Solubility in Z/MZ Filled points: this work; Open pts: [11]; curves: 1998 ln CO E+ 1E CO Loading (gmole/mol Alk.) Figure 9: CO Solubility in DGA Filled points: this work; Open pts: [1]; curves: 757 ln CO 3.3 Heat of Absorption of CO into Aqueous Amines According to the Gibbs-Helmholtz Equation, (),, ΔH abs is the heat of absorption of CO into aqueous amines. able 3 gives ΔH abs derived from the empirical models. Within the precision of these measurements and estimates, the heat of CO absorption is independent of temperature and amine concentration, but varies with CO loading. he statistics of the empirical regression suggest that the heat of absorption is determined with a precision of to 4 kj/mol.

8 8 14 Q. Author Xu, G. name Rochelle / Energy / Energy rocedia rocedia 00 (010) 4 (011) Conclusions and recommendations CO solubility data was obtained using total pressure measurements. Empirical models as a function of temperature and loading were developed for CO solubility from 40 to 191 C in aqueous MEA, Z, 1MZ, MZ, Z/MZ and DGA. he high temperature CO solubility data for MEA is comparable with that by Jou et al. [3]. he high temperature data is also compatible with previous low temperature data. For MEA and Z, amine concentration does not have obvious effects on the CO solubility. he heat of CO absorption derived from these models varies from 66 kj/mol for Z and its derivatives to 71 and 73 for MEA and DGA and does not vary significantly with temperature. able 3: Comparison of the Heat of Absorption of CO Solvent ΔH abs (J/mol CO ) ΔH abs (kj/mol CO )* Mid-loading** m MEA R ( ) 71± m Z R ( ) 66± m 1MZ R ( ) 69± m MZ R ( ) 66± /4 m Z/MZ R ( ) 66± m DGA R ( ) 73± *: he heat of absorption of CO at mid-loading. **: he loading where CO is 1.5 ka at 40 C, calculated from the empirical models. Acknowledgements his work was supported by the Luminant Carbon Management rogram. Some experiments were conducted by the undergraduate assistants Martin Metzner and Mychal Zipper. References [1] Freeman SA, Dugas R, Van Wagener D, Nguyen, Rochelle G. Carbon dioxide capture with concentrated, aqueous piperazine. IJGCC, 010, 4, [] Rochelle G. Amine Scrubbing for CO Capture. Science, 009, 35, [3] Jou, F.-Y.; Mather, A. E.; Otto, F. D., he solubility of CO in a 30 mass percent monoethanolamine solution. he Canadian Journal of Chemical Engineering, 1995, 73, [4] DIR, 1998-rovo, U: BYU DIR, hermophysical roperties Laboratory, 1998-Version [5] Autoclave Engineers, Zipperclave 500&1000 ml stirred reactor, [6] Hilliard MD. A redictive hermodynamic Model for an Aqueous Blend of otassium Carbonate, iperazine and Monoethanolamine for CO Capture from Flue Gas. he University of exas. h.d. Dissertation [7] Sexton AJ. Amine Oxidation in CO Capture rocesses. he University of exas. h.d. Dissertation [8] Dugas R., Rochelle G. Absorption and desorption rates of carbon dioxide with monoethanolamine and piperazine. GHG-9, Washington D.C [9] Ermatchkov V., Kamps A. -S., Speyer D. and Maurer G. Solubility of carbon dioxide in aqueous solutions of piperazine in the low gas loading region. J. Chem. Eng. Data, 006, 51 (5), [10] Nguyen., Hilliard M., Rochelle G. Amine Volatility in CO Capture. IJGCC. 010, in press. [11] Chen, X., Rochelle G. Aqueous iperazine Derivatives for CO Capture: Accurate Screening by a Wetted Wall Column. Chem. Eng. Res. Des. 010, submitted. [1] Chen, X., Closmann, F., Rochelle G. Accurate Screening of amines by the Wetted Wall Column. GHG-10, 010. [13] Frailie,., laza, JM., Van Wagener, DH., Rochelle, G. Modeling piperazine thermodynamics, GHG