Scholarly article on topic 'Capacity and Kinetics of Solvents for Post-Combustion CO2 Capture'

Capacity and Kinetics of Solvents for Post-Combustion CO2 Capture Academic research paper on "Chemical engineering"

CC BY-NC-ND
0
0
Share paper
Academic journal
Energy Procedia
OECD Field of science
Keywords
{"Carbon dioxide" / Absorption / Desorption / "Cyclic capacity" / "Solvent selection"}

Abstract of research paper on Chemical engineering, author of scientific article — Peter Brœder, Hallvard F. Svendsen

Abstract In this work the experimental results of initial solvent characterization for the EU project CESAR are presented. A range of single and mixed amine solutions were screened and the focus was on the effect of the loading capacity kinetics and cyclic capacity depending on the amine concentration, amine ratio, and type of amine. Some of the solutions have much lower water content than the commonly used commercial solutions and the effect of this is discussed. Semi-quantitative screening tests were done enabling estimation of rate of absorption over the whole CO2 loading range and also of the cyclic capacity. For the tests a laboratory scale bubble reactor was used and they are typically early and relatively rapid tests that, together with available environmental data, can be used for the selection of solvents for detailed characteristics. The final target is to find solvents for the pilot plant campaigns in the CESAR project. The results were compared to the commonly used 5M (30 w%) monoethanolamine (MEA) solution.

Academic research paper on topic "Capacity and Kinetics of Solvents for Post-Combustion CO2 Capture"

Available online at www.sciencedirect.com

Procedía

Energy

ELSEVIER

Energy Procedía 23 (2012) 45 - 54

6th Trondheim Conference on CO2 Capture, Transport and Storage (TCCS-6)

Capacity and kinetics of solvents for post-combustion CO2

capture

Norwegian University of Science and Technology (NTNU) Department of Chemical Engineering, Sem Swlands vei 4, Trondheim

NO-7491, Norway

In this work the experimental results of initial solvent characterization for the EU project CESAR are presented. A range of single and mixed amine solutions were screened and the focus was on the effect of the loading capacity kinetics and cyclic capacity depending on the amine concentration, amine ratio, and type of amine. Some of the solutions have much lower water content than the commonly used commercial solutions and the effect of this is discussed. Semi-quantitative screening tests were done enabling estimation of rate of absorption over the whole CO2 loading range and also of the cyclic capacity. For the tests a laboratory scale bubble reactor was used and they are typically early and relatively rapid tests that, together with available environmental data, can be used for the selection of solvents for detailed characteristics. The final target is to find solvents for the pilot plant campaigns in the CESAR project. The results were compared to the commonly used 5 M (30 w%) monoethanolamine (MEA) solution.

© 2012 The Authors. Published by Elsevier Ltd. Selection and/or peer-review under responsibility of SINTEF Energi AS Keywords: Carbon dioxide; Absorption; Desorption; Cyclic capacity, Solvent selection

1. Introduction

Global warming is established as one of the most serious problems facing the earth at present. Anthropogenic CO2 is one of the gases causing this problem, and large sources are traffic, fossil fuel based power production, iron, steel and cement production and the metal and alloy production based on reduction with carbon. CO2 absorption in aqueous amine solutions is the most widely used process in the industry. However, capital and operating costs are high, and in particular the heat required for absorbent regeneration requires large quantities of energy, causing large drops in efficiency for the industry. One of the options to decrease the operational and investment costs is the usage of a highly effective amine solution. The target in 2050 is to double the production of energy compared to today's level with half of the CO2 emitted into the atmosphere. A high rate of absorption is necessary to reduce the height of the

* Corresponding author. Tel.: +47-47-264-478; fax: +47-73-594-080.

E-mail address: peter.bruder@chemeng.ntnu.no

Peter Brüder*, Hallvard F. Svendsen

Abstract

1876-6102 © 2012 The Authors. Published by Elsevier Ltd. Selection and/or peer-review under responsibility of SINTEF Energi AS doi:10.1016/j.egypro.2012.06.028

absorber and save on the investment and maintenance costs and at the same time reach a close approach to equilibrium, thereby saving on regeneration heat. High capacity of solvent of course reduces the electrical energy needed for the solvent recirculation at the same viscosity. The environmental point is to have lower emissions, and less toxic degradation products. The primary and secondary amines have fast kinetics, but the theoretical maximum loading (ratio between the CO2 and amine groups) is only 0.5 mol CO2/NHx In the literature information on the primary monoethanolamine (MEA), which is the most frequently used amine [1-4], is plentiful, but still progress is made in processes using this amine. At present the most promising secondary amine for post combustion processes seems to be the cyclic diamine piperazine (PZ), which is suggested alone and also as an activator for blended systems [5-7].

Table 1. The used chemicals

Amine CAS Purity Structure Producer

2-Amino-2-(hydroxymethyl)-1,3-propanediol (TRIS) 77-86-1 99% Sigma -Aldrich

Dimethyl- monoethanolamine (DMMEA) 108-01-0 99% Riedel-de Haën

2-amino-2-methyl-1-propanol (AMP) 124-68-5 97% .Je - Fluka

Diethyl- ethanolamine (DEEA) 100-37-8 99.5% Sigma -Aldrich

Piperazine (PZ) 110-85-0 98% "O" Fluka

Ethanolamine (MEA) 141-43-5 99% Sigma -Aldrich

3-amino-1-methylaminopropane (MAPA) 6291-84-5 98% Sigma -Aldrich

Tertiary amines do not react directly with CO2 to form carbamate and due to the different reaction mechanism the kinetics are slower. The theoretical capacity is higher because the maximum loading is 1.

They are used mostly in blended systems with primary or secondary amines as activators. On popular amine used for high pressure applications is methyldiethanolamine or MDEA [8]. Sterically hindered primary or secondary amines were first proposed by Sharma [9] and 2-amino-2-methyl-1-propanol or AMh [10,11] is commonly used. They have a very low tendency to form carbamates with CO2 due to geometric protection of the amine group by other functional groups. As a consequence, the reaction kinetics are slower, but the theoretical absorption loading is 1, similar to tertiary amines [12].

Environmentally, generally the primary and secondary amines are more readily bio-degradable than tertiary amines [13]. Sterically and ring-forming amines like AMh and hZ are exceptions being more stable and less biodegradable.

2. Experimental section

The chemicals tested are given in Table 1 and were used for the experiments without further purification and dissolved in deionized water.

2.1. Screening apparatus

A sketch of the apparatus is shown in Fig. 1. The gases, N2 (99.6%) and CO2 (99.999%), were supplied by AGA Gas GmbH. The analysis of CO2 was done with the standard BaCl2 method, see [14]. For the comparison of the absorption rate and the loading capacity a screening apparatus was used [15] and the results were compared to 30 wt% (5 M) monoethanolamine (MEA) solutions. The absorption tests were performed at 40 °C and the flows of N2 and CO2 were controlled by flow controllers (Bronkhorst High Tech— ± 0.1% full scale) and these regulated also their mixing ratio. The CO2 flow was set to 0.5 and the N2 to 4.5 Nl/min (dm3 at normal conditions). This gave a 10 % CO2 concentration in the inlet gas flow through the amine solution. The volume of absorbent solution was set to 750 ml. The CO2 content in the outlet gas after absorption was measured by an IR detector (Rosemount Binos 100 series 10 ± 0.1 %) and continuously logged by LabView software (Ver 6.0). An analyzer calibration run was performed before every measurement, and the two calibration points used were pure N2 and 10% CO2. The program registered the measured remaining CO2 in the gas stream, and the difference between 0.5 Nl/min and the registered value by the IR instrument was saved as a function of time. The measurements were stopped when the measured outlet gas CO2 concentration reached 95% of the initial (inlet) value. After the end of the absorption, a liquid sample was taken for measuring density and CO2 concentration in the end solution.

Desorption tests were performed at 80 °C, and the stripping gas was pure N2.

3. Results and discussion

The aim of the measurements was to get a semi-quantitative measure for the solvent CO2 capacity, the useful cyclic capacity (the difference between the maximum rich and the minimum lean loading at the given conditions) and in addition an estimate of the solvent kinetics over the whole loading range. The solvent capacities can be accurately determined as the program terminates at ~ 95 % of equilibrium values at 40 °C. A comparison of kinetics is more uncertain because of the continuously changing viscosity. This can results in different bubble size distributions, active area and retention time in the solution. The evaporation, mainly during desorption at 80 °C, leads to a somewhat variable solution depth in the reactor. Due to this, the measurement can only be considered semi-quantitative. In some cases it was not

possible to finish desorption, or even absorption, because precipitates were formed. Some operational difficulties were observed during the tests:

AbaorberJDesorber

Fig. 1. Schematic drawing of the screening apparatus

1. Between absorption and desorption 45 - 60 minutes were used to heat up the solution to the desorption temperature. The CO2 concentration was high after stopping the flow and possibly solid particles (or highly viscous solution) could plug the pores of the bubble distributor. In these cases the result was a high backpressure in front of the reactor as the nitrogen gas flow was set to the maximum already from the beginning of the desorption process.

2. If the solution was easy to strip, then, during heating up, the CO2 would be released continuously. After the N2 flow was started the stripping could be very rapid and if some narrow parts of the apparatus were unable to handle the increased flow (e.g. a flow-meter) this could lead to overpressure after the reactor.

3. Solids could form in the gas phase after a temperature decrease. Because the water bath was used only to heat up and keep the reactor temperature constant, sometimes the water coolers were not able to handle all the evaporated amine and water, and solids could form in the narrow parts. This was most visible in the flow-meter which could then not be operated. When this happened desorption could not be run and missing desorption values did not allow the estimation of cyclic capacity.

These problems mostly relate to the used equipment, but the build-ups of crystals in a longer term could be an issue in absorption plants too. The tested systems can be found in Table 2.

Table 2. The tested systems (mole/l)

Piperazine MEA MAPA

5, 100%* 5

TRIS (Trizma) 3:1, 2:2 3:1, 1.5:100%* 1.5:100%** 3:1

DMMEA 3:1, 5:2, 6:2, 4:4, 6:3, 5:3, 4:3, 6.5:1.5 5:2 3:2, 5:2, 5:1, 3:5, 4:4

AMP 2:3.5, 2.5:2.5, 3:1, 3:1.5, 3:2, 3:3, 4:1 3:1

DEEA 3:1, 3:3 5:2

* water was not added to the solution, the liquid phase was only MEA ** only half a mole of water was not added to the solution

—*— 2 AMP + 3,5 PZ —3M AMP + 2 M PZ 2.5M AMP + 2.5M PZ

—l— 3M AMP + 1M PZ —4M AMP + 1M PZ + 30% MEA

+ 3M DMMEA +2 M MAPA 3 M AMP + 1.5 M PZ

0.5 0.4

■a 0.2

■E 0 S

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

loadfmol C02/kg solution]

Fig. 2. Screening comparison of the AMh/hZ systems with DMMEA/MAhA and MEA

In Fig. 2 results for the AMh/hZ systems which could be tested in the screening apparatus are shown. Some of the systems formed solids when CO2 was added. As seen from the graph the three most concentrated systems could not be run to equilibrium and therefore the determination neither of maximum loading nor cyclic capacity could be done. This was because the apparatus clogged during measurement. Some observations on these systems are made. For the 3M AMP/2M hZ system all crystals were dissolved at 22 °C if the solution was mixed. Without mixing, very small amounts of solids were formed

during 12 - 16 hours and this would probably not create problems during stable industrial operation. The 2.5M AMP/2.5M PZ system without heating and mixing crystallizes at 22 °C but the solids can be dissolved. Solid formation starts around 33 - 35 °C. No precipitate formed by loading with CO2 at 40oC but the viscosity was high and the solution sticky. The 2M AMP/3.5M PZ mixture was full of crystals at 22 °C and all the solution became solid. The crystals were formed at 35 °C and the screening at 40°C could be run. No crystals were found during CO2 loading, but the solution appears to be sticky and highly viscous. The solution with the highest AMP/PZ concentration possible to test without problems in the equipment was the 3M AMP/1.5M PZ. From Fig. 2 it is seen that this system performs similarly to the DMMEA/MAPA 3M/2M system [16]. The 4M AMP/1M PZ system obtains a loading similar to 5 M MEA (5 mole/dm3 total solution), but the kinetics are much slower due to the AMP. The AMP/PZ system is characterized by forming precipitates both at high AMP and high PZ concentration.

3 M DMMEA + 1 MPZ 5 M DMMEA + 2MPZ 6.5 M DMMEA + 1.5 M PZ S M DMMEA + 3MPZ

-t- SMDMMEA + 2MPZ -t- 8MDMMEA + 2MPZ 5 M DMMEA + 3MPZ

4 M DMMEA + 3MPZ

5 M DMMEA + 2M MEA SM DMMEA + 2 M MAPA 3 M DEEA + 3 MPZ

-ir 3M DEEA + 1 MPZ 3 M TRIZMA * 1 MPZ

2 5 M TRIZMA + 2.5 M PZ

3 M TRIZMA + 1 M MEA 3 M TRIZMA + 1 M MAPA

— 30 % MEA - 5 M MAPA 5 M DEEA + 2M MEA

0 -1-,-1-1-1-,

0 1 2 3 4 5 6

load [mot C02 kg solutionj

Fig. 3. Screening of various systems, comparison with 5 M MEA

The comparison of most of the systems with 30 wt % MEA is shown in Figure 3. The mixtures in this figure do not form precipitates neither at the test temperatures (40 and 80 °C), nor at room temperature, with or without CO2 addition. The highest capacity per kg of solution was found for the 5 M MAPA single amine system, and was almost twice that of 5M MEA. MAPA is diamine with one primary and one secondary amino group. The solutions with the slowest kinetics and lowest capacity are the mixtures with high concentrations of the sterically hindered amine TRIS and with relatively low concentrations of promoter (MEA, PZ, MAPA). Similarly as the last TRIS (TRIS/PZ 2.5M/2.5M), tertiary amine DEEA (DEEA/PZ 3/1 and DEEA/MEA 5/2) and the 30 w % MEA, they are capturing significantly less then maximum available CO2 even at the initial stage of the measurement (the maximum CO2 flow through the solution is 0.5 Nl/min - Fig. 3.). It seems that MEA is the least effective promoter for all cases. It is also the only mono-amine of the promoters tested. Comparing MAPA and PZ, PZ seems to perform a little better, and the difference is mainly visible at close to the final loading. Examples here are the systems 5 M DMMEA + 2 M MAPA/PZ and 3 M TRIS + 1 M MAPA/PZ.

Fig. 4. Inversion of the absorption rate and loading (DMMEA/PZ)

There is a special effect observed in the DMMEA/PZ systems with constant PZ concentration (3 M) in Fig. 4. DMMEA is a tertiary amine with much slower kinetics then PZ and has relatively high concentration in all tests given in this figure. Comparing the curves with 6, 5 and 4 M of DMMEA respectively it can be seen that the highest DMMEA concentration gives the highest initial absorption rate. However, at a loading of ~2.3-2.4 mole CO2/kg solution the curves cross each other and the highest DMMEA concentration exhibits the slowest kinetics. This is close to the point where the CO2/PZ concentration ratio is nearly one. After this point it seems that the majority of the piperazine is reacted and that the reaction mechanism in changing to a slower one, namely forming bicarbonate with DMMEA. Other factors that will influence the mass transfer are the viscosity, which is generally higher in the systems with high concentration, and also the water content which becomes very low at the highest amine concentrations. In this case solvation of the formed complex may become limiting. The absorption time is the same, ~66 min. This inversion also occurs with the 2 M PZ systems (Fig. 4).

Some mixtures with high amine concentration and fast kinetics are given in Fig. 5. Three systems were tested with no or negligible water content. These were 1.5 M TRIS dissolved in 100 % MEA, the same mixture but with 0.5 M water added and finally 100 % MEA). For the TRIS/MEA systems the water addition does not seem to have any significant effect in the beginning. As TRIS (TRIZMA) is a solid, dissolving it in MEA increases the solution viscosity. This effect is even more significant after reaction with CO2. The small difference between the systems (with and without water) from a loading of about 3 moleCO2/kg solution is within the experimental uncertainty. Both these systems have lower capacity than 100% MEA and are much slower at higher CO2 concentrations. Pure MEA is more than 16 M and this solution has the highest loading capacity. Adding 1.5M of TRIS probably has very little initial effect, but only lowers the active MEA concentration.

Fig. 5. Some mixtures with high amine concentrations

PZ seems to be a better promoter then MAPA at lower amine concentrations when comparing the 5 M DMMEA + 2 M MAPA/PZ and 3 M TRIS + 1 M MAPA/PZ systems. However, at higher amine concentrations, MAPA seems to be the faster and having higher capacity. This is seen when comparing the 4 M DMMEA/4 M PZ/MAPA systems. This effect is believed to be caused by the much higher viscosity of the PZ containing systems resulting in low diffusivities. It is however, worth noting that very high PZ concentration can be dissolved in DMMEA solution with forming precipitates. The 5 M MAPA solution has a loading near twice that of 5 MEA, which is close to the theoretical value since MAPA is a diamine. 5 M MAPA has slower kinetics than 3M DMMEA/ 5M MAPA, but almost the same loading capacity.

In Fig. 6 an overall comparison of the tested systems is presented. The cyclic capacity appears only for the systems where the full absorption/desorption cycle could be carried out. The absorbed CO2 and cyclic capacity in this figure come from the more reliable BaCl2 analysis and it is usually lower than the ones done by cumulative measurements based on the gas phase by the control program. The highest cyclic capacities are found for the DMMEA/PZ solutions with ~2.7 mole CO2/kg solution, i.e. for DMMEA/PZ : 4M/3M, 5M/3M, 4M/4M). The highest ratio of cyclic to total capacity is exhibited by the DMMEA/PZ systems with lower PZ and higher DMMEA concentrations, i.e. DMMEA/PZ: 6.5M/1.5M - 88 %, 6M/2M - 85.8 %, 5M/2M - 83.5 %). The lowest cyclic capacities are found for the TRIS systems and the lowest ratio between cyclic and the total capacity is found for the 5 M MAPA system which has only 25 % of the absorbed CO2 replaceable between 40 - 80 °C.

4. Conclusion

Experimental results of initial solvent characterization for the EU project CESAR are presented. A range of single and mixed amine solutions were screened and the focus was on the effect of the loading capacity, kinetics and cyclic capacity depending on the amine concentration, amine ratio, and type of amine. Some of the solutions had much lower water content than the commonly used commercial solutions and the effect of this was discussed. Semi-quantitative screening tests were done enabling estimation of rate of absorption over the whole CO2 loading range and also of the cyclic capacity. The tested promoters were primary and secondary amines/diamines where the diamines showed better properties for post-combustion CO2 capture due to improved kinetics and capacity. TRIS is a sterically hindered amine similar to AMP, but both kinetics and capacity are much lower compared to AMP in blends with PZ at the same concentrations. At lower amine concentrations PZ was found to be a more efficient promoter then MAPA, but at lower water/higher amine contents the situation is the opposite. At higher amine concentrations the viscosity increases significantly due to CO2 capture and more so for the PZ promoted solutions. Thereby the diffusivities decrease and mass transfer may become the limiting factor for reaching the theoretical maximum loading. In some cases solutions this effect is so strong that even lower amine concentrations can result in more efficient capture. MAPA has lower cyclic capacity than PZ. Pure MEA was found to have a very large CO2 capacity combined with very fast reaction. However the cyclic capacity was limited.

Acknowledgement

This work was performed as part of the European Community Seventh Framework Program FP7/2007-2011 project CESAR, under grant agreement n° 213569.

References

[1] Hikita, H., Asai, S., Katsu, Y., Ikuno, S. Absorption of carbon dioxide into aqueous monoethanolamine solutions. AIChE J. 1979; 25, 793-800.

[2] Crooks, J.E., Donnellan, J.P. Kinetics and mechanism of the reaction between carbon dioxide and amines in aqueous solution. J. Chem. Soc., Perkin Trans. 2 1989; 331-333.

[3] Jou, F.-Y., Otto, F.D., Mather, A.E. Vapor-Liquid Equilibrium of Carbon Dioxide in Aqueous Mixtures of Monoethanolamine and Methyldiethanolamine. Ind. Eng. Chem. Res. 1994; 33, 2002-2005.

[4] Kim, I., Svendsen, H.F. Heat of Absorption of Carbon Dioxide (CO2) in Monoethanolamine (MEA) and 2-(Aminoethyl)ethanolamine (AEEA) Solutions. Ind. Eng. Chem. Res. 2007; 46, 5803-5809.

[5] Bishnoi, S., Rochelle, G.T. Absorption of carbon dioxide into aqueous piperazine: reaction kinetics, mass transfer and solubility. Chem. Eng. Sci. 2000; 55, 5531-5543.

[6] Freeman, S.A., Dugas, R., Van Wagener, D.H., Nguyen, T., Rochelle, G.T. Carbon dioxide capture with concentrated, aqueous piperazine Int. J. Greenh. Gas Con. 2010; 4, 119-124.

[7] Dash, S.K., Samanta, A.N., Bandyopadhyay, S.S. Solubility of carbon dioxide in aqueous solution of 2-amino-2-methyl-1-propanol and piperazine. Fluid Phase Equilib. 2011; 307, 166-174.

[8] Xu, G.-W., Zhang, C.-F., Qin, S.-J., Gao, W.-H., Liu, H.-B. Gas-Liquid Equilibrium in a CO2-MDEA-H2O System and the Effect of Piperazine on It. Ind. Eng. Chem. Res. 1998; 37, 1473-1477.

[9] Sharma, M.M., 1964.

[10] Chakraborty, A.K., Astarita, G., Bischoff, K.B. CO2 absorption in aqueous solutions of hindered amines. Chem. Eng. Sci. 1986; 41, 997-1003.

[11] Zioudas, A.P., Dadach, Z. Absorption rates of carbon dioxide and hydrogen sulphide in sterically hindered amines. Chem. Eng. Sci. 1986; 41, 405-408.

[12] Versteeg, G.F., Van Dijck, L.A.J., Van Swaaij, W.P.M. ON THE KINETICS BETWEEN CO2 AND ALKANOLAMINES BOTH IN AQUEOUS AND NON-AQUEOUS SOLUTIONS. AN OVERVIEW. Chem. Eng. Commun. 1996; 144, 113-158.

[13] Eide-Haugmo, I., 2011. Environmental impacts and aspects of absorbents used for CO2 capture, Phd Thesis, Chemical Engineering. NTNU, Trondheim.

[14] Ma'mun S, Jakobsen JP, Svendsen H.F, Juliussen O. Experimental and modeling study of the solubility of carbon dioxide in aqueous 30 mass % 2-((2-aminoethyl)amino)ethanol solution. Ind. Eng. Chem. Res 2006;45(8),2505-2512

[15] Aronu U.E, Svendsen HF, Hoff K.A, Juliussen O. Solvent selection for carbon dioxide absorption. Energy Procedia 2009;1, 1051-1057

[16] Bruder P, Svendsen H.F, CO2 capture into NN-dimethylethanolamine/3-amino-1-methylaminopropane (DMMEA:MAPA) solution. 2009, 5th Trondheim Conference on CO2 Capture, Transport and Storage (TCCS-5), Poster presentation, Trondheim, Norway, June 16-17