CHEMICAL ENGINEERING SERIES - CO2 capture by Sciencenetwork

CHEICAL ENGINEERING SERIES - CO2 Capture

Managing Editor: Omid Zadakbar Publisher: Science Network ISBN: 978-0-9869554-0-2

Cover Designer: Mona Atari First published in September, 2011 Printed in Canada

A free online edition of this book is available at www.sciencenetwork.ca Additional hard copies can be obtained from reprint@sciencenetwork.ca

Copyright ÂŠ 2011 Science Network All Books published by Science Network are licensed under a Creative Commons AttributionNonCommercial-NoDerivs 3.0 Unported License.

CHEICAL ENGINEERING SERIES

CO2 Capture

 Science Network Online Open Access Publisher

Table of Content Authors

Chapter 1 CO2 Removal by MEA Scrubbing: Process Simulation and Energy Saving Study Introduction

1-1-

Modeling

1-1-1- Thermodynamic and chemical equilibria

1-1-1-1- Non-ideality in the liquid phase

1-1-1-2- Values of Electrolyte-NRTL parameters

1-1-2- Kinetics and mass transfer

1-1-2-1- Film theory

1-1-2-2- Mass transfer with chemical reaction

1-2- Simulation of different configurations

1-2-1- Validation of simulation results with experimental data from a pilot plant

1-2-2- Double column scheme

1-2-3- Flashing feed scheme

1-2-4- Multipressure column scheme

Conclusion

References

Chapter 2 Overall efficiency analysis of the post-combustion CO2 capture using aqueous solution of amines Introduction

2-1- Process description

2-2- Problem formulation

2-3- Hypothesis and mathematical model

2-4- Result and Discussion

2-4-1- Bases for the development of a preliminary short-cut method for cost-effective design of amine scrubbing in post-combustion CO2 capture

Conclusions

Appendix A. Mathematical Model

Acknowledgement

References

Chapter 3 Hydrate-Based CO2 Captures from Flue and Fuel Gases Introduction

3-1- Equilibrium hydrate formation conditions

3-1-1- Research method of the equilibrium hydrate formation

3-1-2- Equilibrium conditions for CO2 /H2/TBAB system

3-2- Hydrate-based CO2 separation and capture from flue gas

3-2-1- Induction time of hydrate formation

3-2-2- Pressure drop during hydrate formation

3-2-3- CO2 concentration in hydrate slurry phase

3-2-4- Two-stage separation for CO2 Capture

3-3- Hydrate-based CO2 separation and capture from fuel gas

3-3-1- Gas uptake

3-3-2- CO2 selectivity

3-3-3- CO2 separation efficiency

3-3-4- Induction time

3-3-5- Hydrate formation rate

3-4- A Hybrid Hydrate/Membrane Process

Conclusion

Reference

Chapter 4 Capture of Carbon Dioxide by Amine Functioned Mesoporous Silicas Introduction

100

1- Character of amine functionalized mesoporous silicas

100

1-1- Low angle XRD patterns for some adsorbents

100

4-1-2 The thermal stability of some adsorbents

102

4-1-3- The textural parameters of adsorbents

104

4-2- CO2 adsorption performances of adsorbents

105

4-2-1- CO2 breakthrough curves of adsorbents

105

4-2-2- CO2 adsorption capacity of adsorbents

107

4-2-3- CO2 adsorption isotherms of adsorbents

109

4-2-4- Effect of Moisture, NO and SO2

110

4-3- Modeling of CO2 adsorption breakthrough curves

111

4-3-1- Analysis of CO2 adsorption on amine-impregnated M

112

4-3-2- Analysis of CO2 adsorption on TEPA-impregnated KT-50

112

4-4- Industrial conceptual process analyses

114

Conclusion

115

References

116

AUTHORS

Laura A. Pellegrini Dipartimento di Chimica, Materiali e Ingegneria Chimica “G. Natta”, Politecnico di Milano, Piazza Leonardo da Vinci 32, I-20133 Milano, Italy Tel: +39 02 2399 3237; Fax: +39 02 7063 8173 E-mail address: laura.pellegrini@polimi.it

Laura A. Pellegrini earned her MSc and PhD in Chemical Engineering at Politecnico di Milano in Italy. She also gained a two years experience at Foster Wheeler Italiana as process engineer. She is presently Full Professor of “Unit Operations of Chemical Plants” at the Department of Chemistry, Materials and Chemical Engineering “Giulio Natta” of Politecnico di Milano. She has a significant scientific production on hydrocracking modelling, due to a long cooperation with Eni R&M, as well as on CO2 capture and sour gas purification (current research activities with Oil & Gas Companies and Engineering Groups - Maire Tecnimont, ENI E&P). Her current research interests regard: - natural gas treatments: dehydration, C2+ recovery, etc.; - CO2 capture: ammine scrubbing, physical absorption and cryogenic distillation; - catalytic hydrocracking of heavy paraffins from GTL/BTL processes; - thermodynamic packages for highly non-ideal systems; - design and simulation of separation units using commercial simulation packages like Aspen Plus®, Aspen Hysys®, Promax®. She can produce more than one hundred publications on international journals, books and memories presented in national and international conferences.

Stefania Moioli Dipartimento di Chimica, Materiali e Ingegneria Chimica “G. Natta”, Politecnico di Milano, Piazza Leonardo da Vinci 32, I-20133 Milano, Italy Tel: +39 02 2399 4704; Fax: +39 02 7063 8173 E-mail address: stefania.moioli@mail.polimi.it

Stefania Moioli was born in 1985 in Monza (MB). She graduated magna cum laude in Chemical Engineering at Politecnico di Milano in 2009. The thesis, named Simulazione del processo di rimozione di anidride carbonica da gas esausti con soluzione acquosa di monoetanolammina was focused on the process of CO2 removal with acqueous monoethanolamine solution from exhaust gases, with the aim of studying energy saving configurations. In 2009 she won a competitive examination for a PhD position in Industrial Chemistry and Chemical Enginnering. She joined the Department of Chemistry, Materials and Chemical Engineering “Giulio Natta” of Politecnico di Milano, where in 2010 she earned a grant sponsored by Maire Tecnimont (Purificazione di gas acidi con lavaggio amminico) whose main research object is the acid gas purification by means of amine scrubbing. She has been in charge of teaching assistantship activity of courses in the field of the Unit Operations for Chemical Plants at the Third School of Engineering of Politecnico di Milano. She is studying the process of acid gas removal from exhaust gases of power plants, natural gases and refinery gases by chemical absorption and regeneration of the solvent, considering also the influence that components such as sulphur compounds, hydrocarbons and mercaptans exert on the considered process. She is co-author of various papers published on international scientific journals.

iii

Simone Gamba Dipartimento di Chimica, Materiali e Ingegneria Chimica “G. Natta”, Politecnico di Milano, Piazza Leonardo da Vinci 32, I-20133 Milano, Italy Tel.: +39 02 2399 4704; fax: +39 02 7063 8173 E-mail address: simone.gamba@mail.polimi.it

Simone Gamba was born in 1982. At present, Simone Gamba is a contracted researcher at the Department of Chemistry, Materials and Chemical Engineering "Giulio Natta" of the Politecnico di Milano where he is working on the “Acidic gases: methods for purification and thermodynamic characterization” program. He has been in charge of teaching assistantship activity of courses in the field of the Unit Operations for Chemical Plants at the Third School of Engineering of the Politecnico di Milano. He holds a cum laude PhD degree in Industrial Chemistry and Chemical Engineering (2010) from Politecnico di Milano earned defending the Thesis entitled “Kinetic Modeling and Thermodynamic Analysis of the Fischer-Tropsch Wax Hydrocracking Process”. For this Thesis, Simone Gamba received the “eni Award 2011 - Debut in Research Prize” presented by eni. During his scientific career, Simone Gamba was recognized as the Best Student in Chemical Engineering of the Politecnico di Milano in the academic year 2003-2004. He is also been awarded the “Isimbardi Prize 2011 – “Young Talents” category” presented by the Province of Milan. He is the first author and a co-author of various papers published on international scientific journals.

Patricia Mores Universidad Tecnológica Nacional – Facultad Regional Rosario. CAIMI (Centro de Aplicaciones Informáticas y Modelado en Ingeniería). Address: Zeballos 1341 - S2000BQA Rosario, Argentina. e-mail: patricia_mores@hotmail.com ; Tel.: (00 54) 341 4480102, int. 144 Patricia Mores holds degree in chemical engineering from the National Technological University of Argentina in 2008. Currently, she is a PhD student in the “Centro de Aplicaciones Informáticas y Modelado en Ingeniería” (CAIMI), a research center of the National Technological University, Rosario Regional Faculty (FRRosario), Argentina. She is a doctoral fellow of the National Technological University and National Agency of Scientific and Technological Promotion (FONCyT IP-PRH 2007) Her main research interest is process system engineering (PSE). More precisely, her PhD Thesis deals with the modeling, simulation and optimization of CO2 capture processes.

Sergio Fabián Mussati INGAR Instituto de Desarrollo y Diseño. (CONICET-UTN); Address: Avellaneda 3657 – (3000), Santa Fe, Argentina. Universidad Tecnológica Nacional – Facultad Regional Rosario. CAIMI (Centro de Aplicaciones Informáticas y Modelado en Ingeniería); Address: Zeballos 1341 - S2000BQA Rosario, Argentina. e-mail: mussati@santafe-conicet.gov.ar ; Tel. (00 54) 342 4534451 , Fax: (00 54) 342 4553439

Sergio Mussati is a Professor in the Department of Chemical Engineering at the National Technological University (Rosario Regional Faculty, Argentina). He obtained his BSc. in Chemical Engineering from the National Technological University, Villa Maria Regional Faculty (Argentina) in 1997 and his PhD. in Chemical Engineering from the National University of Litoral (Argentina) in 2003. After that, He carried out his postdoctoral studies at the Technical University of Berlin with Dr. George Tsatsaronis, from 2003 to 2005. Since 2005, he works as a research staff member at the CONICET in Argentina (Consejo Nacional de Investigaciones Científicas y Técnicas). His work place is INGAR (Instituto de Desarrollo y Diseño, CONICET-UTN). Also, he is a research member of the Research Center named CAIMI (Centro de Aplicaciones Informáticas y Modelado en Ingeniería) at the National Technological University, Rosario Regional Faculty. His main research interest is process system engineering (PSE) including flowsheet optimization and mathematical programming (discrete/continuous optimization problems). He has coauthored about 25 scientific papers. Currently, he participates in different local research projects and also supervises several PhD and MSc theses.

Nicolas José Scenna INGAR Instituto de Desarrollo y Diseño. (CONICET UTN); Address: Avellaneda 3657–(3000), Santa Fe, Argentina - Universidad Tecnológica Nacional – Facultad Regional Rosario. CAIMI (Centro de Aplicaciones Informáticas y Modelado en Ingeniería); Address: Zeballos 1341 - S2000BQA, Rosario, Argentina. e-mail: nscenna@santafe-conicet.gov.ar , Tel. (00 54) 342 4534451 , Fax: (00 54) 342 4553439

Nicolas Scenna is a Professor in the Department of Chemical Engineering at the National Technological University of Argentina, Rosario Regional Faculty (FRRosario). He obtained his BSc. in Chemical Enginnering from the National Technical University, FRRo, in 1980 and his PhD. in Chemical Enginnering from the National University of Litoral (Argentina) in 1987. He has more than 25 years of experience in teaching chemical engineering in the university. He works as a research staff member at the CONICET in Argentina (Consejo Nacional de Investigaciones Científicas y Técnicas). Also, he is head of the Research Center named CAIMI (Centro de Aplicaciones Informáticas y Modelado en Ingeniería) depending of the National Technological University, FRRosario. His main research interest is computer aided process engineering including non lineal process control and reliability. He has coauthored about 70 scientific papers. He also was editor of the book entitled "Modelling, simulation and optimization of Chemical Processes”. He participated in many national and international research projects and in several projects in collaboration with industries. Currently, he is the advisor of several PhD and MSc theses.

vii

Dr. Xiao-Sen Li Professor, Chief Scientist The Center of Gas Hydrate Research Key Laboratory of Renewable Energy and Gas Hydrate Guangzhou Institute of Energy Conversion The Chinese Academy of Sciences No 2, Nengyuan Road., Wushan, Tianhe District, Guangzhou P. R. China, 510640 Tel:+86-20-8705-7037 E-mail: lixs@ms.giec.ac.cn

Xiao-Sen Li , Ph.D., Professor, graduated from Department of Chemical Engineering, Tsinghua University for Doctoral Degree in 2000. From April, 2000 to July, 2005 he worked at University of Alberta and University of British Columbia. Since 2005, he has been awarded by “Hundred Talents Program” of the Chinese Academy of Sciences (CAS). From August, 2005 to now, he works at the Center of Gas Hydrate Research in Guangzhou Institute of Energy Conversion, the Chinese Academy of Sciences, as a chief scientist and professor, and he is a chief scientist of the Guangzhou Center of Gas Hydrate Research, CAS, and an academic leader for the Innovation Program of CAS. In recent years, He has undertaken more than 30 projects in China and on abroad, including the “National High Technology Research and Development Program of China (863 Program)”, the National Basic Research Program of China (973 Program), “Hundred Talents Program of the Chinese Academy of Sciences”, the “National Natural Science Foundation of China”, CAS Knowledge Innovation Program, the “Natural and Science Foundation of Guangdong”, and “Science and Technology Planning Project of Guangdong Province”, etc. His research areas include the equilibrium and kinetics of gas hydrates, the production technology and the utilization of natural gas hydrate and the application technology of gas hydrate. So far, he has published more than 120 academic papers, and has written 3 books. He is an Editorial Board member of Global Journal of Physical Chemistry. He has a long-term cooperation in the gas hydrate research field with the scientists from USA, Canada and South Korea and so on. He has obtained 26 patent of invention (14 has been authorized), and 1 Computer Software Copyright Registration. He has gotten the 2nd class Scientific Technology Progress Award of the China Petroleum and Chemical Industry Association (CPCIA), the 3rd class Scientific Technology Progress Award of Guangdong Province, the 3rd class Scientific Technology Progress Award of the Inner Mongolia Autonomous Region and the 2nd class Scientific Technology Progress Award of the Petroleum Chemistry Industry Department of the Inner Mongolia Autonomous Region.

viii

Chun-gang Xu Senior Engineer The Center of Gas Hydrate Research Key Laboratory of Renewable Energy and Gas Hydrate Guangzhou Institute of Energy Conversion The Chinese Academy of Sciences No 2, Nengyuan Road., Wushan, Tianhe District, Guangzhou P. R. China, 510640 Tel:+86-20-8705-9617 E-mail: xucg@ms.giec.ac.cn

Chun-gang Xu , Ph.D., Senior Engineer, graduated from Guangzhou Institute of Energy Converse, Chinese Academy Sciences. From 2008 to now, he works at the Center of Gas Hydrate Research in Guangzhou Institute of Energy Conversion, the Chinese Academy of Sciences, as a senior engineer. He has undertaken 8 projects in China, including the “National High Technology Research and Development Program of China (863 Program)”, the “National Natural Science Foundation of China”, CAS Knowledge Innovation Program, the “Natural and Science Foundation of Guangdong”, and “Science and Technology Planning Project of Guangdong Province”, etc. He mainly works on the investigations into the thermodynamics and kinetics of gas hydrate and the hydrate-based CO2 capture from flue and fuel gases.

SHI YAO Institute of Industrial Ecology and Environment, Department of Chemical and Biological Engineering, Yuquan Campus, Zhejiang University, Hangzhou, China, 310027 shiyao@zju.edu.cn Phone: +86-571-88273591; Fax: +86-571-88273693

Prof. Dr. SHI YAO received his Ph.D degree from Zhejiang University of China. He has more than three years of visiting research experience at Energy & Environment Division, Lawrence Berkeley National Lab at UC Berkeley. His research work mainly focuses on air pollution control, CCS, resource recycling technology, industrial ecology and environment system. He has published more than 100 papers and received one US patent, ten China patents.

LIU YA-MIN Department of Environment and Equipment, Fujian University of Technology, Fuzhou, China, 350108 mingjing2000@126.com

Dr. LIU YA-MIN received his PhD degree from Zhejiang University of China at 2011-03-30. His research work mainly focuses on air pollution control, CCS, resource recycling technology, industrial ecology and environment system. He has published 7 papers and received two China patents

C O

C A P T U R E

Chapter

1 CO2 Removal by MEA Scrubbing: Process Simulation and Energy Saving Study Laura A. Pellegrini, Stefania Moioli, Simone Gamba Dipartimento di Chimica, Materiali e Ingegneria Chimica “G. Natta”, Politecnico di Milano, Piazza Leonardo da Vinci 32, I-20133 Milano, Italy E-mail :

laura.pellegrini@polimi.it (L.A. Pellegrini) stefania.moioli@mail.polimi.it (S. Moioli) simone.gamba@mail.polimi.it (S. Gamba)

ndustrially CO2 has to be removed from natural gas, refinery gas and exhaust gases of power plants, with the aim of reducing emissions and, thus, mitigating greenhouse effect. CO2 capture is commonly achieved by absorption with chemical solvents, such as aqueous solutions of amines. This work focuses on CO2 capture from exhaust gases of power plants by absorption with monoethanolamine (MEA), which is the most frequently used solvent for this purpose. The study is carried out by means of a commercial simulation software (ASPEN Plus®), using an electrolytic thermodynamic package with parameters calibrated on experimental solubility data for the CO2-MEA-H2O system. The purpose of our research is reducing the high energy requirement of the stripping unit by analyzing alternative configurations.

Introduction Many gas streams commonly present in industrial plants - natural gas, syngas, exhaust gases and refinery gases - contain significant quantities of acid gases, mainly CO2 and H2S. Their presence is very undesirable due to corrosion, operational, economical and/or environmental reasons. As a matter of fact, carbon dioxide should be removed from synthesis gas which is the basis for different productions (ammonia, hydrogen, etc.). Moreover, CO2 is a powerful greenhouse gas, whose massive presence in the atmosphere is the cause of gradual global warming. In order to limit this problem and to accomplish the requirements of the Kyoto protocol, an important CO2 removal is then realized in the treatment of combustion gases at power plants. 1

C O

C A P T U R E

Industrially, absorption is probably the most used gas purification technique, involving the transfer of a substance from the gaseous phase to the liquid one through the phase boundary. The absorbed material may dissolve physically in the liquid or react chemically with it. Using aqueous amine solvents, the mass transfer is promoted by chemical reactions: acid gases can react directly or through a mechanism due to acid-base ionic species in solution. The mass transfer from the bulk of the gas phase to the bulk of the liquid phase is mainly influenced by: 1) diffusion of the component from the bulk of the gas phase to the gas-liquid interface; 2) diffusion of reagents from the gas-liquid interface to the bulk of the liquid phase; 3) simultaneous reaction between dissolved gas and liquid reactant; 4) diffusion of reaction products in the bulk of the liquid phase promoted by the concentration gradient due to chemical reaction. Molecules of alkanolamines are characterized by one hydroxyl group and one amino group. The former helps in reducing the vapor pressure and increasing water solubility, while the latter provides the necessary alkalinity in water solutions to make the acid gas absorption occur (Kohl and Riesenfeld, 1985). Indeed, the solubility of acid gas in water is highly influenced by the presence of the amine. The equilibrium solubility of an acid gas that does not react in the liquid phase is governed by the partial pressure of this gas over the liquid. In a reactive solvent, on the contrary, when an acid gas is absorbed, it is partially consumed by chemical reactions. As a consequence, the CO2 bulk concentration in the liquid phase is low and the rate of absorption of the acid gas is significantly affected by the amine. Chemical reactions enhance the driving force for mass transfer, i.e. the difference between the gas concentration in the liquid at the gas-liquid interface and the unreacted gas concentration in the bulk of the liquid phase. This work focuses on CO2 capture from exhaust gases of power plants by absorption with monoethanolamine (MEA). This amine is the most frequently used solvent for this process, due to its relatively high loading, i.e. the ratio of moles of absorbed acid gas per mole of amine. Experimental data for the system CO2-H2O-MEA have been collected over the past two decades. In order to reproduce these data, many thermodynamic models were developed, in particular the one proposed by Kent and Eisenberg (1976) and the Electrolyte-NRTL model by Chen et al. (1979). The latter can be used to reproduce experimental data for a wider range of temperatures and loadings. An absorption + regeneration plant has been considered and simulated using the RateBased model implemented in the process simulator ASPEN Plus速 (ASPEN Plus速, 2008). This model takes into account mass and heat transfer limitations and appropriate kinetic expressions for chemical reactions that do not reach chemical equilibrium should be provided as well. Experimental data for CO2 removal from exhaust gases of a power pilot plant are reported in Dugas (2006) and have been used to validate simulation results. The purification of acid gases by means of aqueous solutions of monoethanolamine is energy-intensive. The major requirement is in the regeneration section, when usually a distillation column or a reboiled stripper is used to obtain the lean solvent. The energy for 2

C O

C A P T U R E

generating the steam represents the main operating cost of the plant. In coal-fired power plants the energy consumption in the reboiler of the regeneration section can be 30% of the net power production of the plant (Oyenekan and Rochelle, 2005). This high requirement can be reduced by regenerating the solvent according to different schemes (Moioli, 2009), that are more complex than the simple distillation column. Among the most likely configurations, the double column, the flashing feed column and the multipressure column have been studied. The schemes are based on different operating pressures, which significantly influence the desorption phenomenon. By increasing pressure, the molar fraction of CO2 in vapor phase increases more than the molar fraction of water: as a consequence, to obtain the same amount of acid gas in the vapor phase a smaller quantity of water has to be vaporized so reducing energy consumption. Pressure variations can take place in the same unit (different operating sections) or in different units (different columns).

1-1-

Modeling

1-1-1- Thermodynamic and chemical equilibria A good description of vapor-liquid equilibrium is crucial for a correct design of unit operations involving mass transfer from one phase to the other (Pellegrini et al., 2010). Chemical reactions in the liquid phase should be taken into account when describing the VLE. The generation of ionic species, moreover, makes the system highly non ideal. As a matter of fact, acid gases and amines partially dissociate in the aqueous phase because they are weak electrolytes. The liquid phase is then composed of a moderately volatile solvent (water), a non volatile solvent (MEA), very volatile molecular species (acid gas) and non volatile ionic species (Pellegrini et al., 2011). Physical VLE involves only molecular species, i.e., CO2, H2O and MEA. Chemical reactions occurring in the liquid phase involve both molecular and ionic species and can be described according to the following equilibrium relations: 2

+ â&#x2021;&#x152;

2 +

(1)

, +

â&#x2021;&#x152;

(2)

+ , +

â&#x2021;&#x152;

(3)

, +

â&#x2021;&#x152;

(4)

+ + â&#x2021;&#x152;

(5)

The equilibrium constants Kj are strongly dependent on temperature T: 3

C O

C A P T U R E

! = ! +

#$ %

+ ! & + '! &

(6)

where Aj, Bj, Cj, Dj are parameters whose values can be found in literature (Edwards et al., 1978; Maurer, 1980). Values used for the simulation are reported in Table 1. Table 1. Parameters for the equilibrium constants of Eq. (1)-(5). Reaction

Source

Maurer (1980)

132.89

-13446

-22.47

Edwards et al. (1978)

231.46

-12092

-36.78

Edwards et al. (1978)

216.05

-12432

-35.48

ASPEN PlusÂŽ

-3.038

-7008.3

-0.0031

ASPEN PlusÂŽ

-0.52

-2545.5

Both literature sources (for dissociation constants of water, carbon dioxide and bicarbonate) and ASPEN PlusÂŽ default values (for dissociation constants of protonated monoethanolamine and carbammate) are taken into account. As already stated, the CO2-MEA-H2O system is highly non ideal in the liquid phase. However, the behavior of the gas phase is not so far from ideality. This is justified by the low pressure operating conditions, since the maximum pressure considered is lower than 5 bar. To describe such VLE systems, a Îł/ÎŚ approach is used. The vapor phase is represented by means of an Equation of State (EoS), in particular the SRK EoS proposed by Soave (1972). The liquid phase, on the other hand, is described by means of an activity coefficient model. For the considered system, the ElectrolyteNRTL model (Chen et al., 1979; Chen et al., 1982; Chen and Evans, 1986; Mock et al., 1986) is used. The activity coefficient is calculated from the excess Gibbs free energy: this thermodynamic function can be obtained as the sum of two contributions, one due to short range or local interactions (NRTL, Renon and Prausnitz (1968)) that exist in the immediate vicinity of a component and one due to long range (LR) ion-ion interactions.

1-1-1-1-

Non-ideality in the liquid phase

The liquid phase is characterized by high departure from ideality due to intermolecular forces. When dealing with acid gas treatment, ionic forces are present and have great influence on the interactions between species. Excess Gibbs free energy models for non-electrolytic systems take into account the local composition in the immediate neighborhood of each species. The considered local compositions are different from the average system composition and reflect the

C O

C A P T U R E

short-range interactions that provide a contribution to the excess Gibbs free energy. Applying these concepts to electrolytic systems, contribution to the Gibbs free energy consists of both short-range interactions among species and long-range interactions around a certain species. The representation of the system, then, is strongly influenced by the ability of predicting ionic forces. The ElectrolyteNRTL model (Chen et al., 1979, 1982; Chen and Evans, 1986; Mock et al., 1986) is based on the two following assumptions: 1) like-ion repulsion assumption: the local composition of cations around other cations is zero and, similarly, the local composition of anions around other anions is zero. This idea is based on the assumption that repulsive forces between ions of the same charge are strong, therefore they are extremely relevant for near species; 2) the local electroneutrality assumption: the distribution of anions and cations around a central molecule makes the net local charge null. The considered model provides an expression for the excess Gibbs free energy, taking into account molecular and ionic interactions among all species in liquid phase. It is based on the concept that the excess Gibbs free energy gE of an electrolytic system can be described as the sum of two contributions, one related to short-range local interactions (NRTL), that exist in the immediate neighborhood of a component, and another related to long-range ion-ion interactions (LR), that exist beyond the immediate vicinity of a central ionic species: ( = ()*%+ + (+*

(7)

The long-range interaction contribution is described by means of the extended form of Pitzer-Debye-H端ckel theory and of the Born equation. The short-range interaction contribution is represented by means of the local composition concept by Renon and Prausnitz (1968). According to like-ion repulsion assumption and local electroneutrality assumption, three types of cell should be considered, as shown in Figure 1.

C O

C A P T U R E

a g ac

m c

m g mc

c g cm

m Cation at center

g am

m c

g mm

m a

g ma g ca

Molecule at center

Anion at center

Figure 1. Different types of cell according to Electrolyte-NRTL theory.

One type of cell is composed of a neutral molecule m in the center, with other molecules, anions a and cations c, around. To this cell the local electroneutrality assumption applies: it is assumed that the distribution of anions and cations around the central molecule is such that the total net charge is zero. The other two types of cell are characterized by a cation or an anion in the center and solvent molecules and ions of opposite charge around. Repulsion between ions of the same charge makes the local concentration of cations (or anions) around cations (or anions) zero. The effective local mole fractions ,!- and ,-- of species j and i in the immediate neighborhood of a central species i are related by: .$/ .//

= 0 1 2!.

(8)

Where: ,! = 3! !

(9)

C O

C A P T U R E

2!- = exp (â&#x2C6;&#x2019;9!- :!- ) :!- =

(10)

<$/ <//

(11)

with ! = =! for ions (=! is the absolute value of ionic charge) and ! = 1 for molecules. 9!- is the nonrandomness factor, (!- and (-- are energies of interactions between ? â&#x2C6;&#x2019; @ and @ â&#x2C6;&#x2019; @ species. Both (!- and 9!- are symmetric: (!- = (-! and 9!- = 9-! . The effective local mole fractions ,!- and ,A- of species j and k in the neighborhood of ionic species i are related through the expression: .$/

.B/

= 0 1 2!-,A.

(12)

where: 2!-,A- = exp (â&#x2C6;&#x2019;9!-,A- :!-,A- ) :!-,A- =

(13)

<$/ <B/

(14)

and 9!-,A- the nonrandomness factor. The effective local mole fractions are related as follows: - for the central solvent cells: , C + , C + ,CC = 1

(15)

- for the central cation cells: ,C + , = 1

(16)

- for the central anion cells: ,C + , = 1 (17) The effective local mole fractions in terms of the effective local fractions can then be derived: ,-C =

./ D/E

(18)

(.F DFE .G DGE .E DEE )

where @ = H, I, J. , =

(19)

(.F .E DEG,FG ) 7

C O

C A P T U R E

, =

(20)

(.G .E DEF,GF )

The residual Gibbs energies per mole of cells of central cation ((( ) ), anion ((( ) ) and solvent (((C) ) are related to the effective local mole fractions according to the following expressions: (( ) = = (,C (C + , ( )

(21)

(( ) = = (,C (C + , ( )

(22)

((C) = , C ( C + , C ( C + ,CC (CC

(23)

The reference Gibbs energy per mole is determined for the reference state of completely dissociated liquid electrolyte and of pure solvent. Then: ( )

(24)

( )

(25)

(C)

(26)

( -K = = ( ( -K = = ( ( -K = (CC

The molar excess Gibbs energy is the sum of all variations in residual Gibbs energy per mole that result when the electrolyte and solvent in their reference state are mixed and form the existing electrolytic system: (C)

( )

= 3C 0((C) â&#x2C6;&#x2019; ( -K 1 + 3 0(( ) â&#x2C6;&#x2019; ( -K 1 + 3 0(( ) â&#x2C6;&#x2019; ( -K 1 ()*%+ The excess Gibbs energy is then: P <LMNO

= ,C , C : C + ,C , C : C + , ,C :C , + , ,C :C ,

(27)

(28)

The assumption of local electroneutrality applied to cells containing a molecule in the center can be described as: , C = , C

(29)

Combining the above expression with the one for the effective local mole fractions, the following equality applies: 2 C = 2 C

(30)

C O

C A P T U R E

and for nonrandomness factors: 9 C = 9 C = 9 ,C

(31)

9C , = 9C , = 9C,

(32)

9 ,C = 9C,

(33)

Then: : C = : C = : ,C

(34)

:C , = :C , = :C,

(35)

Considering a binary system of a single electrolyte and a single solvent, the nonrandomness factor 9 ,C and the energy parameters : ,C and :C, are the adjustable binary parameters. The expression for the excess Gibbs energy is normalized to the infinite dilution reference state: P,∗ <LMNO

P <LMNO

− 3 R ∗ − 3 R ∗

(36)

and the following asymmetric expression is obtained: P <LMNO

= ,C (, C + , C ): ,C + , ,C :C, + , ,C :C, − , S:C, +

2 C : ,C T − , S:C, + 2 C : ,C T

(37)

The equations for binary electrolyte systems can be generalized to solutions containing a greater number of electrolytes. Since the local composition concept considers only interactions between two species, no further assumptions are required for generalization. The reference state for the reference Gibbs energy of molecules is the pure component; the reference state for electrolytes is the hypothetical homogeneously mixed and completely dissociated liquid electrolyte mixture. The reference Gibbs energies per mole are: ( )

(. <FVG )

(38)

( )

(.GV <GVF )

(39)

( -K = = ∑ W ∑ FV ( -K = = ∑ W

FVV .FVV

∑GVV .GVV

(C)

( -K = (CC

(40)

C O

C A P T U R E

The excess Gibbs energy expression for aqueous multicomponent electrolyte systems is: P <LMNO

∑$ .$ D$G,FVG X$G,FVG . + ∑ , ∑ W 0∑ FV 1 ∑ ∑B .B DBE . FVV FVV B .B DBG,FVG ∑$ .$ D$F,GVF X$F,GVF .GV ∑ , ∑ W 0 1 ∑ . D ∑GVV .GVV B B BF,GVF

∑C ,C

∑$ D$E X$E

+ (41)

The expression for activity coefficient for molecular component is: )*%+ = RC ∑$ .$ D$E X$E

∑ . D . D X + ∑CW ∑EV EEV 0:CCW − B∑ B BEV BEV 1 + ∑B .B DBE D D . B BEV B B BEV ∑B .B DBG,FVG XBG,FVG .G DEG,FVG .FV ∑ ∑ W Y:C , W − ∑ Z+ ∑FVV .FVV ∑B .B DBG,FVG B .B DBG,FVG ∑ . D .F DEF,GVF X . ∑ ∑ W GV Y:C , W − B ∑B BF,GVF BF,GVF Z ∑ . ∑ . D . D GVV GVV

B BF,GVF

(42)

B BF,GVF

The activity coefficient for cation is: [

R )*%+ = ∑ W ∑

∑ ∑ W

.FV

FVV .FVV

.F DGF,GVF .GV ∑GVV .GVV ∑B .B DBF,GVF

∑B .B DBG,FVG XBG,FVG

. D + ∑C ∑ E GE 0: C ∑B .B DBG,FVG B .B DBE ∑B .B DBF,GVF XBF,GVF Y: , W − ∑ Z . D B

−

∑B .B DBE XBE ∑B .B DBE

(43)

B BF,GVF

The activity coefficient for anion is: [

R )*%+ = ∑ W ∑

∑ ∑ W

.FV

.GV

GVV .GVV .G DFG,FVG

∑FVV .FVV ∑B .B DBG,FVG

∑B .B DBF,GVF XBF,GVF

. D + ∑C ∑ E FE 0: C ∑B .B DBF,GVF B .B DBE ∑B .B DBG,FVG XBG,FVG Y: , W − ∑ Z . D B

−

∑B .B DBE XBE ∑B .B DBE

B BG,FVG

(44)

Taking into account the local electroneutrality assumption and the two-body interaction assumption of the local composition concept, the following expressions apply: 2 C =

∑F .F DGF,E

(45)

2 C =

∑G .G DGF,E

(46)

∑FV .FV ∑GV .GV

For the nonrandomness factor a molar average mixing rule is used: 10

C O

C A P T U R E

9 C =

â&#x2C6;&#x2018;F .F ]GF,E

(47)

9 C =

â&#x2C6;&#x2018;G .G ]GF,E

(48)

â&#x2C6;&#x2018;FV .FV â&#x2C6;&#x2018;GV .GV

:-C parameter can be computed from 2-C . :C , and :C , parameters can be obtained from :-C parameters: :C , =

(<EF <GF )

(<EF <EE )

:C , =

(<EG <FG )

(<EG <EE )

(<EE <GF )

= : C â&#x2C6;&#x2019; : ,C + :C,

(49)

= : C â&#x2C6;&#x2019; : ,C + :C,

(50)

(<EE <FG ) *%

The nonrandomness parameters 9 ,C , 9 , V , 9 , V and 9CCV and the energy parameters : ,C , :C, , : , V , : V, , : , V , : V , , :CCV and :CV C are the adjustable parameters for the model. : , V and : V , are related according to: : , V =

S<GV F <GF T *%

=â&#x2C6;&#x2019;

S<GF <GVF T *%

= â&#x2C6;&#x2019;: V ,

(51)

= â&#x2C6;&#x2019;: V,

(52)

and similarly : , V and : V , : : , V =

S<GFV <GF T *%

=â&#x2C6;&#x2019;

S<GF <GFV T *%

1-1-1-2- Values of Electrolyte-NRTL parameters The model is characterized by a large number of parameters, that take into account interactions between molecule and molecule, molecule and ion pair, ion pair and ion pair (see Section 1.1.1). Binary parameters vary with temperature according to the following expression: :-! = I-! +

(53)

The influence of each species is related to its concentration in the liquid phase: hence, parameters associated to low quantities species do not significantly affect the performance of the vapor-liquid equilibrium model. For this reason, according to Chen and Evans (1986) and Mock et al. (1986), the aij parameters were set to 8.0 and -4.0 for water-ion pair and ion pair-water and to 15 and -8.0 for amine-ion pair and ion pair-amine. The bij parameters are equal to 0 for each interaction, neglecting the influence of temperature for low concentrated species. 11

C O

C A P T U R E

Interactions between H2O, MEACOO-, MEAH+ and HCO3-, on the other hand, must be well described, in order to obtain a reliable model. To calibrate parameters related to these species experimental VLE data available in literature, in detail those from Jou et al. (1995) and Ma’mun et al. (2005) were used. The optimization was performed using the “Data Regression System” tool of ASPEN Plus®, forcing the obtained value to respect the maximum likelyhood method. Figure 2 shows results obtained with new parameters and compared with experimental data (Jou et al., 1995; Ma’mun et al., 2005) and with other parameters (Freguia, 2002; ASPEN Plus®). The two graphs are related to the temperatures of interest in the absorption column (60°C) and in the regeneration section (120°C). The good prediction of optimized parameters can be observed in Figure 3, where the ratio of the partial pressure to the square of the loading is shown. Since loading is smaller than unity, the Figure 3 enhances the difference between the curves. Values of aij and bij parameters for H2O, MEACOO-, MEAH+ and HCO3- used for simulation are reported in Table 2. Table 2. Values of parameters of the Electrolyte-NRTL used for the simulation. Molecule i or Ion Pair i Molecule j or Ion Pair j aij parameter H2O MEAH+ MEACOO9.86785257 MEAH+ MEACOO- H2O -4.9610383 H2O MEAH+ HCO35.34335279 MEAH+ HCO3H2O -4.0786707

bij parameter 10.7912952 0.09979927 963.302488 -11.089215

CO2 partial pressure [kPa]

ASPEN Plus parameters

Freguia (2002) parameters optimized parameters

experimental (Jou et al., 1995)

6 4 2 0 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 loading [(mol CO2) / (mol MEA)]

(a)

C O

C A P T U R E

450

CO2 partial pressure [kPa]

400 350 300

ASPEN Plus parameters Freguia (2002) parameters optimized parameters experimental (Jou et al., 1995) experimental (Ma'mun et al., 2005)

250 200 150 100 50 0 0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

loading [(mol CO2) / (mol MEA)]

(b) Figure 2. Comparison of vapor-liquid equilibrium results obtained with optimized parameters, with Freguia (2002) parameters and with ASPEN Plus® default parameters with experimental data at a) 60°C and b) 120°C.

CO2 partial pressure / (loading2) [kPa]

ASPEN Plus parameters

Freguia (2002) parameters

optimized parameters experimental (Jou et al., 1995)

30 25 20 15 10 5 0 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 loading [(mol CO2) / (mol MEA)]

Figure 3. Comparison of vapor-liquid equilibrium results obtained with optimized parameters, with Freguia (2002) parameters and with ASPEN Plus® default parameters with experimental data at 60°C in terms of ratio of CO2 partial pressure to the square of the loading.

C O

C A P T U R E

1-1-2- Kinetics and mass transfer When dealing with the absorption + regeneration system involving CO2 and amines, mass and heat transfer limitations should be taken into account. There are two main approaches to modeling: the Equilibrium and the Rate-based models. The Equilibrium model divides the column in different segments, each considered well mixed in the liquid and vapor phases. The departure from equilibrium is taken into account by introducing efficiency. The Rate-based model is a non-equilibrium model, where the rate of absorption or desorption is finite, as in a real process. This model is already implemented in ASPEN Plus速 process simulator and is used in this work. Several theories were developed to describe transfer limitations (Lewis and Whitman, 1924; King, 1966). Among these, film theory by Lewis and Whitman (1924) is used.

1-1-2-1- Film theory Film theory was proposed by Lewis and Whitman in 1924 and is the first approach to mass transfer analysis at vapor-liquid interface. Lewis and Whitman (1924) considered that in most cases the absorption rate of a solute from the gas phase to the liquid phase is limited by the diffusion phenomenon. Reactions that occur in the liquid phase are faster than diffusion rate and do not affect significantly the absorption rate. This theory assumes stationary state, i.e. time dependent variations are not taken into account. Film theory is based on the hypothesis that all the mass transfer resistances are located in two films of a finite thickness. These films are near the interface on the gas side and on the liquid side. When contact between liquid and gas phase occurs, on the gas phase of the interface there exists a layer in which motion by convection is slight compared to that in the bulk of the gas phase. A similar behavior is seen on the liquid side of the interface, where there is a layer of liquid which is free from mixing by convection. By assuming that these films are free from convection currents, any transfer of solute through these films is affected by the process of diffusion. Therefore, they represent the controlling resistances to mass transfer from one phase to another. In the bulk of the liquid or of the gas phase, on the contrary, mixing by convection is so rapid that the concentration of solute in the phase can be considered uniform everywhere. In Figure 4 physical absorption is shown.

C O

C A P T U R E

PCO2,bulk

x=0 PCO2,i

x=Î´

CCO 2,i Film

Film

CCO 2,bulk

Figure 4. Film theory

When absorption occurs without chemical reaction, film theory predicts a straight line concentration gradient both in the gas and in the liquid film, for species which transfer between phases. The rate of diffusion through the gas film is proportional to the difference between solute concentrations at the bulk and at the interface. Diffusion through the liquid film, on the contrary, is controlled by the difference between the concentration of solute in the liquid at the interface and its concentration in the bulk of the liquid. Being the surface films very thin, the actual amount of solute there contained can be considered at any time negligible compared with the amount diffusing through them. Indeed, all the solute that passes through one film must also pass through the other: the two films are two diffusional resistances in series. Sometimes one of the two films exerts a greater influence than the other one that can be negligible. In this case, the system can be considered as composed of only one film. The driving force that allows diffusion depends on the concentration difference through either of the two films. The amount of solute absorbed n per ^_ unit time t due to diffusion through the two films is . It is proportional to the ^` surface of the interface, A, and can be expressed as diffusional current density: ^_

= a< Sb< â&#x2C6;&#x2019; b- T = a+ ( - â&#x2C6;&#x2019; + )

(54)

where P is gas concentration and C is liquid concentration; g is the gas bulk, L the liquid bulk and i the interface. Interface conditions are determined considering two factors: - vapor-liquid equilibrium; - all the solute that diffuses through the gas film must diffuse through the liquid film. Then: b- = c( - )

(55)

a< Sb< â&#x2C6;&#x2019; b- T = a+ ( - â&#x2C6;&#x2019; + )

(56) 15

C O

C A P T U R E

1-1-2-2- Mass transfer with chemical reaction If chemical reactions take place, a further contribution to mass transfer should be taken into account. Besides diffusion limitations also kinetics of reactions between CO2 and MEA (Eq. (57)) and between CO2 and OH- (Eq. (58)) should be taken into account, since chemical equilibrium conditions are not attained. The reaction that are considered are then: + + +

a , +

â&#x2020;&#x2019;

(57)

a , â&#x2020;&#x2019;

(58)

For these two reactions, rate equations can be written as follows: e = a , f gf g

(59)

e = a , f gf g

(60)

Rate constants are expressed according to the Arrhenius relationship: a = h3i 0â&#x2C6;&#x2019;

Fjj *%

(61)

Values of the pre-exponential factor and of the activation energy are taken from literature (Hikita et al., 1977; Pinsent et al., 1956) and are reported in Table 3. Table 3. Values of kinetic constants. Reaction 57 58

Source Hikita et al. (1977) Pinsent et al. (1956)

A [m3/(kmol s)]

Eatt [kJ/mol]

4.32Ă&#x2014;1013

55.4603

41.2564

9.77Ă&#x2014;10

C O

C A P T U R E

1-2- Simulation of different configurations 1-2-1- Validation of simulation results with experimental data from a pilot plant The model was tested by simulating an experimental pilot plant (Dugas, 2006), consisting of an absorber and a regenerator. The scheme is shown in Figure 5. VAPOR FLASH

W PUMPOUT

LIQUID

PUREGAS PUMP

MIXER MIXEROUT LEAN

HEAT1 RICHIN

FLUEGAS ABSORBER MIXER2

STRIPPER

RICHOUT

HEAT2 PUMP2OUT

MMAKEUP

CO2OUT Q

PUMP2 LEANOUT

HEATOUT

WMAKEUP

Figure 5. Absorption + regeneration scheme in ASPEN Plus速.

The pilot plant has two columns of the same dimensions. Each column has an inside diameter of 42.7 cm and two 3.05 meter beds of packing with a collector plate and a distributor between the beds. The main characteristics of the absorber and the regenerator are reported in Table 4. Table 4. Characteristics of the absorption and regeneration sections of the experimental pilot plant (Dugas, 2006). Parameter diameter [m] packing height [m]

Absorber

Regenerator

0.427

6.10

packing type

IMTP #40

Flexipac 1Y

pressure [bar]

1.01

0.69

The absorber removes CO2 from the flue gas (18.46 kmol/h) by means of a solution of monoethanolamine 30% w/w (90.78 kmol/h). The composition of the feed entering the absorption column is reported in Table 5. Table 5. Composition of the flue gas fed to the absorber. Component Molar fraction H2O 0.0160 CO2 0.1841 N2 0.7528 O2 0.0471

C O

C A P T U R E

The lean solvent is fed to the absorption column with a loading equal to 0.281. After CO2 removal from flue gas, the loading becomes equal to 0.468. The carbon dioxide capture rate of the process is then 2.1053 kmol/h. In the absorption process considerable heat is released, due to the exothermic reaction of the acid gas in the amine solution. The temperature profile along the column, see Figure 6, presents a bulge due to the cold inlet gas absorbing heat from the rich solution. The position of the bulge depends on the value of the liquid to gas ratio. Since a detailed heat loss distribution along the column is not provided, simulation results are fairly in agreement with experimental data. 360

temperature [K]

350 340 330 320 310 300

experimental data from a pilot plant (Dugas, 2006) results of simulation

290 0.00 0.61 1.22 1.83 2.44 3.05 3.66 4.27 4.88 5.49 6.10 packing height [m] Figure 6. Comparison of simulation results and experimental data (Dugas, 2006) in terms of temperature profile along the absorption column.

The obtained duty at the reboiler (187.5 kW) is in good agreement with the experimental datum of 186.4 kW. The heat required at the reboiler of the regeneration section represents the main expense of the process. A proper description of this amount is fundamental to study energy minimization configurations.

1-2-2- Double column scheme Energy saving is a critical point in the process of gas purification. The energy demand of the carbon dioxide removal process with aqueous solutions of amine from flue gases of a power plant is equal to about 15-30 % of the produced power of the plant (Oyenekan and Rochelle, 2005). Most of the required energy is used in the reboiler of the regeneration column, where steam is needed to vaporize part of the liquid phase. A minimization of heat consumption in the regeneration section is then of upmost importance. In order to reach this goal, different 18

C O

C A P T U R E

configurations have been studied. An important factor in det determining ermining the temperature range in the column is pressure. Under pressure, the amount of carbon dioxide in vapor phase increases. The same CO2 removal from the rich amine solution can then be obtained by vaporizing a less quantity of solvent than the base ccase. ase. This cause a reduction in energy consumption, with a saving of about 20 20-40 40 % (Moioli, 2009). Moreover, the amine is regenerated without excessive loss of solvent. As for the acid gas rich stream coming out of the regeneration section, the content of w water ater is lower than in the base case, so itâ&#x20AC;&#x2122;s purer in CO2. The higher pressure is an advantage because less energy consumption is required downstream to compress gas. The solutions here presented are the double column, the flashing feed column and the multipressure ipressure column. All of them are characterized by different sections with different pressures. To compare the results from the different configurations, the pressure of the outlet carbon dioxide stream has been set to 3.30 bar. The double column scheme is composed of one absorber and two packed stripping columns, each with a reboiler, operating at different pressures (Figure 7).

Figure 7. Double column scheme.

The rich amine solution coming from the absorber is divided into two streams, with a ratio 80/20. 80/20. The bigger stream is sent to the first stripping column where a great part of CO2 is removed, but not totally. The 10% of the regenerated solvent is cooled and recycled to the absorber. The other regenerated solvent is fed at the middle of the second stripping column. The 20% of the rich solvent is fed at the top of the second column where the desired loading for the lean solvent, i.e. 0.281, is reached. The characteristics of the two columns are reported in Table 6.

C O

C A P T U R E

Table 6. Characteristics of the two columns in the regeneration section (double column). Parameter

column REG1

column REG2

diameter [m]

0.427

packing height [m]

2.04

4.06

packing type

Flexipac 1Y

pressure [bar]

2.95

1.60

The sum of the heights of the packing of the two columns is equal to the total packing height of the regenerator in the base case, so that the comparison is possible. The regenerating columns are both reboiled stripper columns, without reflux. Water coming out with CO2 in the gas phase is recovered in a partial condenser and added to the lean amine solution. In the first column pressure is higher, in order to obtain a substantial removal of carbon dioxide with low energy requirement; in the second column the pressure is lower and the desired CO2 removal is obtained. Temperature profiles are shown in Figure 8. 420

temperature [K]

410 400 390 380 370

first column

second column

360 0.00 0.61 1.22 1.83 2.44 3.05 3.66 4.27 4.88 5.49 6.10

packing height [m] Figure 8. Temperature profile of the two regeneration columns of the double column scheme.

Since CO2 is removed mainly in the first column, this equipment is characterized by a higher temperature gradient than the low pressure column. In this second column the temperature profile presents a variation only at the top, where most of desorption occurs. The partially regenerated lean solution has a loading equal to 0.283 and is fed to the absorber just below the lean solvent stream, which is characterized by a loading of 0.281. The split 4 stream has a higher CO2 content, so a higher amount of circulating amine than in the base case is required, although the increase is negligible (0.2 %). According to this scheme, 100.83 kW are required in the first column and 1.8 kW in the second column. 20

C O

C A P T U R E

The equivalent work per unit time, which takes into account all energy requirements, is (Oyenekan and Rochelle, 2006): klm = 0.75 75 r s

(%t/uvww/ ) [ t/uvww/jvtx [y) [ (% ( t/uvww/jvtx [y))

z + k = 20.27 ak

(62)

It is calculated considering that: â&#x20AC;˘ vapor condenses at 10 K above the liquid temperature in the reboiler; â&#x20AC;˘ turbine downstream temperature is equal to 313 K; â&#x20AC;˘ gas stream rich in CO2 coming out of the process is at 3.30 bar; â&#x20AC;˘ 75 % efficiency is both for turbine and compressor.

1-2-3- Flashing feed scheme The flashing feed configuration consists of one absorber and two regeneration columns. Unlike the double column scheme, howeve however, r, only one of the two columns is provided with a reboiler. In the second column a simple contact between vapor and liquid phase occurs (Oyenekan and Rochelle, 2007). The scheme is reported in Figure 9.

Figure 9. Flashing feed scheme.

The rich amine sol solvent vent entering the regeneration section is splitted into two streams, with a ratio 80/20. The split 1 stream is fed to the first column, which operates as a reboiled stripper. The regeneration allows the solvent to reach the desired loading, i.e. 0.281. The acid gas rich gaseous stream coming out of the REG1 column enters a second column. At the top of this equipment the split 2 stream is fed. Countercurrent contact between the two streams helps in removing some of the carbon dioxide from the liquid phase wi without thout energy consumption. By 21

C O

C A P T U R E

operating in this way, the reboiler has to vaporize only 80 % of the rich solvent to be regenerated. The characteristics of the two columns are reported in Table 7. Table 7. Characteristics of the two columns in the regeneration section (flashing feed). Parameter

column REG1

column REG2

diameter [m]

0.427

packing height [m]

4.06

2.04

packing type

Flexipac 1Y

pressure [bar]

1.58

Heat required at the reboiler is 124.06 kW. Considering also the compression work, equal to 2.67 kW, the equivalent work per unit time results: klm = 0.75 r s

7%t/uvww/jvtx [y; [ 7%t/uvww/jvtx [y;

z + k = 21.8 ak

(63)

As in the double column case, part of the solvent is not fully regenerated and is not at a desired loading, but is characterized by a higher CO2 content (loading = 0.398). It is then fed to the absorber at a lower point than the completely regenerated solvent. The amount of amine solution needed to obtain the same CO2 removal as the base case is higher, but not so relevant. The proper height for feeding the partially regenerated amine solution is chosen on the basis of the amount of carbon dioxide removed in the absorber. The temperature profile of the absorption column is drawn in Figure 10. 360

temperature [K]

350

340

330

320

310 0.00 0.61 1.22 1.83 2.44 3.05 3.66 4.27 4.88 5.49 6.10

packing height [m] Figure 10. Temperature profile of the absorption column of the flashing feed columns scheme.

C O

C A P T U R E

The curve is different from the one of the base case. Indeed, the partially regenerated solvent is fed to the hottest section of the column and temperature decreases. Consequently, kinetics is favored, mass transfer is enhanced and a better carbon dioxide removal is obtained. An increase of temperature occurs, then, in the section immediately below the feed.

1-2-4- Multipressure column scheme The multipressure column configuration allows a removal of CO2 from the head of the regeneration column at higher concentration and higher pressure. The scheme consists of one absorber and one reboiled stripper (Figure 11).

A R

Figure 11. Multipres Multipressure sure column scheme.

The regeneration column is composed of three sections, each operating at different pressure (Oyenekan and Rochelle, 2006). The highest zone has the highest pressure, so that the amount of water in the vapor phase is few. The liquid coming ing out of this section is fed to the lower zone: being the lower zone at a lower pressure, it is partially vaporized without any duty. As in other schemes, water is recovered from the CO2 rich gas stream by means of separation equipment. The recovered liq liquid uid is mixed with the lean solvent. The column is characterized by the same height of the base case, but it is provided with equipments for conveying and compressing the steam and for removing and introducing the liquid between the zones. The characteristi characteristics cs of the three zones of the column are reported in Table 8.

C O

C A P T U R E

Table 8. Characteristics of the regeneration column (multipressure column). Section

Pressure [bar]

High

3.30

Middle

2.30

Low

1.60

The liquid stream coming out of the middle section is divided into two streams: 10 % is fed to a proper height of the absorber (Oyenekan and Rochelle, 2007) without being further regenerated. The remainder enters the low section of the reboiled stripper. The compression work is equal to 9.72 kW, while the heat required at the reboiler amounts to 85.5 kW. The total equivalent work per unit time is: klm = 0.75 r s

7%t/uvww/jvtx [y; [ 7%t/uvww/jvtx [y;

z + k = 23.42 ak

(64)

Conclusion Energy saving is one of the major features in CO2 removal plants (Idem et al., 2006; Jassim and Rochelle, 2006; Oyenekan and Rochelle, 2006, 2007; Moioli, 2009; Pellegrini et al., 2010; Pellegrini et al., 2011). In this work, different schemes have been studied and compared on the basis of the power demand and of the total equivalent work per unit time. To compare different configurations, the same packing height of the base case has been considered: in one case only for one column (multipressure column), in the cases of double column and flashing feed column it has been split between the two regeneration columns. The multipressure column allows a lower consumption of heat at the reboiler, but it is not so advantageous in terms of electric power. Moreover, the investment costs are high, since equipment for conveying and compressing the steam and for removing and reintroducing the liquid in the column is needed. The double column configuration is the most profitable scheme: it is characterized by a quite simple structure and requires the lowest equivalent work per unit time. The present work is relevant to CO2 removal process from exhaust gases of power plants. However, it can be easily extended to the capture of carbon dioxide and other acid gases (such as hydrogen sulfide) from natural gas or refinery gas.

C O

C A P T U R E

References Aspen Plus, V7.0 (2008). Aspen Physical Property System: Physical PropertyMethods, Physical Property Models; Aspen Plus Input Language Guide, in: AspenONE V7.0 Documentation, Aspen Technology, Burlington, MA. Bates R.G., Pinching G.D. (1951). Acidic dissociation constant and related thermodynamic quantities for monoethanolammonium ion in water from 0° to 50°C, Journal of Research of the National Bureau of Standards, 46, 349-352. Chen C.C., Britt H.I., Boston J.F., Evans L.B. (1979). Extension and application of the Pitzer equation for vapor-liquid equilibrium of aqueous electrolyte systems with molecular solutes, AIChE Journal, 25, 820-831. Chen C.C., Britt H.I., Boston J.F., Evans L.B. (1982). Local composition model for excess Gibbs energy of electrolyte systems. Part I: single solvent, single completely dissociated electrolyte systems, AIChE Journal, 28, 588-596. Chen C.C., Evans L.B. (1986). A local composition model for the excess Gibbs energy of aqueous electrolyte systems, AIChE Journal, 32, 444-454. Dugas R.E. (2006). Pilot Plant Study of Carbon Dioxide Capture by Aqueous Monoethanolamine, Master of Science in Engineering Thesis, The University of Texas at Austin, Austin, TX. Edwards T.J., Maurer G., Newman J., Prausnitz J.M. (1978). Vapor-liquid equilibria in multicomponent aqueous solutions of volatile weak electrolytes, AIChE Journal, 24, 966-976. Freguia S. (2002). Modeling of CO2 Removal from Flue Gas with Monoethanolamine, Master of Science in Engineering Thesis, The University of Texas at Austin, Austin, TX. Hikita H., Asai S., Ishikawa H., Honda M. (1977). The kinetics of reactions of carbon dioxide with monoethanolamine, diethanolamine and triethanolamine by a rapid mixing method, Chemical Engineering Journal, 13, 7-12. Idem R., Wilson M., Tontiwachwuthikul P., Chakma A., Veawab A., Aroonwilas A., Gelowitz D. (2006). Pilot plant studies of the CO2 capture performance of aqueous MEA and mixed MEA/MDEA solvents at the University of Regina CO2 capture technology development plant and the Boundary Dam CO2 capture demonstration plant, Industrial and Engineering Chemistry Research, 45, 24142420. Jassim, M.S., Rochelle, G.T., 2006. Innovative absorber/stripper configurations for CO2 capture by aqueous monoethanolamine, Industrial and Engineering Chemistry Research, 45, 2465-2472. Jou F.Y., Mather A.E., Otto F.D. (1995). The solubility of CO2 in a 30 mass percent monoethanolamine solution, The Canadian Journal of Chemical Engineering, 73, 140-147. Kent R.L., Eisenberg B. (1976). Better data for amine treating, Hydrocarbon Processing, 55(2), 87-90. King C.J. (1966). Turbulent liquid phase mass transfer at free gas-liquid interface, Industrial and Engineering Chemistry Fundamentals, 5, 1-8. 25

C O

C A P T U R E

Kohl A.L., Riesenfeld F.C. (1985). Gas Purification, fourth ed, Gulf Publishing Company, Houston. Lewis W.K., Whitman W.G. (1924). Principles of gas absorption, Industrial and Engineering Chemistry, 16, 1215-1220. Maâ&#x20AC;&#x2122;mun S., Nilsen R., Svendsen H.F., Juliussen O. (2005). Solubility of carbon dioxide in 30 mass % monoethanolamine and 50 mass % methyldiethanolamine solutions, Journal of Chemical and Engineering Data, 50, 630-634. Maurer G. (1980). On the solubility of volatile weak electrolytes in aqueous solutions. American Chemical Society Symposium Series, 133, 139-172. Mock B., Evans L.B., Chen C.C. (1986). Thermodynamic representation of phase equilibria of mixed-solvent electrolyte systems, AIChE Journal, 32, 1655-1664. Moioli S. (2009). Simulazione del Processo di Rimozione di Anidride Carbonica da Gas Esausti con Soluzione Acquosa di Monoetanolammina, Master of Science in Chemical Engineering Thesis, Politecnico di Milano, Milano, Italy. Oyenekan B.A., Rochelle G.T. (2005). Performance of innovative stripper options for CO2 capture, 2005 AIChE Annual Meeting Conference Proceedings, AIChE Annual Meeting and Fall Showcase, Cincinnati, OH, Oct. 30-Nov. 4, 2005. Oyenekan B.A., Rochelle G.T. (2006). Energy performance of stripper configurations for CO2 capture by aqueous amines, Industrial and Engineering Chemistry Research, 45, 2457-2464. Oyenekan B.A., Rochelle G.T. (2007). Alternative stripper configurations for CO2 capture by aqueous amines, AIChE Journal, 53, 3144-3154. Pellegrini L.A., Gamba S., Moioli S. (2010). Using an adaptive parameter method for process simulation of nonideal systems, Industrial and Engineering Chemistry Research, 49, 4923-4932. Pellegrini L.A., Moioli S., Gamba S. (2011). Energy saving in a CO2 capture plant by MEA scrubbing, Chemical Engineering Research and Design, 89, 1676-1683. Pinsent B.R.W., Pearson L., Roughton F.J.W. (1956). The kinetics of combination of carbon dioxide with hydroxide ions. Transanctions of the Faraday Society, 52, 1512-1520. Renon H., Prausnitz J.M. (1968). Local compositions in thermodynamic excess functions for liquid mixtures, AIChE Journal, 14, 135-144. Soave G. (1972). Equilibrium constants from a modified Redlich-Kwong equation of state. Chemical Engineering Science, 27, 1197-1203.

C O

C A P T U R E

C O

C A P T U R E

Chapter

2 Overall efficiency analysis of the postcombustion CO2 capture using aqueous solution of amines Patricia Mores a , Sergio Mussati a , b, Nicolas Scenna a , b

a UTN – FRRo. - Zeballos 1341 - S2000BQA - (Rosario), Argentina. Tel. (00 54) 341 4480102 b INGAR/CONICET –Instituto de Desarrollo y DiseñoAvellaneda 3657 – (3000), Santa Fe-Argentina Tel. (00 54) 342 4534451 ; Fax: (00 54) 342 4553439 E-mail :

patricia_mores@hotmail.com (P. Mores) mussati@santafe-conicet.gov.ar (S. Mussati) nscenna@santafe-conicet.gov.ar (N. Scenna)

his chapter deals with the optimization of the post-combustion CO2 capture using a monoethanolamine (MEA) solution. The overall efficiency of the whole capture process is optimized using a detailed NLP mathematical model. The proposed model takes into account the effect of equilibrium reactions on the mass transfer, thermodynamic non-idealities and the hydraulics of the random packing. Specific correlations to compute enthalpies, fugacity coefficients, surface tension and effective interfacial area for mass transfer are also considered. The following are some of the main process variables which are optimized simultaneously: temperature, composition and flow-rates profiles of liquid and flue-gas streams along the column, heat duties in the reboiler, condenser and heat exchangers, electricity consumed by pumps and compressors. The main goal is to obtain a set of optimal “thermodynamic” solutions useful to establish valuable relationships between the most important process variables in order to find optimal economic solutions. Thus, the optimization results presented in this chapter can be then used to obtain basic relationships between thermodynamic and 28

C O

C A P T U R E

economic solutions. Certainly, the set of thermodynamic solutions may be efficiently used in a first phase to get systematically initial feasible solutions and bounds to guarantee the convergence of economic optimization problems.

Introduction The main greenhouse gases considered to be the main responsible of the global warming are water vapor, carbon dioxide, methane and nitrous oxide. Several studies indicate that the combustion of fossil fuels is the largest source of CO2 emissions to the atmosphere which have been drastically increased during the last 50 years. In addition for the next decades, experts predict that the level of greenhouse emissions will increase as the population grows. Certainly the growing of worldÂ´s population leads to higher energy demands which can only be satisfied using fossil fuel. Despite that renewable energies (wind, solar, biomass) are attractive from environmental criterion, they could not replace the use of fossil fuel in order to satisfy the predicted energy demands. The coal is the most plentiful fossil fuel and it is found over much of the world. Consequently, coal is the dominant fuel in the electricity-generation sector. Power plants that process coal, natural gas, and oil to generate electricity account for about 40% of the world's total carbon dioxide emissions. However, there are not commercial technologies available to treat greenhouse gases emissions coming from fossil fuel power plants. Certainly, the Carbon Capture and Storage (CCS) is a relatively new technology, and its implementation in a large-scale has not yet been done. Despite that each component of CCS systems is commercially available, there is not enough experience at industrial scales with the configuration of all of these components into fully integrated CCS systems. At the moment, the following are the main technologies to capture CO2 from energy conversion processes: pre-combustion; post-combustion and oxyfuel combustion. In the pre-combustion process, the fossil fuel is firstly processed in a gasifier instead of a conventional combustor, to convert the solid coal into H2 and CO2 (syngas). Then, the CO2 is easily separated and captured and the hydrogen-rich gas is burned in a â&#x20AC;&#x153;combined cycleâ&#x20AC;? gas turbine to produce electricity at low greenhouse gases emissions. In fact, the combustion of hydrogen does not lead to any production of CO2. This capture technology is particularly suitable for new fossil fuel power plants because it requires strong integration between the gasifier and the remaining pieces of equipment. In contrast to the pre-combustion technology, in the post-combustion CO2 capture the CO2 is removed from a gas mixture composed mainly by N2 and H2O after burning the fossil fuel in conventional combustors. The following are the main technologies considered for post-combustion capture: 1) scrubbing with aqueous amine solutions (MEA, MDEA, DEA); 2) combined amine-membrane techniques; 3) molecular sieves 29

C O

C A P T U R E

and 4) pressure and temperature swing adsorption (PSA and TSA) using zeolites. The possibility to retrofit existing fossil fuel plants is one of the greatest advantages of this technology because it can be placed at the end of the pipe treatment without requiring significant modifications on the process arrangement. However, most of the mentioned techniques involve high operating costs. For example, one of the main drawbacks of scrubbing using aqueous amine solutions are the high amount of energy required to regenerate the amine solvent, corrosion and excessive solvent degradation. Finally, in the oxyfuel combustion the fossil fuel is burnt in presence of concentrated oxygen and recycled exhaust gases, instead of regular air. Thus, the flue gas contains mainly CO2 and H2O and the CO2 can be easily separated and captured from the water vapor in a direct contact cooler. Conventional boiler technology can be used for the combustion and has the potential to be retrofitted to existing fossil fuel plants. However, the high investment and operating costs to obtain pure oxygen (air separation unit) are the main drawbacks this technology. At the moment, there exist several research projects concerned with the mentioned capture technologies in order to increase the CO2 absorption efficiency and/or to minimize the total cost of the CO2 capture process. Prestigious international universities, research institutes and industries are working in different research areas: 1) Selection and development of amine solvents. The development of new absorbents (primary, secondary, tertiary, sterically-hindered alkanolamines including their mixtures) with low environmental impact and high absorption efficiency is receiving considerable efforts at the present time. This area also involves research activities on degradation, corrosion, chemical kinetics and reaction mechanisms (Lepaumier et al., 2009(a); Lepaumier et al., 2009(b); Lepaumier et al., 2009 (c); Bello and Idem, 2006; Dubois et al., 2010; Lepaumier et al., 2010; Davis and Rochelle, 2009; Sexton and Rochelle, 2009; Dawodu et al., 2009; Supap et al., 2009; Edali et al., 2009; Kittel et al., 2009; Puxty et al., 2009; Manjula et al., 2009). 2) Experimental works at pilot plant scales. Many demonstration projects, focused on comprehensive monitoring and verification of the mentioned capture technologies have recently started. Certainly, the installation of pilot plants is essential to develop cost-effective process flow-sheets of CO2 capture for use in middle to large scale plants. 3) Modeling, simulation and optimization of CO2 capture processes. The application of the mathematical programming techniques as well as advanced process modeling tools for simulation and optimization of chemical processes are invaluable in gaining insights about the complex processes in order to determine feasible and optimal process designs. Certainly, process alternatives and modifications can be easily evaluated in short times without the need for pilot-scale experimentations which are generally costly.

C O

C A P T U R E

Despite the valuable results achieved in the mentioned research lines, more intensive R&D efforts are still required on all research areas for improved performance, lower costs and reliability of operation on large and commercial scales. The current chapter deals precisely with the modeling and optimization of postcombustion CO2 capture via mathematical programming. Specifically in postcombustion CO2 absorption process using aqueous amine solution, a great number of mathematical models have been recently developed according to different purposes: steady state models [Bhattacharyya et al., 2011; Khan et al., 2010; Oyenekan et al., 2009] and dynamic models [(Kvamsdal et al., 2009; Chalmers et al., 2009; Ziaii et al., 2009; Greer et al., 2010)]. Most of them deal with the simulation and only few articles focus on the simultaneous optimization of the whole process. This chapter presents a nonlinear mathematical programming model (NLP) in order to study the overall efficiency of the whole post-combustion CO2 capture using MEA amine. Precisely, a family of optimal solutions obtained from a thermodynamic point of view (overall efficiency) is discussed. As will be explained later, an objective function defined as the ratio between the amount of CO2 captured and the total area required for the whole process is proposed to be maximized. The obtained solutions are very useful for many purposes: - They provide the needed information to predict optimal operating conditions when some processÂ´s specifications (flue-gas flow-rate, % CO2 recovery, among others) vary. - They can be directly related to the economic solutions. In fact, they provide the basis for the development of a short-cut method to find optimal designs when the total cost of the process is considered. The chapter is outlined as follows. Section 1 briefly describes the post-combustion process. Section 2 introduces the problem formulation. Section 3 summarizes the assumptions and the mathematical model. Section 4 presents applications of the developed NLP model and the analysis of results of the obtained solutions. Finally, the conclusions and future work are outlined.

2-1- Process description Figure 1 shows a schematic flow-sheet of a conventional post-combustion CO2 capture process considered for the optimization. As shown in Fig. 1, the acid gas containing CO2 enters the absorber [ABS] at the bottom and flows upwards. The sweet gas leaves the top of the absorber tower. The lean amine, without acid gases (coming from the regenerator unit [REG]), enters the 31

C O

C A P T U R E

top of the absorber. The rich amine, carrying the acid gases, leaves the bottom of the absorber, passes through the lean/rich amine heat exchanger [EXCH [EXCH-1], 1], and then flows through the filter to remove solid impurities. The rich solution then flows downwards through this stripper column [REG]. Acid gases are removed from the process and condensed steam returns to the regenerator as reflux. Purified amine leaves the regenerator and goes through the amine amine-amine amine heat exchanger [EXCH [EXCH-1] and solution coo cooler ler [COOLER 2] before returning to the absorber.

Figure 1. Post combustion CO2 capture process

Flue gases volume to be treated in common thermoelectric generation plants becomes extremely high in capture process columns. It is common to use one, two, thr three ee or more capture trains in real plants depending on the plant`s capacity (Chapel et al., 1999; Iijima, 2002; Desideri and Paolucci, 1999). On the other hand, the volume of recovered CO2 strongly influences on the selection of compressors (number of units and capacities). The ratio between total moles CO2 /total moles amine referred to liquid phase (hereafter named CO2 loading) is one important parameter widely used to analyze CO2 capture with amine. This parameter strongly depends on the type of amine. Primary and secondary amines react in a 2/1 mol amine/mol CO2 proportion, limiting loading to 0.5 mol (CO2) to mol of amine, while 1 mol of tertiary amines reacts with 1 mol of CO2, considerably improving loading limitation. Finally, the reaction of tertiary ry amines leads to the formation of bicarbonate and protonated amine instead of carbamate. Despite that the reactions between tertiary amines and CO2 do not produce carbamate ions, they indirectly contribute to bicarbonate formation and amines protonation. These reactions are relatively slower than carbamate formation from 32

C O

C A P T U R E

primary and secondary amines, so high residence times are necessary to achieve the same performance. On the other hand, the reboiler heat duty of the regenerator unit is strongly influenced by amine type and composition. From the process configuration point of view, it is important to mention that heat integration between both columns (absorption and regeneration) as well as other structural changes such as bifurcations of the lean amine stream to feed different points in both columns are being discussed in the literature [Ali 2004; Jassim and Rochelle (2006)] but they are still in a conceptual phase and are difficult to implement at an industrial scale. According to the above comments, it is easy to observe the strong relationship among all process variables; and they should be necessarily considered to optimize the entire process.

2-2- Problem formulation The optimization problem proposed to analyze the overall efficiency of the postcombustion CO2 capture illustrated in Fig. 1 can be stated as follow. Given the heat duty supplied to the reboiler, the problem consists in finding the optimal operating conditions and distributions of the heat transfer areas between coolers, heaters, pumps and compressors in order to obtain the maximum efficiency of the whole process. More precisely, the objective function (efficiency) is defined as the ratio between the total mole of CO2 captured and the total area required for the main process-units. Certainly, the total area involves the heat transfer area of heat exchangers (coolers and reboiler) and the superficial area related to the absorber and regenerator columns. The objective here is to maximize the overall efficiency of the system for different heat loads supplied to the reboiler. Thus, the idea is based on the efficiency optimization of the process subject to a finite energy constraint. Temperature, composition and flow-rate profiles of aqueous solution and flue gas streams along the columns as well as the heat transferred in the reboiler, condenser and heat exchanger form part of the optimization variables. In addition, the dimensions of each pieces of equipment (absorber, regenerator, heaters and pumps) are also optimized simultaneously.

C O

C A P T U R E

2-3- Hypothesis and mathematical model The following are the main hypothesis assumed for the absorber and regenerator units. The packing is divided into N stages (known value). The liquid and gas phases are considered to be well-mixed and consequently there is no concentration and temperature gradients in single liquid and gas phases and point efficiency is equivalent to Murphree efficiency. Also, it is assumed that liquid and gas phases are in thermal equilibrium. Non-ideal behavior in the gas phase is assumed and the fugacity coefficients are computed by using Peng-Robinson equations of state (multi-component mixture). In contrast to this, ideal behavior is assumed for the liquid phase. The effect of the chemical reaction on the CO2 transfer is considered by an enhancement factor. Dimensions and pressure drop along the units are considered. It is assumed that the following chemical reactions take place at the liquid and vapor interface:

2 H 2O ↔ H 3O + + OH −

(1)

2 H 2O + CO2 ↔ H 3O + + HCO3−

( 2)

H 2O + HCO3− ↔ H 3O + + CO32−

(3)

H 2O + MEAH + ↔ H 3O + + MEA

( 4)

MEACOO − + H 2 O ↔ MEA + HCO 3−

( 5)

MEA + CO2 + H 2O ↔ MEACOO − + H 3O +

( 6)

CO2 + OH − ↔ HCO3−

(7 )

Heat exchangers

The pressure drop along the heat exchangers is neglected. Special equations to avoid temperature crosses (mathematical restrictions in the optimization problem) are taken into account. Additionally, the hot-side temperature difference of the EXCH-1 is an optimization variable. It is assumed a constant flow rate and fluid thermal properties along the exchanger. Therefore the heat transfer area is computed by using overall heat transfer area and LMTD.

C O

C A P T U R E

Compressors and pumps

Isentropic efficiency of pump, blower and compressors are assumed to be 0.75. The number of compressor stages is assumed to be 4 and the final pressure is fixed into 8.6 MPa. Finally, it is assumed that the flue gas to be treated comes from a combined cycle power plant. Therefore a low CO2 concentration in the flue gases is assumed. In addition, an aqueous MEA amine solution is used as the solvent (30 wt. %).

Mathematical Model

The optimization problem defined in section 2 can be written formally as a non linearprogramming model (NLP) according to the following structure:  Kg of CO 2 captured(x)  Maximize   Total Area( x )  

subject to: Reboiler heat duty(x) – E0 ≤ 0

(a)

Gi(x) = 0 Hj(x) ≤ 0 x∈X

X is an open nonempty set

Gi(x)

equality constraint “i”

Hj(x)

inequality constraint “j”

where Gi(x) refers to the mass and energy balances and design equations [eq. (A1) to (A34) listed in Appendix A]. Hj(x) are inequality constraints used, for example, to avoid over-crosses of temperatures. Some of the inequality constraints are listed in Appendix A [eq. (A35) to (A40)]. Total Area(x) involves total heat transfer area (cooling and heating) and superficial area of absorber and stripper which depend on the dimensions of columns (height and diameter). Thus, according to the above assumptions a complete mathematical model for the post-combustion CO2 capture using aqueous solution of amines is developed which is fully described in Appendix A.

C O

C A P T U R E

Basically, the mathematical model also includes mass and energy balances corresponding to the absorber, regenerator, condensers, heaters, reboiler and compressors. Also, it considers constraints to compute the Henry's law constant (H), enthalpies, reaction heats ( â&#x2C6;&#x2020;H R ), viscosity, vapor pressures, fugacity coefficients and surface tension among others. The resulting model involves approximately 3500 variables and constraints. The model was implemented in General Algebraic Modeling System GAMS (Brooke et al., 1996). The generalized reduced gradient algorithm CONOPT 2.041 was here used as NLP solver (Drud, 1992). It should be noticed that global optimal solutions can not be guaranteed due to some non-convex constraints involved in some of the mathematical model (bilinear terms, logarithms etc.).

2-4- Result and Discussion In this section the optimal solutions obtained by the proposed optimization mathematical model are presented. The packing and flue-gas specifications are listed in Table 1 and 2. Table 1. Specifications for absorption and regeneration units Packing specifications

Type of packed

Ceramic Intalox Saddles 2

Specific area (m /m )

118

Nominal packing size (m)

0.05

Packing factor (m2/m3)

121.4

Void fraction

0.78

Reboiler pressure (kPa)

202.6

Inlet Temperature of the lean solvent into the 313.15 absorber (K) Clean gas exit pressure (kPa)

101.3

Table 2. Flue-gas specifications Operating conditions

Flue gas flow rate (mol/s) 10000 Flue gas temperature (K)

323.15

CO2 molar fraction (%)

4.22

H2O molar fraction (%)

8.45

N2 molar fraction (%)

75.67

O2 molar fraction (%)

11.66

C O

C A P T U R E

Figure 2 to 11 show the optimal values of the main process variables for different values of specific heat duties in the reboiler. In order to obtain the mentioned solutions, successive optimization problems have been solved using the following strategy. The solution obtained for the first optimization problem [E0(k=1)] was used as initialization to solve the next optimization problem [E0 (k=2)], similarly to the homotopy perturbation method. It should be mentioned, however, that it was necessary to try with several initial points for the convergence of the first optimization problem [E0(k=1)].

Kg of CO2 captured / Total Area [Kg of CO2 / m ]

It should be mentioned that the solutions illustrated in the figures were obtained systematically by varying the model parameter E0 (specific reboiler heat duty, Qreb).

2,7 2,6 2,5 2,4 2,3 2,2 2,1 4,8 5,0 5,2 5,4 5,6 5,8 6,0 Specific reboiler heat duty [MJ/Kg of CO2 captured] Figure 2. Optimal values of objective function vs Qreb

56000 Amount of CO2 captured [Kg/hr]

26000

54000

Total Area [m ]

24000

52000

22000

50000 20000

48000 18000 46000 4,8 5,0 5,2 5,4 5,6 5,8 6,0 Specific reboiler heat duty [MJ/Kg CO2 captured] Figure 3. Amount of CO2 captured and Total Area vs Qreb

C O

C A P T U R E

1750 1700

1650 20000

1600 1550

18000

1500 16000 1450 1400 4,8 5,0 5,2 5,4 5,6 5,8 6,0 Specific reboiler heat duty [MJ / Kg of CO2 captured] Figure 4. Total heat transfer area and packing volume vs. Qreb

Total heat trasnfer area / Total Area

14000

0,890

0,150

0,885

0,145

0,880

0,140

0,875

0,135

0,870

0,130

0,865

0,125

0,860

0,120

0,855

0,115 2

0,850

Total Area [m ]

0,110 18000 20000 22000 24000 26000 Figure 5. Optimal distributions of heat transfer area and superfical area vs. total area

Superficial area of absorber and stripper / Total Area

Total heat transfer area [m ]

22000

Total packing volume [m ] (absorber and stripper)

Figure 2 to 3 clearly show how the objective function defined as the ratio between the amount of CO2 captured and the total area decreases as the increasing of the specific reboiler heat load (Qreb). This is because the total area increases faster than the amount of CO2 captured as shown in Fig 3.

Figure 4 shows the increase of the total heat transfer area and the total volume packing (absorber and stripper) with the increase of Qreb. According to the reported values, the total heat transfer area increases approx. 37 % from 4.8 to 6.0 MJ/Kg CO2 captured, while the total packing volume increases approx. 13 %.

C O

C A P T U R E

0,588

COND-1

0,585

REB-1

0,175 0,150

0,582

0,125 EXCH-1

0,579 0,576

0,100 0,075

COOLER-2

0,573 Inter-coolers (4) in Compression stage

0,570 14000

0,050 0,025

Heat trasnfer area / Heat total trasnfer area

Figure 6 illustrates how the total area is distributed between the heat transfer area and superficial area related to the absorber and stripper.

16000

18000 20000 22000 2 Total heat transfer area [m ] Figure 6. Optimal distribution of the total heat transfer area

19500

CO2 loading to absorber [ÎąLean]

Solvent flow rate to absorber [mol/s]

0,181 18000 0,180 16500

0,179 15000

0,178

13500

0,177

12000

0,176

10500 4,8

5,0 5,2 5,4 5,6 5,8 6,0 Specific reboiler duty (MJ/ Kg of CO2 captured) Figure 7. Optimal of the flow-rate and CO2 loading to absorber vs. Qreb

The distribution of the total heat transfer area between coolers and heaters is illustrated in Fig. 6. As shown, it can be clearly seen that more than the half of the total heat transfer area is used by EXCH-1 (57.00 %). The 31 % of the total heat transfer area is required by the condenser and the reboiler of the stripper (COND-1 and REB-1). The heat transfer area involved by four inter-coolers used in the compression stage only requires 2.5 % of the total area. According to Fig. 7, the solvent flow rate to absorber and CO2 loading factor increase with the increasing of the specific reboiler duty. An increase of CO2 loading factor is 39

C O

C A P T U R E

attractive because it leads to increase the efficiency of the CO2 absorption in the absorber column. In order to illustrate the capability of the developed model, the following figures illustrate different profiles of the main process variables inside the columns. More precisely, the optimal CO2 loading and temperature profiles for liquid and gas phases along the absorber and regenerator obtained for two values of specific reboiler duties are illustrated from Fig. (8) to (11). That is, these profiles correspond to only two points illustrated previously in Fig. 2 to 9. The optimal height of absorber and stripper varies with the specific reboiler heat duty slightly. Certainly, the optimal height of absorber is 16.7 m and 17.0 m for 5.3 and 6.0 MJ/Kg CO2 captured respectively. For stripper, the optimal heights corresponding to 5.3 and 6.0 MJ/Kg CO2 captured are, respectively, 6.0 and 7.0 m.

0,40

395

Liquid temperature [K]

CO2 loading [mol CO2/ mol MEA]

390

0,35

385

0,30

380 T (5.3 MJ/ Kg CO2) T (6.0 MJ/ Kg CO2)

375

CO2 loading (5.3 MJ/ Kg CO2) CO2 loading (6.0 MJ/ Kg CO2)

0,25

0,20 2 4 6 8 Stripper height from the bottom [m] Figure 9. Profiles of liquid temperature and CO2 loading along the stripper

370

0,40 CO2 loading [mol CO2 / mol MEA]

330 Liquid temperature [K]

0,35 325

0,30

320

0,25 T (5.3 MJ/ Kg CO2) T (6.0 MJ/ Kg CO2)

315

0,20

Îą (5.3 MJ/ Kg CO2) Îą (6.0 MJ/ Kg CO2)

0,15 4 6 8 10 12 14 16 18 Absorber height from the bottom [m] Figure 8. Profiles of liquid temperature and CO2 loading factor along the absorber -2

C O

C A P T U R E

According to the results shown in Fig. 8 and 9, it can be concluded that the specific reboiler heat duty has slight influence on the liquid temperature and CO2 loading factor behaviors at the absorber and stripper. However, despite that the profiles of the gas and liquid flow-rate follow similar trends in both columns for different specific reboiler heat duties, they essentially differ on the numerical values. For example, the liquid flow rate in the absorber for 5.3 MJ/Kg CO2 captured is 13.51 % lower than that required for 6.0 MJ/Kg CO2 captured. In the same way, the liquid flow rate in the stripper for 5.3 MJ/Kg CO2 captured is approx. 21.02 % lower than that predicted for 6.0 MJ/Kg CO2 captured. Finally, Fig. 10 and 11 illustrate the corresponding profiles of liquid and gas flow-rates in both columns.

Vapour phase (6.0 MJ/ Kg CO2 captured)

1800

Liquid phase (6.0 MJ/ Kg CO2 captured)

Liquid phase (5.3 MJ/ Kg CO2 captured)

20000 19000

1650 18000

1500 1350

17000

1200 16000 1050

Liquid flow-rate [mol/s]

Vapour flow-rate [mol/s]

Vapour phase (5.3 MJ/ Kg of CO2 captured)

1950

15000

900 0

2 4 6 8 Stripper height from the bottom [m] Figure 11. Profiles of liquid and vapour flow-rate along the stripper

19500 10500 10400

18500

10300

18000 Vapour; 5.3 MJ/Kg of CO2

10200

Vapour; 6.0 MJ/ Kg of CO2 Liquid; 5.3 MJ/ Kg of CO2

10100

Liquid; 6.0 MJ/ Kg of CO2

10000

17500 17000 16500 16000

-2

4 6 8 10 12 14 16 18 Absorber height from the bottom [m] Figure 10. Profiles of liquid and vapour streams along the absorber

Liquid flow-rate [mol/s]

Vapour flow-rate [mol/s]

19000

C O

C A P T U R E

2-4-1- Bases for the development of a preliminary short-cut method for costeffective design of amine scrubbing in post-combustion CO2 capture

Realistic models of a wide variety of chemical and industrial processes often lead to large and non-convex programming models (NLP) which are difficult to solve. For this type of models, good starting points and lower-upper bounds are essential to find feasible solutions in order to guarantee the model`s convergence towards finding optimal solutions. In this sense, it is clear that the development of short-cut methods to determine feasible starting points (or least cuasi-feasible solutions) in a few iterations is valuable to help to solve rigorous models. As follows, the development of a shortcut method to accurately determine feasible (or least cuasi-feasible) starting points to solve economic optimization problems successfully is highlighted. The final design of any process is often obtained by considering economic aspects. For example, the total cost of the CO2 capture process with amine can be computed by the objective function proposed for problem P1 (Table 3). As shown, P1 consists of the minimization of total annual cost including the basic items to compute the main operating costs and investment. The proposed objective function for P1 is only used for a preliminary designs; a more complete objective function should be used for final designs. As shown in Table 3, the problem P2 is similar to the optimization problem solved in the previous section. In fact, it only differs from the previous problem on the constraint (a) which is now extended to include the total energy supplied to the process (cooling utilities and electricity consumed by pumps and compressors). In addition, it is important to mention that the criterion proposed in P2 can be easily extended for the different CO2 capture technologies involving the more important process variables in an appropriate manner.

Table 3. Proposed optimization problems Problem P1 (objective function: total cost) Min.[CAK1 Total_Area(x) + CEK2 Total_Energy(x*) + CRK3 CO2_Recovery(x)] subject to: Gi(x) = 0 Hj(x) ≤ 0 x∈X X is an open nonempty set Gi(x) equality constraint “i” Hj(x) inequality constraint “j”

Problem P2 (objective function: efficiency) Max. [CO2 Recovery(x)/Total_Area(x)] subject to: Total_Energy(x) - E0 ≤ 0 (a) Gi(x) = 0 (same constraints of P1) Hj(x) ≤ 0 (same constraints of P1) x∈X Gi(x) Hj(x)

As introduced, CO2_Recovery(x), Total_Energy(x) and Total_Area(x) refer, respectively, to the CO2 Recovery, total heating-cooling utilities as well as total electricity requirements and total area involved by coolers, heaters, absorber and regenerator. CA and CE are, respectively, the unit cost of area and energy. CR refers to the unit cost of the CO2 recovery. K1 to K3 are conversion factors. E0 is a model parameter while x represents the model variables. 42

C O

C A P T U R E

As stated earlier, for each values of E0, an optimal distribution of Total_Energy(x) between heaters and coolers leading to the maximum value of [CO2_Recovery(x)/Total_Area(x)] is obtained by solving problem P2. Thus, an optimal solution family characterized by their high efficiencies can be obtained. A relevant and interesting point is that these optimal solutions can be then related to economic solutions by applying the Karush Kuhn Tucker –KKT– optimal conditions and the corresponding Lagrangian dual problems for P1 and P2. Certainly, it is possible to group the more representative process variables and relate them with the unit costs from the application of optimal conditions for P1 and P2. In this sense, it is possible to compute an optimal cost set for each value of E0 which are directly related with the main process variables. Then once the unit costs are given, initial values and bounds can be established in a systematic and easy way to solve the detailed optimization model. Thus, the greatest advantages of these relationships are: a) they provide feasible solutions to guarantee the convergence of the detailed optimization models and b) they provide valuable insights to predict the direction to which economic solutions will move when the specific costs varies. Figure 12 illustrates a basic framework proposed to solve detailed optimization problems involving an objective function considering both investment and operating costs. This methodology is under development and is tested for different economic optimization problems.

Input data:

Specific costs: CA ; CE ; CR ; PR

To compute the “efficiency costs” using the introduced input data. This will give initial values to solve P1. Pre-processing phase to provide initial values (feasible solutions) to be used for the rigorous optimization To solve problem P1 (simplified objective function) using starting points and bounds computed in previous step.

To solve problem P1 (detailed and rigorous economic model) using starting points and bounds computed in previous step.

Economic Optimal Design (detailed) Figure 12. Basic steps of an alternative solution procedure

C O

C A P T U R E

According to Figure 12, once the unit costs are known, initial values of the main process variables are computed using the optimal relationships between P1 and P2. Then, these values are used to initialize the majority of the remaining optimization variables in order to solve problem P1 (simplified objective function). Then, the obtained optimal solution can be used as initialization to solve the same mathematical model of P1 but involving a detailed and a more complete objective function (total annual cost). Finally, it is important to mention that the proponed method is an alternative way to solve detailed optimization problem and is being developed in order to be applied for any chemical/industrial process.

Conclusions

Overall efficiency analysis of the post-combustion CO2 capture using aqueous solution of amines has been performed using a non linear programming model (NLP). The proposed model is flexible enough to be used as optimizer and simulator as well. Also an alternative solution procedure for rigorous optimization of detailed models has been presented. As a first step, the proposed method involves a pre-processing phase in order to provide initial values which guarantee the convergence of the detailed and rigorous economic models. More precisely, initial values and bounds can be obtained when the optimization problem proposed in section 2 is related with an economic optimization problem. In this sense, a formal context for the use of â&#x20AC;&#x153;thermodynamicâ&#x20AC;? models in solving complex optimization problems that arise in the area of design of chemical processes was introduced. The preliminary results are promising and the proposed method is being analyzed and tested for different problem specifications. A complete discussion about its capability and limitations will be presented in future articles. In addition, the coupling of the proposed model into combined cycle power plants will also be addressed. Thus, the resulting model will allow to optimize the power plant and CO2 capture process simultaneously.

C O

C A P T U R E

Appendix A. Mathematical Model Equality constraints Absorber and stripper

Figure 13 schematically shows a non-equilibrium stage “z” used to model the absorber and regenerator columns. In this figure, z (z = 1,…, N), j (j = CO2, N2, H2O, O2) and i (i = CO2, MEA, H2O, MEAH+, MEACOO-, HCO3-, CO32-, H3O+, HO-) denote each stage and either gas and liquid component, respectively. As shown, the vapour goes up into stage z from stage z-1 and the liquid flows down into stage z from stage z+1.

Stage z (Pz ; Tz)

Figure 13. Generic stage “z”.

By considering all presented hypotheses the following rate based model was derived. Overall mass and energy balances in stage “z”

Lz +1 − Lz + G z −1 − G z = 0

(A1)

(

L z +1 H zL+1 − L z H zL + G z −1 H Gz −1 − G z H Gz + (∆H R )z − ∆H H 2O

)

(A2)

where L, G, HL and HG refer to the liquid and vapor flow-rates and enthalpies respectively. ∆HR and ∆HH2O are the heat released by the reaction and vaporization heat of water and the corresponding correlations are taken from Oyenekan et al. (2007) and Hilliard et al. (2008). Species mass balance in stage “z” L z +1 x i ; z +1 − L z x i ; z + G z −1 y j ; z −1 − G z y j ; z = 0

(A3)

C O

C A P T U R E

xi and yj refer to the mole fraction of component “i” or “j” in liquid and vapor phases respectively.

∑y

∑x

j = CO2 , H 2 O, N 2 , O 2

(A4)

i = CO 2 , H 2 O, MEA , MEAH + , MEACOO − , H 3 O + , HO − , HCO 3− , CO 32− (A5)

Ionic mass balance relationship in stage “z”:

[MEAH+]z + [H3O+ ]z = [MEACOO-]z + [HCO3-]z + 2 [CO32-]z + [OH-]z (A6) α [MEA]0z = [CO2]z + [MEACOO-]z + [HCO3-]z + 2 [CO32-]z

(A7)

[MEA]0z = [MEA]z + [MEAH+]z + [MEACOO-]z

(A8)

The superscript (0) refers to the initial condition. CO2 loading (α) is defined as the ratio between total CO2 and total amine. Chemical reactions and phase equilibrium relationships:

The dependences of the equilibrium constants K m of reactions R1 to R5 and Henry´s H coefficient ( CO 2 , i , z ) with the temperature are computed as follows:

(K m )z = ∏ (a i ; z )ν = ∏ (x i ; z γ i; z )ν i

∀m, m = R 1 , R 2 , R 3 , R 4 , R 5

(A9)







 Tz 



(K m )z = exp  A +  B  + C ln (Tz )

∀m, m = R 1 , R 2 , R 3 , R 4 , R 5

(A10)

  B H CO2 , i , z = exp  A +   + C ln (Tz ) + D Tz   Tz    ∀i,

i = MEA, H 2 O (A11)

T refers to the absolute temperature (K) and aiz, γiz, νi are, respectively, activity, activity coefficient and stoichiometric coefficient for component “i” in reaction “m” at stage “z”. Liquid phase has ideal behavior, therefore the activity coefficients are set to one (Kent-Eisenberg model). 46

C O

C A P T U R E

yCO2 z ϕCO2 z Pz = H CO2 z [CO2 ]z

(A12)

y H 2O z ϕ H 2O z Pz = p H 2O z xH 2O z

(A13)

where [i]z is the molar concentration of specie “i” in stage “z”. φz, Pz and pH2Oz refer to fugacity coefficient, total pressure and partial pressure of water, respectively. Solubility of CO2 in MEA solution (HCO2), which is corrected for solution ionic strength (I), is calculated as follow (Greer et al. (2008)):

(

)

 xH O z H CO2 MEA z + xCO2 z H CO2 H 2O z  H CO2 z = 100.152 I z  2  ρ zL  

(

Iz = 1

)

2∑ i

ψ i [i ]z

(A14)

i = MEAH + , MEACOO − , H 3O + , OH − , CO32− , HCO3− (A15)

∀i,

where ψi is the ion charge. Values of the coefficients used in eq. (7) and (8) are taken from Aboudheir et al. (2003) and Liu et al. (1999). Antoine equation is used to predict the partial vapour pressure of water (pH2O). Enhancement Factor

The influence of the reactions on the CO2 transfer is considered by an enhancement factor (E) which is defined as follows: E z=

(D ) [(k L CO 2 z

) [MEA] + (k

r , CO 2 − MEA z

) [CO ] ]

r , CO 2 − OH z

L z

2 z

(A16)

The forward constants (kr, CO2-MEA and kr, CO2-OH) of the parallel and kinetically controlled reactions (R6-R7) are taken from Kucka et al. (2002) and Aboudheir et al. (2003) and are computed as follows:

(k (k

)

r ,CO 2 − MEA z

r ,CO 2 −OH

)

 44940   = 4.495 × 1011 exp  −  R Tz 

 6658   = exp  31.396 − Tz  

(A17)

(A18)

C O

C A P T U R E

The constraints used to compute CO2 diffusivity (DLCO2) and liquid-side mass transfer coefficient (kLz) are taken from Greer et al. (2008) and Onda et al. (1968), respectively. They depend on the viscosity and mass density, the nominal packing size, specific dry area and the effective interfacial area for mass transfer. Effective Interfacial Area of Packing for mass transfer

A number of correlations exist for calculating the effective interfacial area for mass transfer. In this work, the correlation proposed by Onda et al. (1968) is considered.    σ  a z = at 1 − exp − 1.45 c  σz   

   

0.75

 L'z     a µL   t z 

0.1

( ) ( )

 L' 2 a  t   z  ρ 'L 2 g   z 

−0.05

( )

 L' 2 z   ρ z'L σ z at 

   

0.2

    

(A19)

Where G´and L´ are gas and liquid mass velocities (Kg/ m2 s), ρ’G and ρ’L are gas and liquid mass densities (Kg/m3); h is the stage height, σ is the liquid surface tension and, σc and at are the surface tension and specific area (m2/m3) of the packing material. Pressure Drop

The following correlation computes the total pressure drop along the stage “z”. ∆Pz= (∆Pd)z + (∆Pl)

(A20)

where (∆Pd)z and (∆Pl)z refer, respectively, to the dry pressure drop and pressure drop due to the liquid presence and depend on liquid and gas flow rates, liquid and gas densities, liquid viscosity, gas velocity, dry packing factor and operative pressure (Robbins et al. (1990)). Finally, the total pressure drop along the column is given by: ∆P = ∑ ∆Pz hz

(A21)

Dependence of the stage efficiency with the absorber height and process variables

According to the hypothesis, the stage efficiency can be computed as:    HTU z    = 1 − exp − η z = 1 − exp − h z     

    G 'z L'z + λ   z L G 'G  'L  R Tz a z k z ρ z   k z az ρ z Ez    hz   

(A22)

C O

C A P T U R E

where:

h z = HTUz × NTUz

(A23)

NTUz and HTUz are the number of transfer units based in the total mass transfer coefficient and the transfer unit height, respectively. λ is the stripping factor (λ=m/(L/G)), and kG is the gas-side mass transfer coefficient (Kmol/Pa s m2). Finally, as mentioned earlier, h, a, kL and E are the height of the stage (m), effective interfacial area for mass transfer (m2/m3), liquid-side mass transfer coefficient (m/s) and dimensionless Enhancement factor, respectively. Column diameter.

The diameter of each stage (DTz) is computed as follows:

DTz =

4 Gz u sg,z π ρGz

(A24)

where ρG, G and usg,z refer to the gas density, flow-rate and superficial velocity respectively. Then, usg,z is related to the flooding velocity (ufz) by eq. (25) usg,z = fz ufz

(A25)

where fz ranges from 0.7 to 0.8 (lower and upper bounds). Flooding velocity is estimated in terms of the packing factor, which depends on the type and size of packing, viscosities, densities and flow rates through Leva’s correlation (Leva et al. (1992)). Heat exchangers

Reboiler and condenser units are modeled as equilibrium stages (η=1), therefore similar equations to column stages are considered. On the other hand, rich-lean heat exchanger and coolers are computed by the following equations: Mass and energy balances

(x i L )cold , in , e = (x i L )cold , out , e

e = EXCH − 1, COOLER2, COOLER − C

(x i L )hot , in , e = (x i L )hot , out , e

(A26) (A27)

Q e = (L h )hot , in , e − (L h )cold , out , e = (L h )hot , out , e − (L h )cold , in , e 49

(A28)

C O

C A P T U R E

Qe is the transferred energy (KJ/s) on each exchanger (e). Heat transfer area Q e = U e × A e × LMTD e

e = EXCH − 1, COOLER2, REB − 1, COND − 1, COOLER − C (A29)

where U, A and LMTD are the overall heat transfer coefficient (KW / m2 K), the transfer area (m2) and the logarithmic mean temperature difference (K) for each heat exchanger (e). Heat transfer coefficients are adopted from the literature. Pumps

Energy required for amine pumping is estimated using the following equation:

Wp =

L  ×  pL  × (Pout , p − Pin , p ) η p  ρ p  1

(A30)

where W, L, ρL, Pin, Pout and η are the pumping work (KW), molar liquid flow rate (mol/s), molar density (mol/m3), in and out pressures (kPa) and efficiency of each pump (p) respectively. Compressors and blower

The temperature change due to compression work (Tout,c) , the compressor power requirement (Wc) for each stage (c) and the total compression power requirement (WTC) are calculated by eq. (31)-(34).

Tout , c − Tin , c

1 1−   γ c   Pout , c 1   = Tin , c  − 1  ηc P  in , c  

(A31)

1 1−   γ 1  R × Tin , c × Z c × γ c   Pout , c  c    × Wc =   × Vin , c ×  − 1    γ c −1  ηc     Pin , c   

 Pout ,c −1   Pout ,c  =  P    in,c −1   Pin ,c

   

(A32)

(A33) 50

C O

C A P T U R E

WTC =

∑W

(A34)

c =1

where Z and γ are the compressibility factor and the isentropic efficiency, respectively. Similar equations are used to model the flue gas blower. The model also includes constraints to compute the enthalpies, fugacity coefficients and surface tension, among others, which can be found elsewhere (Austgen, (1989), Freguia et al. (2002), Dugas (2006), Greer (2008)).

Inequality constraints Sizing constraints h ×N  0.625 ≤  z ≤ 1.6  DTz  Tower

T = ABS, REG

(A35)

Temperature constraints

10 ≤ (TRH − TR ) ≤ 15

(A36)

Thot ,in , e ≥ Thot , out , e + 0.5 ∀e, e = EXCH − 1, COOLER2, REB − 1, COND − 1, COOLER − C

(A37)

Tcold ,in , e ≤ Tcold , out , e + 0.5 ∀e, e = EXCH − 1, COOLER2, REB − 1, COND − 1, COOLER − C

(A38)

Thot ,in , e − Tcold , out , e ≥ 5 ∀e, e = EXCH − 1, COOLER2, REB − 1, COND − 1, COOLER − C

(A39)

Composition constraints y H 2O ,COND −1 ≤ 0.1

(A40)

TRH , TR and yH2O,COND-1 are the hot rich amine temperature, the reboiler temperature and the water mole fraction of the gas stream leaving the condenser, respectively.

C O

C A P T U R E

Acknowledgement Financial supports obtained from the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET) and the Universidad Tecnológica Nacional Facultad Regional Rosario (UTNFRRo) Argentina are greatly acknowledged.

References - Lepaumier H.; Picq D.; Carrette P.L. (2009a). New Amines for CO2 Capture. I. Mechanisms of Amine Degradation in the Presence of CO2. Ind. Eng. Chem. Res., 48 (20): 9061–9067. - Lepaumier H., Picq D., Carrette P.L. (2009b). New Amines for CO2 Capture. II. Oxidative Degradation Mechanisms. Ind. Eng. Chem. Res., 48 (20): 9068–9075. - Lepaumier H., Picq D., Carrette P.L. (2009c). Degradation Study of New Solvents for CO2 Capture in Post-Combustion. Energy Procedia, 1 (1): 893-900. - Bello A., Idem R.O. (2006). A Comprehensive Study of the Kinetics of the Oxidative Degradation of CO2 Loaded and Concentrated Aqueous Monoethanolamine (MEA) with and without Sodium Metavanadate During CO2 Absorption from Flue Gases. Ind. Eng. Chem. Res., 45(8): 2569–2579. - Dubois L., Mbasha P., Thomas D. (2009). CO2 Absorption into Aqueous Solutions of a Polyamine (PZEA), a Sterically Hindered Amine (AMP), and their Blends. Chemical Engineering & Technology, 33(3): 461-467. - Lepaumier H., Martin S., Picq D., Delfort B., Carrette P.L. (2010). New Amines for CO2 Capture. III. Effect of Alkyl Chain Length between Amine Functions on Polyamines Degradation. Eng. Chem. Res., 49 (10): 4553-4560. - Davis J., Rochelle G. T. (2009), Degradation of Monoethanolamine at Stripper Conditions. Energy Procedia, 1: 327. - Dexton A. J., Rochelle G. T. (2009). Catalysts and Inhibitors for MEA Oxidation. Energy Procedia, 1: 1179. - Dawodu O. F., Meisen A. (2009). Degradation of Alkanolamine Blends by Carbon Dioxide. The Canadian Journal of Chemical Engineering, 74(6): 960-966. - Supap T., Idem R., Tontiwachwuthikul P., Saiwan C. (2009). Kinetics of Sulfur Dioxide- and Oxygen-Induced Degradation of Aqueous Monoethanolamine Solution During CO2 Absorption from Power Plant Flue Gas Streams. International Journal of Greenhouse Gas Control, 3(2): 133-142. 52

C O

C A P T U R E

- Edali M., Aboudheir A., Idem R. (2009). Kinetics of Carbon Dioxide Absorption into Mixed Aqueous Solutions of MDEA and MEA Using a Laminar Jet Apparatus and a Numerically Solved 2D Absorption Rate/Kinetics Model. International Journal of Greenhouse Gas Control, 3(5): 550-560. - Kittel J., Idem R., Gelowitz D., Tontiwachwuthikul P., Parrain G., Bonneau, A. (2009). Corrosion in MEA Units for CO2 Capture: Pilot Plant Studies. Energy Procedia, 1(1): 791-797. - Puxty G., Rowland R., Allport A., Yang Q., Bown M., Burns R., Maeder M., Attalla M. (2009). Carbon Dioxide Postcombustion Capture: A Novel Screening Study of the Carbon Dioxide Absorption Performance of 76 Amines. Environ. Sci. Technol., 43(16): 6427–6433 - Manjula N., Amornvadee V. (2009). Corrosion in CO2 Capture Process Using Blended Monoethanolamine and Piperazine. Ind. Eng. Chem. Res., 48(20): 9299– 9306. - Bhattacharyya D., Turton R. Zitney S. (2011). Steady-State Simulation and Optimization of an Integrated Gasification Combined Cycle Power Plant with CO2 Capture. Ind. Eng. Chem. Res., 50, 1674–1690. - Khan F.M., Krishnamoorthi V., Mahmud T. (2010). Modelling reactive absorption of CO2 in packed columns for post-combustion carbon capture applications. Chemical Engineering Research and Design, In Press. (http://www.sciencedirect.com/science/article/pii/S026387621000287X). - Oyenekan B.; Rochelle G. (2009). Rate modeling of CO2 stripping from potassium carbonate promoted by piperazine. International Journal of Greenhouses Gas Control (3), 121-132. - Kvamsdal H.M., Jakobsen J.P., Hoff K.A. (2009). Dynamic Modeling and Simulation of a CO2 Absorber Column for Post-combustion CO2 Capture. Chemical Engineering and Processing: Process Intensification, 48(1):135-144. - Chalmers H., Lucquiaud M., Gibbins J., Leach M. (2009). Flexible Operation of Coal Fired Power Plants with Postcombustion Capture of Carbon Dioxide. J. Envir. Engrg., 135(6): 449-458. - Ziaii S., Rochelle G.T., Edgar T. (2009). Dynamic Modeling to Minimize Energy Use for CO2 Capture in Power Plants by Aqueous Monoethanolamine. Ind. Eng. Chem. Res., 48: 6105–6111. - Greer T., Bedelbayev A., Igreja J.M., Gomes J.F., Lie B. (2010). A Simulation Study on the Abatement of CO2 Emissions by De-Absorption with Monoethanolamine. Environmental Technology, 31(1): 107 – 115. 53

C O

C A P T U R E

- Chapel D., Ernst J., Mariz C. (1999). Recovery of CO2 from Flue Gases: Commercial Trends, Canadian Society of Chemical Engineers Annual Meeting, Saskatoon, Saskatchewan, Canada - Iijima M. (2002). CO2 Recovery/Effective Commercialization. MHI Technical Journal, 39(5): 286.

Utilization/Fixation

and

- Desideri U., Paolucci A. (1999). Performance Modeling of a Carbon Dioxide Removal System For Power Plants. Energy Conversion and Management, 40: 18991915. - Ali, C. (2004). Simulation and optimisation of a coal power plant with integrated CO2 capture using MEA scrubbing. Master Thesis. University of Waterloo. - Jassim M.S., Rochelle G.T. (2006). Innovative absorber/stripper configurations for CO2 capture by aqueous monoethanolamine. Ind & Eng Chem. Res., 45(8):2465– 2472. - Brooke A., Kendrick D., Meeraus A. (1996). GAMS – A User’s Guide (Release 2.25), The Scientific Press, San Francisco, CA. - Drud A.S. (1992). CONOPT, A GRG Code for Large Scale Nonlinear Optimization. Reference manual, ARKI Consulting and Development A/S, Bagsvaerd, Denmark. - Oyenekan B. (2007). Modeling of Strippers for CO2 Capture by Aqueous Amines. Ph.D. Dissertation, University of Texas at Austin. - Hilliard M.D. (2008). A Predictive Thermodynamic Model for an Aqueous Blend of Potassium Carbonate, Piperazine and Monoethanolamine for Carbon Dioxide Capture from Flue Gas. Ph.D. Dissertation, University of Texas of Austin. - Aboudheir A., Tontiwachwuthikul P., Chakma A., Idem R. (2003). Kinetic oft Reactive Absorption of Carbon Dioxide in High CO2-Loaded, Concentrated Aqueous Monoethenolamine Solutions. Chemical Engineering Science, 58: 5195-5210. - Liu Y., Zhang L., Watanasiri S. (1999). Representing vapor-liquid equilibrium for an aqueous MEA-CO2 system using the Electrolyte Nonrandom-Two-Liquid model. Industrial & Engineering Chemistry Research 38, 2080-2090. - Kucka L., Kenig E.Y., Górak A. (2002). Kinetics of the Gas−Liquid Reaction between Carbon Dioxide and Hydroxide Ions. Industrial & Engineering Chemistry Research, 41: 1105-1112. - Greer T. (2008). Modeling and simulation of post combustion CO2 capturing . Ph. D. Thesis, Telemark University College, Faculty of Technology, Porsgrunn, Norway.

C O

C A P T U R E

- Onda K., Takeuchi H., Okumoto Y. (1968). Mass Transfer Coefficients between Gas and Liquid Phases in Packed Columns. J. Chem. Eng. Jpn., 1(1): 56-52. - Robbins L.A. (1990). Improved pressure drop prediction with a new correlation. Chemical Engineering Progress 87, 91. - Leva M. (1992). Reconsider packed-tower pressure-drop correlations. Chemical Engineering Progress 88, 65-72. - Austgen D. M, Rochelle G.T. (1989). Model of vapor-liquid equilibria for aqueous acid gas-alkanolamine systems using the electrolyte-NRTL equation. Ind. Eng. Chem. Res. 28: 1060-1073. - Freguia, S. (2002). Modeling of CO2 Removal from Flue Gases Using MEA. M.S. Thesis, University of Texas at Austin. - Dugas E.R (2006). Pilot Plant Study of Carbon Dioxide Capture by Aqueous onoethanolamine. M.S.E. Thesis, University of Texas at Austin.

C O

C A P T U R E

C O

C A P T U R E

Chapter

3 Hydrate-Based CO2 Captures from Flue and Fuel GasesXiao-Sen Li and Chun-Gang Xu Key Laboratory of Renewable Energy and Gas Hydrate, Guangzhou Institute of Energy Conversion, Chinese Academy of Sciences, Guangzhou 510640, P. R. China E-mail :

xucg@ms.giec.ac.cn (P. Mores) lixs@ms.giec.ac.cn (S. Mussati)

he basis for the hydrate-based carbon dioxide (CO2) captures from flue and fuel gases is the selective partition of the CO2 component between the hydrate phase and the gaseous phase. Compared to hydrogen (H2) or nitrogen (N2), the equilibrium hydrate formation pressure of CO2 is much lower than that of H2 or N2 at the same temperature. It is expected that CO2 encages into the hydrate crystal phase preferentially. Then, the hydrates are separated and subsequently decomposed to create the CO2-rich steam, while the rest constitutes the CO2-lean one. The main challenges for hydrate-based CO2 captures from flue and fuel gases are to moderate the operating condition, to accelerate the hydrate formation, to increase the gas uptake, to improve the CO2 separation efficiency. In order to meet the challenges, the promoters of gas hydrate formation are employed to improve the gas separation based on the thermodynamic and kinetic characteristics of the gas hydrate formation. Tetran-butyl ammonium bromide (TBAB) as a promoter is investigated in the hydratebased CO2 capture from flue and fuel gases. TBAB can form a semi-clathrate hydrate [6(512)•4(51262)•4(51263)•38H2O], and CO2 can be encaged into the dodecahedral cavities (512) of the TBAB hydrate at favorable stability conditions. It is found that the TBAB of 0.29 mol% is optimum for effectively lowering the equilibrium hydrate formation pressure at the same temperature for capturing CO2 from both flue and fuel gases. Furthermore, quite small quantities of adjuvant additives are added into the TBAB solution to further enhance the hydrate-based CO2 captures. The addition of the 0.028 mol% Dodecyl trimethyl 57

C O

C A P T U R E

ammonium chloride (DTAC) into the 0.29 mol% TBAB solution make the large amount of the gas hydrates form instantaneously at as low as 0.66 MPa and 277.15 K for post-combustion capture of CO2. At the condition, the induction time is shortened considerably. CO2 is purified from 17.0 mol% to 99.4 % with the two-stage hydrate separation process. In addition, the addition of cyclopentane (CP) into the 0.29 mol% TBAB solution can remarkably enhance the CO2 separation efficiency for fuel gas and shorten the induction time. The induction time in the system with the optimum CP/TBAB solution volume ratio of 5 vol% is shortened to 0.2 minutes at 4.0 MPa/274.65 K for pre-combustion capture of CO2, the gas uptakes of more than 80 mol% are obtained within 6 ~ 7 minutes, and a 96.0 mol% CO2-rich and a 86.5 mol% H2-rich gas can be obtained from one-stage pre-combustion separation. The results provide the data and the supports for the developments of the processes for the CO2 captures from flue and fuel gases.

Introduction Carbon dioxide (CO2) as one main greenhouse gas contributes itself to greenhouse effect accounted for about 60% [1]. In order to deal with the challenge of global warming and carry out the Kyoto protocol, to reduce the emission of CO2, especially for the CO2 emitted from fossil fuel power plants, is imperative. Various methods such as cryogenic fractionation, selective adsorption, gas absorption and membrane process, have been proposed. However, the above methods have their individual issues of either high corrosion, large energy consumption, high cost, or low capacity [2]. Accordingly, a new efficient and more cost-effective technology which is different from the conventional methods needs to be explored. The hydrate separation method for gas mixtures is a novel gas separation technique. Gas hydrates are crystalline hydrogen-bonded networks of water molecules that form polyhedral cavities large enough to accommodate small molecules of gases, such as hydrogen (H2), methane (CH4), carbon dioxide (CO2), nitrogen (N2), sulfureted hydrogen (H2S), etc.[3] Up to now, there are four hydrate structures which have been reported in the literatures, including structure I (sI), structure II (sII), structure H (sH) and semi-clathrate hydrate (sc) [4, 5]. In the pure water system, CO2 forms sI hydrate, while H2 and N2 form sII hydrates. Tetra-n-butyl Ammonium Bromide (TBAB), as an excellent hydrate promoter, forms semi-clathrate hydrates with water molecules and small gas molecules.[ 6, 7]. Hammerschimdt E G. [8] firstly found that the gas components released from the gas hydrates in the natural gas lines were different from those in the feed gases, and the concentrations of C3H8 and i-C4H10 increased among all the components. It illustrated that the gas could be separated from the gas mixtures in the process of hydrate formation. After than, the gas separation via hydrate-based technology was given serious attention. In 1997, Spencer D F. [9] developed one process of separating CO2 58

C O

C A P T U R E

from the gas mixtures via hydrate hydrate-based based technology. In the process, the CO2 was finally removed from the gas mixtures after the CO2 in the gas phase was adsorbed by CO2 nucleated water and then formed gas hydrates. The process was then used for removing CO2 from different gas mixtures, such as flue gas, IGCC syngas, etc. The basic mechanism of hydrate hydrate-based based CO2 separation and capture from fuel gases is the selective partition of CO2 component between in the hydrate phase and in the gaseous phase [10]. To take CO2/N2 and CO2/H2 gas mixtures for examples, bbecause ecause the equilibrium hydrate formation pressure of CO2 is much lower than that of N2 or H2 at the same temperature, it is expected that CO2 preferentially encaged into the hydrate crystal phase. The hydrates crystals are separated and subsequently decomp decomposed osed to create the CO2-rich rich stream, while the rest constitute the CO2-lean lean one one.. Figure 1 and Figure 2 show two typical P P-xx diagrams of CO2/N2/H2O and CO2/H2/H2O [11, 12].. As seen from Figure 1 and Figure 2, once the gas hydrate forms at one temperature and pressure, the mole fractions of CO2 in the hydrate phase and in the vapor phase are evidently different, and the mole fraction of CO2 in the hydrate phase is observable larger than that in the vapor phase, thereby, CO2 is separated from the gas mixtures successfully. uccessfully.

Figure 1 P-xx diagram of CO2/N2/H2O mixture measured at three temperatures of 274 K, 277 K and 280 K.

C O

C A P T U R E

Figure 2 P-xx diagram of CO2/H2/H2O at 282 K. The mole fraction is water water-free.

The study of hydrate hydrate-based CO2 separation technology main mainly ly focuses on CO2 separation efficiency, the gas uptake, hydrate formation induction time and hydrate hydratebased CO2 separation process. Among them, CO2 separation efficiency can be given from the two following equations [13].. The CO2 recovery or split fraction (S.Fr) of carbon dioxide is expressed as: S .Fr. =

H nCO 2 Feed nCO 2

(1)

The separation factor (S.F. (S.F.) is expressed as:

S .F =

H nCO Ă&#x2014; nXgas2 2 gas nXH2 Ă&#x2014; nCO 2

(2)

n gas n H n Feed where CO2 , CO2 and CO2 are the moles of CO2 in the gas phase phase,, in the hydrate slurry phase at the end of the hydrate formation formation,, and in the initial gas mixture, nXgas2 n XH2 respectively. , are the moles of bala balanced nced component of the binary gas mixture in the residual gas phase and in the hydrate slurry phase at the end of the hydrate 60

C O

C A P T U R E

formation, respectively. The gas uptake refers to the gas consumed from in the process of the hydrate formation, including the gas dissolved in the solution and the gas entrapped into the cavities of the hydrates. The number of moles of gas uptake at any time during hydrate formation ( ∆nH ) can be calculated by the following equation:  PV   PV  ∆n H = nG ,0 − nG, t =   −   zRT  G,0  zRT  G,t

(3)

n n Where G, 0 and G,t are the number of moles of free gas in the cell before and after hydrate formation, respectively. P, T, V, z, and R are the system pressure, the system temperature, the volume of the cell occupied by free gas, the gas compressibility factor and the universal gas constant, respectively. In this chapter, we mainly introduce the technologies of hydrate-based CO2 separation from flue gas and fuel gas and the relative methods of the study.

3-1- Equilibrium hydrate formation conditions 3-1-1- Research method of the equilibrium hydrate formation The method is similar to the T-cycle method, which is reliable to determine the equilibrium hydrate formation temperatures [14]. The crystallizer is cleaned using deionized water and allowed to dry firstly. TBAB aqueous solution prepared at a desired concentration is introduced into the high-pressure hydrate crystallizer. Subsequently, the hydrate crystallizer with the solution is evacuated with a vacuum pump and flushed with the mixture gas at least four times to ensure that it was air-free. In the study, at first, the hydrate crystallizer is pressurized up to the desired pressure by supplying the mixture gas. Then, the system temperature is dropped and remained constant at the point of pressure depression which is caused by the gas/TBAB mixed gas hydrate formation. The system is heated quite gradually with each temperature step of 0.1 K until there is an infinitesimal amount of gas hydrate left. The interval time is taken at least for one day to determine the equilibrium state at each temperature step. When no particle of hydrates appeared during the gradual heating, the point is determined as the equilibrium point. In addition, the formation conditions for pure TBAB hydrate were also measured using the T-cycle method.

3-1-2- Equilibrium conditions for CO2/H2/TBAB system Prior to the experiment for the CO2/H2/TBAB hydrate system, five equilibrium hydrate formation data for the CO2/H2 gas mixture of the mole fraction of 39.2 % CO2 were measured in the temperature range of (274.05 to 278.75) K and compared with 61

C O

C A P T U R E

the values given by Kumar et al. [15] As shown in Figure 3, the results are in excellent agreement with the literature values. To determine the operating temperature and the minimum hydrate formation pressure for CO2 separation for the first stage IGCC fuel gas (CO2/H2) with the addition of TBAB, the measurements of the equilibrium hydrate formation conditions for CO2/H2/TBAB hydrate system in the pressure range of (0.25 to 6.23) MPa and the temperatures in the range of (275.15 to 288.55) K were carried out. TBAB aqueous solutions containing the mole fraction of (0.14, 0.21, 0.29, 0.50, 1.00, and 2.67) % of TBAB and the CO2/H2 gas mixture containing a mole fraction of 39.2 % CO2 were used to form the hydrates. The phase equilibrium data are summarized in Table 1 and plotted in Figure 3. As shown in Figure 3, the addition of TBAB reduces considerably the pressure required to form the hydrate, and the equilibrium formation pressure decreases with the increase of the concentration of TBAB at the same temperature. For example, the equilibrium hydrate formation pressure for the gas hydrate compositions of a mole fraction of 39.2 % CO2 and a mole fraction of 60.8 % H2 is 11.01 MPa at 278.75 K. The pressure is reduced to 3.15 MPa at the same temperature in the presence of a mole fraction of 0.14 % TBAB in the liquid phase. As shown, there is an approximate 71.38 % reduction of the equilibrium pressure on account of the addition of only 0.14 % mole fraction of TBAB, and with the increase of the mole fraction of TBAB from (0.14 to 0.29) %, the pressure decreases to 0.85 MPa. In addition, at the same TBAB concentration, the equilibrium hydrate formation pressure increases with the increase of the equilibrium temperature.

12.00

10.00

P/MPa

8.00

6.00

4.00

2.00

0.00 273.00

276.00

279.00

282.00

285.00

288.00

T/K

Figure 3 Equilibrium hydrate formation pressure for CO2 (1) + H2 (2) gas mixture with x1 = 39.2 % and x2 = 60.8 % in the presence of TBAB (3) + H2O (4). ■, x3 = 0.00, Kumar et al.[16]; □, x3 = 0.00; ●, x3 = 0.14 %; ▲, x3 = 0.21 %; ▼, x3 = 0.29 %; ○, x3 = 0.50 %; △, x3 = 1.00 %; ▽, x3 = 2.67 %.

C O

C A P T U R E

Table 1 Equilibrium hydrate formation pressure of CO2 (1) + H2 (2) gas mixture with x1 = 39.2 % and x2 = 60.8 % in the presence of TBAB (3) + H2O (4) 100 x3

T/K

P/MPa

0.00

274.05

5.75

275.45

6.75

276.85

8.03

277.85

9.84

278.75

11.01

275.15

0.51

276.25

1.20

277.15

1.71

278.25

2.67

279.55

3.88

281.15

5.21

276.25

0.50

277.15

1.00

278.35

1.95

279.55

3.10

0.14

0.21

0.29

0.50

1.00

2.67

280.65

4.19

277.35 277.85 278.15 278.65

0.25 0.40 0.50 0.71

279.55

1.34

280.55

1.96

281.15

2.34

281.55

2.66

282.85

4.05

284.55

6.23

279.55

0.25

280.35

0.50

281.95

1.55

283.25

2.41

284.25

3.50

285.05

4.58

282.45

0.52

282.75

0.80

283.80

1.42

284.95

2.22

286.25

3.20

287.35

4.63

285.95

0.50

286.55

1.17

287.25

2.01

287.95

3.11

288.55

4.20

C O

C A P T U R E

After our first-stage separation, the residual gas contains an approximate mole fraction of 18 % CO2. Thus, the residual gas requires further separation as the second-stage feed gas. Experimental measurements of equilibrium hydrate formation conditions were measured for the CO2/H2 gas mixture of the mole fraction of 18.5 % CO2 in the presence of TBAB with the mole fraction of (0.14, 0.21, 0.29, 0.50, and 1.00) %. The experiments were carried out in the pressure range of (0.48 to 7.26) MPa and the temperature range of (274.45 to 285.45) K. The experimental data are tabulated in Table 2 and plotted in Figure 4. As shown in Figure 4, the equilibrium hydrate formation conditions obtained have the similar characteristics with that for the CO2/H2 gas mixture of a mole fraction of 39.2 % CO2 with the different concentration of TBAB. The equilibrium formation pressure decreases with the increase of the concentration of TBAB at a certain temperature.

8.00

P/MPa

6.00

4.00

2.00

0.00 274.00

276.00

278.00

280.00

282.00

284.00

286.00

T/K

Figure 4 Equilibrium hydrate formation pressure for CO2 (1) + H2 (2) gas mixture with x1 = 18.5 % and x2 = 81.5 % in the presence of TBAB (3) + H2O (4). ■, x3 = 0.14 %; ●, x3 = 0.21 %; ▲, x3 = 0.29 %; ▼, x3 = 0.50 %; ◆, x3 = 1.00 %.

C O

C A P T U R E

Table 2 Equilibrium hydrate formation pressure of CO2 (1) + H2 (2) gas mixture with x1=18.5 % and x2=81.5 % in the presence of TBAB (3) + H2O (4) 100 x3

T/K

P/MPa

0.14

274.45

0.48

275.15

1.19

276.75

3.01

277.75

5.05

278.65

7.04

275.65

0.51

276.15

1.17

277.55

3.03

278.45

5.05

0.21

0.29

0.50

1.00

279.45

7.07

276.85

0.50

277.45

1.27

279.35

3.04

280.55

5.21

281.65

7.18

279.55

0.52

280.05

1.15

281.55

2.97

282.45

5.04

283.35

7.26

281.75

0.51

282.15

1.07

283.65

3.00

284.55

5.02

285.45

7.04

Figure 5 gives the typical curve of the effect of the concentration of CO2 on the equilibrium hydrate formation conditions for the CO2/H2 gas mixture in the presence of TBAB with the same concentration. As shown in Figure 5, the equilibrium hydrate formation pressures for the two CO2/H2 mixture containing a mole fraction of 60.8 % H2 and a mole fraction of 81.5 % H2 are (1.96 and 5.21) MPa, respectively, in the presence of a mole fraction of 0.29 % TBAB, at 280.55 K. The same behavior can be observed for the other given TBAB concentration and temperature. This illustrates that, as the relative amount of H2 to CO2 increases in the gas mixture, the equilibrium shifts to higher pressures at a given temperature with the addition of TBAB under the given concentration.

C O

C A P T U R E

Figure 6 shows the hydrate formation conditions for CO2/H2 mixtures containing a mole fraction of approximately (38.3 to 39.9) % CO2 with the four different additives, propane (C3H8) [15], THF [17], cyclopentane (CP) [11], and TBAB. With the effect of different additives, it can be found that the four additives all can effectively reduce the equilibrium hydrate formation pressure of the CO2/H2 mixture, to the different degrees, compared with that of the pure CO2/H2 mixture. For example, at 277.85 K, the phase equilibrium pressure of the CO2/H2 mixture is 9.84 MPa without addition of any additive, 5.10 MPa [15] with the addition of a mole fraction of 3.20 % C3H8 in the vapor phase, 1.43 MPa [17] with the addition of a mole fraction of 1.00 % THF in the liquid phase, and 0.40 MPa with the addition of a mole fraction of 0.29 % TBAB in the liquid phase, respectively. That is, the presence of a mole fraction of 0.29 % TBAB, a mole fraction of 1.0 % THF, and a mole fraction of 3.2 % C3H8 can make the hydrate formation pressure of the CO2/H2 mixture decrease by (95.93, 85.46, and 48.17) %, respectively, at 277.85 K.

8.00

P/MPa

6.00

4.00

2.00

0.00 276.00

278.00

280.00

282.00

284.00

T/K

Figure 5 Equilibrium hydrate formation pressure for CO2 (1) + H2 (2) gas mixture with different CO2 concentration in the presence of TBAB (3) + H2O (4) with x3 = 0.29 %. â&#x2013; , x1 = 39.2 %; â&#x2014;?, x1 = 18.5 %.

C O

C A P T U R E

12.00

10.00

P/MPa

8.00

6.00

4.00

2.00

0.00 273.00

276.00

279.00

282.00

285.00

288.00

291.00

T/K

Figure 6 Effects of different additives on equilibrium hydrate formation pressure of CO2 + H2 gas mixture. ■, pure water; □, 3.2 mol% C3H8 (dry gas), Kumar et al. [16]; △, 1.0 mol% THF (liquid), Lee et al.[17]; ▽, CP (the volume ratio of CP to water = 1.5), Zhang et al.[11]; ●, 0.29 mol% TBAB (liquid); ▲, 2.67 mol% TBAB (liquid).

3-2- Hydrate-based CO2 separation and capture from flue gas The main components of the flue gas emitted from fire-power plant are CO2 and N2. CO2/N2 gas mixture containing 17.0 mol% CO2 and 65.0 mol% CO2 are used in the work to simulate flue gas mixture.

3-2-1- Induction time of hydrate formation Figure 7 gives the change of the induction time of the hydrate formation for the CO2/N2 gas mixture with 17.0 mol% CO2 in the 0.29 mol% TBAB aqueous system in the presence of DTAC with the different concentrations vs. the initial pressure at 274.95K. It can be seen from Figure 7 that the addition of DTAC results in a remarkable reduction of the induction time, and the induction time decreases with the increase of concentration of DTAC at the fixed initial pressure. For example, when the initial pressure is 1.66MPa, the induction time for pure TBAB aqueous solution is 31.0 minutes. However, when DTAC with the concentration from 0.014 mol% to 0.056 mol% are added into the 0.29 mol% TBAB aqueous solution, the induction time is shortened from 7.1 minutes to 3.0 minutes. It may be due to the fact that DTAC as a 67

C O

C A P T U R E

surfactant can change the surface activity of the solution and lower its surface tension. As a result, the addition of DTAC promotes the dissolving ability of mixture gases in the TBAB solution, and furthermore enhances the degree of supersaturation of CO2. According to the expression for the rate of hydrate nucleation (J) given by Nataraja n [18], J = k × (S − 1) , in which k and n are the constants, and S is the super-saturation ratio, it can be seen that the nucleation rate increases with the degree of supersaturation resulting in the decrease of the induction time. However, as shown in Figure 3, the reduction of the induction time is quite small, when the concentration of DTAC changes from 0.028 mol% to 0.056 mol% at the fixed initial pressure. According to the characteristics of the surfactant [19, 20], the surface tension of solution is the smallest when the concentration of DTAC is critical micellar concentration (CMC) of 1.6×10-2 mol/L which is estimated by use of the way of Zhong [21]. Its mole percentage is approximately 0.028 mol%. However, when the concentration exceeds its CMC, the increase of concentration cannot change surface activity any more, and the surface tension cannot be also further decreased. Thus, the super-saturation of CO2 in the solution can also not rise any more. Hence, the induction time of the hydrate formation cannot be shortened in spite of the increase of the DTAC concentration. This is the reason why the induction time has little change with the increase of the DTAC concentration from 0.028 mol% to 0.056 mol%. The similar phenomenon was found by Watanabe.K et al. [22]. In addition, it can be also seen from Figure 3 that the induction time reduces with the increase of the initial pressure. This is because the higher initial pressure creates the higher super-saturation which results in the higher nucleation rate.

0.00 % DTAC 0.014% DTAC 0.028% DTAC 0.056% DTAC

Induction time (min)

0 0.5

1.0

1.5

2.0

2.5

3.0

Initial pressure (MPa)

Figure 7 Induction time of the hydrate formation for 17.0 % CO2/N2 gas mixture in TBAB aqueous system in the presence of DTAC with different concentrations vs. initial pressure at 274.95K

C O

C A P T U R E

Likewise, Figure 8 gives a typical comparison for the induction time of hydrate formation for 65.0 mol% CO2/N2 gas mixture as the feed gas for the second stage separation with and without the addition of DTAC. It can be seen from Figure 8 that a small amount of DTAC reduces the induction time significantly at 277.15K. For example, the induction time with 0.028 mol% DTAC is about 0.4minutes and is one thirteenth of that without DTAC at initial pressure of 1.66MPa. In addition, it can be found from Figures 7 and 8 that for 65.0 mol% CO2/N2 mixture gas as the feed gas for the second stage separation, the addition of DTAC can more remarkably reduce the induction time of the hydrate formation, compared with that for 17.0 mol% CO2/N2 mixture gas. It is because 65.0 mol% CO2/N2 mixture gas has the higher component of CO2. Furthermore, it is easier to be induced to form the hydrate under the function of DTAC.

without DTAC 0.028 mol% DTAC

Induce time (min)

0 0.5

1.0

1.5

2.0

2.5

3.0

Initial pressure (MPa)

Figure 8 Induction time of the hydrate formation for 65.0 mol% CO2/N2 mixture gas in 0.29 mol% TBAB aqueous system in the presence of 0.028 mol% DTAC vs. initial pressure at 277.15K.

3-2-2- Pressure drop during hydrate formation In this study, the gas uptake is reflected by the pressure drop during the hydrate formation due to the experiment is carried out volume constantly. Figure 9 gives the pressure drops in the system vs. the DTAC concentration in 0.29 mol% TBAB aqueous solution with the different initial pressures at 274.95 K using 17.0 mol% CO2/N2 mixture gas as the feed gas. As shown in Figure 9, with the increase of the concentration of DTAC, the pressure drop (â&#x2013;łP1) increases obviously. In the process of 69

C O

C A P T U R E

formation of gas hydrate, when the concentration of DTAC is less than CMC, the gas super-saturation in the solution increases, that is the amount of gas dissolving in water increase, with the increase of DTAC. When the concentration of DTAC is more than CMC, the DTAC molecules associate as the micelles containing solubilized gas, and the concentration of the micelles also increase with the increase of the concentration of DTAC resulting in the increase of solubilized gas corresponding to the decrease of the pressure in the system and that is the increase of △P1. Likewise, as shown in Figure 5, the higher initial pressure causes the bigger pressure drop in the process of the hydrate formation. It attributes to the fact that more gas dissolves in the solution and forms the hydrate with the increase of the initial pressure, resulting in bigger pressure drop.

Pressure drop during formation (MPa)

0.45 0.40

initial pressure 1.66 MPa initial pressure 2.16 MPa initial pressure 2.66 MPa

0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0.00

0.02

0.04

0.06

0.08

DTAC (mol%)

Figure 9 Pressure drops (∆P1) in the system vs. DTAC concentration in 0.29 mol% TBAB aqueous solution with different initial pressures at 274.95 K with 17.0 mol% CO2/N2 mixture gas.

3-2-3- CO2 concentration in hydrate slurry phase The effect of DTAC on promoting the gas to go into the hydrate slurry phase can be seen from Figure 10, which shows the pressure increase (∆P2) in the system after the hydrate dissociation vs. the initial pressure in the presence of DTAC with the different concentrations in 0.29 mol% TBAB aqueous solution at 274.95K with 17.0 mol% CO2/N2 feed gas. As seen, the pressure increase (∆P2) rises with the increases of the concentration of DTAC and initial pressure. It is noted that there is a remarkable 70

C O

C A P T U R E

increase of â&#x2C6;&#x2020;P2 due to the presence of DTAC with the concentration of 0.014 mol% 0.056 mol%, compared with that for the pure TBAB solution. It demonstrates that either DTAC or the increase of initial pressure can conduce to the increase of the gas storage in the hydrate slurry phase.

Pressure increases after dissociation (MPa)

The changes of CO2 concentration in the hydrate slurry vs. the initial pressures in the presence of DTAC with the different concentrations in 0.29 mol% TBAB aqueous solution at 274.95K with 17.0 mol% CO2/N2 feed gas are shown in Figure 11. As seen, the concentration of CO2 in hydrate slurry phase decreases with the increase of DTAC at the fixed initial pressure, and a substantial reduction occurs at the DTAC concentration of more than 0.028 mol%. As mentioned on the above, the addition of DTAC lowers the surface tension of TBAB solution, and promotes the dissolving ability of the mixture gas. Thus, it enhances not only the amount of CO2 but also the amount of N2 to dissolve into the TBAB solution. However, relative to CO2, more N2 goes into the solution. This is due to the hydrophobic groups of DTAC molecules preferentially adsorb N2 molecules. This results in the reduction of the CO2 concentration in the hydrate slurry phase with the increase of the DTAC concentration. In addition, when the concentration of DTAC is more than its CMC, the DTAC molecules in the solution form the Water/Oil micelles, with the function of the hydrophobic group of the surfactant molecules micellized, substantial N2 can be preferentially enclosed into the micelles compared to CO2. This causes the remarkable reduction of concentration of CO2 at the DTAC concentration of 0.056 mol%, as shown in Figure 11.

0.24

0.014 mol% DTAC 0.028 mol% DTAC 0.056 mol% DTAC 0.000 mol% DTAC

0.22 0.20 0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.5

1.0

1.5

2.0

2.5

3.0

Initial pressure (MPa)

Figure 10 Pressure increase (â&#x2C6;&#x2020;P2) in the system after hydrate dissociation vs. initial pressure in the presence of DTAC with the different concentrations in 0.29 mol% TBAB aqueous solution at 274.95K with 17.0 mol% CO2/N2 mixture gas. 71

C O

C A P T U R E

0.68

0.000 mol% DTAC 0.014 mol% DTAC 0.028 mol% DTAC 0.056 mol% DTAC

0.66

CO2 concentration (%)

0.64 0.62 0.60 0.58 0.56 0.54 0.52 0.50 0.48 1.0

1.5

2.0

2.5

3.0

3.5

Initial pressure (MPa)

Figure 11 CO2 concentration in hydrate slurry vs. initial pressures in the presence of DTAC with different concentrations in 0.29 mol% TBAB aqueous solution at 274.95K with 17.0 mol% CO2/N2 mixture gas.

It is noted, as shown in Figure 11, that the CO2 concentration goes up with the increase of initial pressure, and then it goes down after an inflexion point occurs. The inflexion point is also the largest point of CO2 concentration. According to the phase equilibrium data given by Deschamps et al. [23] and Arjmandi et al. [7], CO2 is prior to form CO2 hydrate at low initial pressure, compared to N2. The driving force increases with the increase of initial pressure. Furthermore, N2 can compete with CO2 for hydrate cage (512) occupancy with higher driving force [24] resulting in the dramatic decrease of the concentration of CO2 in the slurry phase after the inflexion point. The similar experimental phenomenon is also found by Fan et al. [25]. However, it can be found from Figure 11 that the inflexion point for pure TBAB solution lags behind that for TBAB+DTAC solution. Compared with those inflexion points occurring at the initial pressure of 1.66MPa in the presence of DTAC, the inflexion point in absence of DTAC is postponed to 2.66MPa. It may attributes to the fact the addition of DTAC actually lower the equilibrium hydrate formation pressure of mixture gas. Meanwhile, the equilibrium pressures of CO2 and N2 are also lowered, resulting from the reduction of the surface tension with the function of the surfactant [26-30]. Thus, N2 can begin to compete with CO2 for hydrate cage (512) occupancy at the relative lower driving force. As discussed on the above, in order to shorten the induction time of hydrate and increase CO2 storage in the hydrate slurry phase to meet the requirement of CO2 72

C O

C A P T U R E

separation from the flue gas (17.0 mol% CO2/N2 mixture gas), it is found that the 0.29 mol% TBAB+0.028 mol% DTAC aqueous solution is chosen as optimizing joint promoter for the CO2 separation at the initial pressure of 1.66MPa. It is also verified by the experimental results obtained from the change of CO2 concentration in the decomposed gas in hydrate slurry phase with 65.0 mol% CO2/N2 gas mixture in the presence of 0.028 mol% DTAC at 277.15K, as shown in Figure 12. Under the function of DTAC, the concentration of CO2 can be purified from 65.0 mole% to 99.2 mol% by forming the semi-clathrate hydrates in the condition of initial pressure 1.66MPa and temperature 277.15K. However, that the highest concentration of CO2 in the hydrate slurry phase at absence of DTAC is only 94.0 mol%. Thereby, it is possible to realize the target of capturing CO2 through a two-stage hydrate separation in industry.

1.00 0.98

CO2 concentration (%)

0.96 0.94 0.92 0.90

without DTAC 0.028 mol% DTAC

0.88 0.86 0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

Initial Pressure (MPa)

Figure 12 CO2 concentration in hydrate slurry vs. initial pressures in the presence of 0.028 mol% DTAC in 0.29 mol% TBAB aqueous solution at 277.15K with 65.0 % CO2/N2 mixture gas.

3-2-4- Two-stage separation for CO2 Capture Figure 13 shows the change of S.Fr. vs. the initial pressure in the presence of DTAC with the different concentrations in the aqueous solution of 0.29 mol% TBAB at 274.95K with 17.0 mol% CO2/N2 feed gas. As seen from the Figure 4, there is a significant enhancing of S.Fr. in the presence of DTAC, compared with that without DTAC. The value of S.Fr increases with the increase of the concentration of DTAC. However, the values of S.Fr. reach an extreme point with the four different DTAC 73

C O

C A P T U R E

concentrations of 0 mol%, 0.014 mol%, 0.028 mol% and 0.056 mol% when the initial pressure is 1.66MPa, as shown in Figure 13, and the S.Fr values with 0.028 mol% and 0.056 mol% DTAC are almost the same at the initial pressure of 1.66MPa. Hence, the operational condition at the DTAC concentration of 0.028 mol% and the initial pressure of 1.66 MPa is an optimal one for the CO2 separation via the hydration crystallization. Figure 14 gives CO2 concentration in the feed gas, the hydrate slurry phase and the residual gas phase for two stage separation with 0.028 mol% DTAC at 1.66MPa. The results indicate that after the separation for the CO2/N2 mixture gas (simulated flue gas) at 274.95K and 1.66MPa, a CO2-rich gas containing more than 65.0 mol% CO2 can be obtained, as shown in Figure 11 and Figure 14. As the industrial objective is to obtain purified CO2 with the high concentration, 65.0 mol% CO2 gas needs to be treated further with a second-stage hydrate separation. As seen in Figure 12 and Figure 14, the CO2 content is increased from initial 65.0 mol% to 99.2% after the second separation stage at 277.15K and 1.66MPa.

0.80

without DTAC 0.014 mol% DTAC 0.028 mol% DTAC 0.056 mol% DTAC

0.75 0.70 0.65 0.60 0.55

S.Fr

0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

Initial pressure (MPa)

Figure 13 Split fraction (S.Fr) vs. initial pressure in the presence of DTAC with different concentrations in 0.29 mol% TBAB aqueous solution at 274.95K with 17.0 mol% CO2/N2 mixture gas.

C O

C A P T U R E

1.0 0.9

Initial gas phase composition Hydrate composition Final gas phase composition

CO2 concentration (%)

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 First Stage

Second stage

Figure 14 CO2 concentration in feed gas, hydrate slurry phase and residual gas phase for two stage separation with 0.028 mol% DTAC in 0.29 mol% TBAB aqueous solution at 1.66MPa

Table 3 shows the values of the split fraction and the separation factor for the twostage CO2 separation with the DTAC concentration of 0.028 mol% at the initial pressure of 1.66MPa. It can be seen from Table 1 that CO2 split fractions for Stage 1 and Stage 2 are 0.54 and 0.39, respectively, and the separation factors are 9.60 and 62.25, respectively, which are superior to those given by Linga et al. [31]

Table 3 Split Fraction and Separation Factor for the two stage CO2 separation Feed gases 17.0 mol% CO2/N2

65.0 mol% CO2/N2

Experimental Conditions Stage 1 274.95K 1.66MPa DTAC 0.028 mol% Stage 2 277.15K 1.66MPa DTAC 0.028 mol%

Split Fraction

Separation Factor

0.54

9.60

0.39

62.25

C O

C A P T U R E

3-3- Hydrate-based CO2 separation and capture from fuel gas The CO2/H2 gas mixture containing 39.0~40.0 mol% CO2 is used in the work to simulate fuel gas mixture.

3-3--1- Gas uptake Figures 15(a) and 15(b) show the gas uptake profiles for the different systems with the same feed gas with the above components at 274.65 K and 4.0 MPa. As seen, the gas uptake for the CO2/H2/CP/H2O system is quite low, compared to those for the CO2/H2/TBAB/H2O and CO2/H2/TBAB/CP/H2O systems. Furthermore, it can be found that the gas uptake for the CO2/H2/TBAB/CP/H2O system is more than the sum of those for the CO2/H2/TBAB/H2O and CO2/H2/CP/H2O systems at any time. Therefore, the addition of the CP in the pure water/gas system has little effect on the gas uptake. This illustrates that the hydrate phase existing in the CO2/H2/CP/H2O system is mainly the CP hydrate rather than the CP-gas hydrate, and the addition of the CP into the TBAB solution remarkably enhances the gas uptake. This may be attributed to the fact that the synergistic effect of CP and TBAB results in the more amount of gas going into the hydrate phase. The mechanism of the promotion of the TBAB in conjunction with the CP still requires to be further investigated in our next work.

0.30 0.28

CP/TBAB/gas/H2O

0.26

TBAB/gas/H2O

0.24

CP/gas/H2O

Gas uptake (mol)

0.22 0.20 0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0

500

1000

1500

Time (s)

(a)

2000

2500

C O

C A P T U R E

0.20

CP/TBAB/gas/H2O TBAB/gas/H2O

0.007

CP/gas/H2O

0.15

0.006

Gas uptake (mol)

0.005

0.10

0.004 0.003 0.002 0.001 0.000 -0.001 0

Time (s)

0.05

0.00 0

100

150

200

250

300

Time (s)

(b) Figure 15 Gas uptake change for hydrate formation vs. time from different systems at 276.15 K and 4.0 MPa. ■:0.29 mol% TBAB solution of 180 ml with CP/TBAB ratios of 5 vol%; ★:0.29 mol% TBAB solution; ▲: pure water of 180 ml with CP/H2O ratios of 5 vol%.

Figure 16 shows the final gas uptake change vs. the CP/TBAB ratio with the 0.29 mol% TBAB solution of 180 ml at 4.0 MPa and the different temperatures. As shown in Figure 16, the value of the gas uptake firstly rises and then drops with the increase of the CP/TBAB ratio from 0 to 20 vol% at the fixed temperature. The peak value of the gas uptake occurs when the CP/TBAB ratio is 5 vol%, and the moles of gas consumed are less than ones in the pure TBAB solution when the ratio is higher than 15 vol%. In addition, it is found that the moles of gas consumed decrease with the increase of the temperature at the fixed CP/TBAB ratio. Because the equilibrium hydrate formation pressure increases with the temperature, the driving force for the gas hydrate formation decreases with the increase of the temperature at the fixed pressure, resulting in the decrease of the moles of the gas consumed.

C O

C A P T U R E

0.30

273.65 K 274.65 K 275.65 K 276.65 K 277.65 K 277.65 K

Gas uptake (mol)

0.25

0.20

0.15

0.10

0.05 0.0

5.0

10.0

15.0

20.0

CP/TBAB ratio ( vol%)

Figure 16 Gas uptake change vs. CP/TBAB ratio in 180 ml TBAB solution at 4.0 MPa and the different temperatures. (â&#x2013;ł:TBAB solution of 150 ml at 4.0 MPa)

Compared to the results in othersâ&#x20AC;&#x2122; work at the similar conditions [16, 32], the gas uptake in this work is enhanced remarkably by approximately 2 times on account of the synergistic effect of the TBAB and CP. For example, the gas uptake obtained by Lee et al. [17] from the system with the 135 ml THF aqueous solution is 0.06 mol at 4.12 MPa and 279.6 K , However, that obtained in this work is 0.12 mol at 4.0 MPa and 277.65 K when the volume of the CP/TBAB/H2O mixture with the CP/TBAB ratio of 5 vol% is 150 ml, as seen from Figure 15. Figure 17 shows the change of the gas uptake vs. the amount of the 0.29 mol% TBAB solution with the corresponding CP/TBAB ratio at 274.65 K and 4.0 MPa. As seen from Figure 17, the gas uptake increases with the increase of the amount of the TBAB solution at the given CP/TBAB ratio. However, the increase is little when the amount of the TBAB solution exceeds 180 ml. It is due to the fact that the more amount of liquid phase means the more moles of TBAB existing in the solution for the hydrate formation. Thus, at the given condition, the larger number of the TBAB results in the formation of the more hydrate containing the TBAB with more enclosed gas. This is why the moles of the gas consumed increase with the amount of the TBAB solution. On the other hand, for the given CP/TBAB ratio, the increase of the amount of the TBAB solution accompanies the corresponding increase of the mount of CP liquid on the top of the TBAB solution. Thus, as the experiment proceeds, the more CP hydrate 78

C O

C A P T U R E

can form, and further the thicker layer of the CP hydrate occurs at the CP-TBAB solution interface. The layer dramatically prevents the gas from contacting with the TBAB solution and dissolving into the TBAB solution to form the gas hydrate containing the TBAB, and the degree of the hindrance is bigger with the increase in the layer of the CP hydrate. However, when the amount of the TBAB solution exceeds 180 ml, the effect of the hindrance resulting from the continuous increase of the thickness of the corresponding CP hydrate layer reaches to a maximum value, and, at the time, little gas permeates the CP hydrate. Therefore, the amount of the gas uptake has little increase when the solution exceeds 180 ml. It can be also been seen from Figure 17 that when the amount of the solution is fixed, the gas uptake firstly rises with the increase of the CP/TBAB ratio from 3 vol% to 5 vol%, then drops with the increase of the CP/TBAB ratio from 5 vol% to 20 vol%. The same phenomenon has been explained for Figure 15.

0.22 0.20

Gas uptake (mol)

0.18 0.16 0.14 0.12 0.10

CP/TBAB ratio 3 vol% CP/TBAB ratio 5 vol% CP/TBAB ratio 8 vol% CP/TBAB ratio 15 vol% CP/TBAB ratio 20 vol%

0.08 0.06 150

160

170

180

190

200

Volume of TBAB solution (ml)

Figure 17 Gas uptake change vs. volume of 0.29 mol% TBAB solution with different CP/TBAB ratios at 274.65 K and 4.0 MPa.

3-2- CO2 selectivity Figure 18 shows the change of the CO2 mole concentration in the gas phase at the end of the hydrate formation vs. the CP/TBAB ratio at 274.65 K and the different pressures. As seen from Figure 18, the CO2 mole concentration reduces with the increase of the driving force. In addition, each of the CO2 mole concentration curves is saddle-shaped with the increase of the CP/TBAB ratio from 0 to 20 vol%. Compared to the pure TBAB solution, the addition of a small amount of CP can efficiently reduce the mole concentration of CO2 in the gas phase as the CP/TBAB ratio increases 79

C O

C A P T U R E

from 0 to 5 vol%. However, the tendency of the change turns in opposite direction when the CP/TBAB ratio exceeds 5 vol% and even the mole concentration of CO2 is higher than that of the pure TBAB solution when the CP/TBAB ratio is more than 15 vol%. As described in Zhang et al.â&#x20AC;&#x2122;s work, the CP hydrate formation at the CP-water interface can catalyze CO2 going into the hydrate phase [33]. Similarly, in this work, with the addition of a small amount of the CP, the occurrence of the CP enclathration may catalyze the gas, which is mainly CO2 because CO2 easily forms the hydrate, compared to H2, transferring into the gas hydrate containing the TBAB other than the CP/gas hydrate, because, as shown in Figure 15, the addition of the CP has little effect on the gas uptake in the H2/CO2/CP/H2O system. This also means that there is little gas enclathrated in gas/CP hydrate. However, with the further increase of the amount of the CP, the amount of gas consumed and the selectivity of CO2 over H2 entrapping into the hydrate containing the TBAB significantly reduces. It is because the CP hydrate sharply forms and substantively agglomerates at the gas/liquid interface when the CP/TBAB ratio is relatively high. The layer of the agglomeration prevents the gas from contacting with water and transferring into the solution, and further results in the relatively small amount of gas dissolving into the solution. However, H2 molecule is smaller than CO2 molecule in the CO2/H2 gas mixture, and thus H2 goes through the layer of the agglomeration diffusing into the aqueous solution more easily. Consequently, this causes the increase of the CO2 composition of the residual gas phase.

CO2 concentration in gas phase (mol%)

28.0

2.5 MPa 3.0 MPa 3.5 MPa 4.0 MPa 4.5 MPa

26.0 24.0 22.0 20.0 18.0 16.0 14.0 12.0 10.0 0.0

5.0

10.0

15.0

20.0

CP/TBAB ratio (vol%)

Figure 18 Change of CO2 concentration in gas phase vs. CP/TBAB ratio in 180 ml TBAB solution at 274.65 K and different pressures.

Figure 19 shows the change of the CO2 concentration in the gas phase vs. the amount of the TBAB solution with the different CP/TBAB ratios at 274.65 K and 4.0 MPa. As 80

C O

C A P T U R E

shown in Figure 19, the CO2 concentration decreases with the increase of the amount of TBAB solution at the fixed CP/TBAB ratio. On one hand, because CO2 has much more solubility than H2 in the TBAB solution at the same condition and the equilibrium hydrate formation pressure of CO2 is remarkably lower than that of H2, the CO2 with the higher solubility and lower equilibrium hydrate formation pressure is more easily incorporated in the hydrate [34]. Therefore, once the CO2 dissolved in the solution is incorporated in the hydrate containing the TBAB, CO2 in gas phase is supplied into the TBAB solution to keep the thermodynamic equilibrium of the mass transfer. On the other hand, the volume of the gas phase reduces with the increase of the TBAB solution at the fixed effective volume of the cell in the CR. In addition, the gas with the constant component from the SV is continuously supplied into the gas phase to maintain the pressure constant as the gas hydrate forms. The more amount of the solution results in the more gas supplied from SV going into the system enclosed in the hydrate. Furthermore, according to the law of mass balance, the increased amount of the TBAB solution leads to the lower CO2 concentration in the gas phase, based on the above given conditions. It can also be seen from Figure 19 that when the volume of the solution is fixed, the change trend of the CO2 concentration with the CP/TBAB ratio is the same with that showed in Figures 17 and 18.

CO2 concentration in the gas phase (mol%)

26.0 No CP CP/TBAB ratio 3 vol% CP/TBAB ratio 5 vol% CP/TBAB ratio 8 vol% CP/TBAB ratio 15 vol% CP/TBAB ratio 20 vol%

24.0

22.0

20.0

18.0

16.0

14.0

12.0 150

160

170

180

190

200

Volume of TBAB solution (ml)

Figure 19 Change of CO2 concentration vs. volume of 0.29 mol% TBAB solution with different CP/TBAB ratios at 274.65 K and 4.0 MPa.

Figure 20 shows the CO2 content change in the hydrate phase vs. the amount of 0.29 mol% TBAB solution with the CP/TBAB ratio of 5 vol% at 274.65 K and the different pressures. As shown in Figure 20, the CO2 content in the hydrate phase 81

C O

C A P T U R E

CO2 concentration in the decomposed gas (mol%)

decreases with the increase of the pressure at the fixed amount of the TBAB solution. It is due to the fact that H2 can compete with CO2 for the hydrate cage occupancy with the higher driving force [35], and thus this leads to the decrease of the CO2 content in the hydrate phase. The similar phenomenon and explanation can also be found elsewhere [32]. It can be also seen from Figure 20 that the CO2 content in the hydrate slurry increases with the increase of the amount of TBAB solution, but the degree of the increase is quite small compared to the effect of the pressure. By the law of conversation of mass, the CO2 content in the hydrate phase has to be high if it is low in the gas phase [36]. Thus, the reason for the change trend of the CO2 content in the hydrate slurry phase with the amount of TBAB solution is the same as that explained for Figure 19. It is noted from Figure 20 that the CO2 mole concentration in the decomposed gas from the hydrate phase reaches approximately 93 mol% after the first stage separation at 274.65 K and 2.5 MPa. It illustrates that the process provided in this work can be a promising one for separating CO2 from IGCC syngas in the future.

98.0

2.5 MPa 3.0 MPa 3.5 MPa 4.0 MPa 4.5 MPa

96.0

94.0

92.0

90.0

88.0

86.0 150

160

170

180

190

200

Volume of TBAB solution (ml)

Figure 20 Change of CO2 concentration in the decomposed gas vs. the volume of TBAB solution with CP/TBAB ratio of 5 vol% at 274.65 K and different pressures.

C O

C A P T U R E

3-3-3- CO2 separation efficiency The values of the CO2 recovery fractions (S.Fr.) and separation factors (S.F.) for all 25 experimental runs can be calculated via Eqs. (1) and (2), based on the experimental data. The results are shown in Table 4. It is found from Table 2 that the values of the CO2 recovery (S.Fr.) and separation factor (S.F.) for run 9 are optimum at the conditions of the volume ratio of CP/TBAB of 5 vol%, volume of TBAB solution of 180 ml, 274.65 K and 4.0 MPa. As a typical example, Figure 21 shows the curves of the S.Fr. and S.F.for CO2 vs. the amount of 0.29 mol% TBAB solution with the CP/TBAB ratio of 5 vol% at 274.65 K and 4.0 MPa after the first stage separation. Both the S.Fr. and S.F. increase with the increase of the amount of the 0.29 mol% TBAB solution and they reach the maximum values, 0.58 and 30.61, respectively, when the amount of TBAB solution is 180 ml. However, when the amount of the TBAB solution exceeds 180 ml, the S.Fr. has a remarkable reduction and the S.F. has little change.

0.590 0.585 0.580 0.575

0.565 0.560 0.555 0.550

CP/TBAB ratio: 5 vol% Temperature: 274.65 K Pressure: 4.0 MPa

0.545 0.540 150

160

170

180

190

S.F

S.Fr

0.570

35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15

200

Volume of TBAB solution (ml)

Figure 21 Changes of CO2 recovery (S.Fr.) and separation factor (S.F.) vs. volume of TBAB solution with CP/TBAB ratio of 5 vol% at 274.65 K and 4.0 MPa.

C O

C A P T U R E

Table 4 Experimental plans along with CO2 separation efficiencies Temperature

Pressure

MPa

150

273.65

160

CP/TBAB ratio (vol%)

Volume of solution (ml)

Run

TBAB

S.Fr.

S.F.

3.0

0.49

17.3

274.65

3.5

0.50

19.1

170

275.65

4.0

0.51

22.4

180

276.65

4.5

0.52

24.2

200

277.65

2.5

0.49

19.3

150

277.65

2.5

0.54

16.9

160

276.65

3.0

0.56

18.2

170

275.65

3.5

0.57

26.5

180

274.65

4.0

0.58

30.6

200

273.65

4.5

0.55

30.2

150

276.65

4.5

0.47

16.1

160

275.65

2.5

0.46

17.9

170

274.65

3.0

0.53

25.4

180

273.65

3.5

0.52

25.7

200

277.65

4.0

0.49

21.0

150

275.65

4.0

0.47

14.5

160

274.65

4.5

0.45

16.7

170

273.65

2.5

0.44

17.2

180

277.65

3.0

0.49

16.3

200

276.65

3.5

0.48

16.6

150

274.65

3.5

0.44

13.3

160

273.65

4.0

0.41

15.4

170

277.65

4.5

0.42

12.1

180

276.65

2.5

0.43

12.9

200

275.65

3.0

0.44

14.9

C O

C A P T U R E

In addition, the optimum results for the S.Fr. and S.F. in the work are compared with those determined by Kumar et al. [16] at the similar condition. In the work of Kumar et al. [16], they are 0.47 and 27.84, respectively, after the first stage separation with C3H8 as a promoter at 273.7 K and 3.8 MPa. Therefore, it is found that the separation efficiency in this work is relatively better than that given by Kumar et al. [16]. Meanwhile, the above results indicate that for the purpose of obtaining an optimum process of the hydrate-based pre-combustion CO2 capture in an IGCC plant, it is necessary to find out the one appropriate ratio of the volume of the TBAB solution to the volume of the CR. In this work, the 0.29 mol% TBAB solution of 180 ml with the CP/TBAB ratio of 5 vol% loaded in the CR of 336 ml is the optimum. According to the theory of scaling up [37], it means the volume ratio of TBAB solution/CR of approximately 0.54 may be the suitable ratio for the potential industrial application.

3-3-4- Induction time Figure 15(b) shows the gas uptake profiles for the three systems, CP/gas/H2O, TBAB/gas/H2O, CP/TBAB/gas/H2O, at 274.65 K and 4.0 MPa. The induction time for the above three systems is 15, 160 and 19 seconds, respectively. Therefore, it can be found from the above results that the addition of CP can efficiently shorten the induction time of the hydrate formation for the CP/TBAB/gas/H2O system, and remarkably enhance the gas uptake. Figure 22 gives the typical curves of the temperature and gas uptake changes for the hydrate formation over time from the 180 ml TBAB solution with the CP/TBAB ratio of 5 vol% and the 180 ml water with the CP at 274.65 K and 4.0 MPa. For the CP/TBAB/gas/H2O mixture, there may be at least 2 different hydrate (the CP hydrate and the gas hydrate involving in the CP and TBAB, at least including the gas hydrate containing the TBAB) forming at the certain experimental condition, and thus resulting in the temperature rise on account of the heat released by their individual hydrate formation. As seen from Figure 22, the temperature abrupt point (point a) for the CP/TBAB/gas/H2O system is at the 15th second. The time is in agreement with that for the CP/gas/H2O system. On the other hand, it can be also seen from Figure 22 that the abrupt point (point b) of the gas uptake curve is at the 19th second. Therefore, it can be identified from the above analysis that the CP hydrate firstly forms, and then the gas hydrate involving in the CP and TBAB forms in the process of the hydrate formation. As seen from Figures 24 and 25, the gas uptake for the hydrate formation from the CP/gas/H2O system is quite low. Thus, it can be determined that the induction time for the formation of the gas hydrate involving in the CP and TBAB is 19 seconds, as shown from the abrupt points of the gas uptake curves in Figure 15(b) and Figure 22. In addition, it is found that the temperature rises immediately after nucleation due to the hydrate formation is an exothermic process. The temperature reaches a high level and then gradually drops to the surrounding temperature because the temperature controller brings the temperature back to the set point value. The similar phenomenon is also observed in other experimental runs. 85

C O

C A P T U R E

0.18

276.0

0.16 275.9

0.14

gas uptake of CP/TBAB/gas/H2O

275.8

temperature of CP/TBAB/gas/H2O

0.10

temperature of CP/gas/H2O

0.08

275.7

0.06 275.6

0.04 0.02

Temperature (K)

Gas uptake (mol)

0.12

a 275.5

0.00 0

100

150

200

250

300

350

400

450

Time (s)

Figure 22 Changes of the temperature and gas uptake for the hydrate formation over time from 180 ml TBAB solution with CP/TBAB ratio of 5 vol% and 180 ml water with CP at 274.65 K and 4.0 MPa.

Figure 23 and Table 5 show the induction time of the hydrate formation from the systems with the fresh and memory 0.29 mol% TBAB solution with the various CP/TBAB ratios at 4.0 MPa and 274.65 K. The induction time with the memory TBAB solution is much smaller than that with the fresh TBAB solution and the phenomenon has been explained elsewhere [38, 39]. Additionally, the induction time decreases with the increase of the amount of the CP at the fixed amount of the TBAB solution. This illustrates that the addition of the CP promotes the nucleation of gas hydrate containing the TBAB and the nucleation rate increases with the amount of CP. The explanation of the reason is not too clear, and it may due to the fact that the CP hydrate crystals, which firstly form, catalyze the gas going to the TBAB solution, and furthermore enhance the degree of the gas super-saturation, resulting in the increase of the nucleation rate. It can be also seen from Figure 23 that decreasing the amount of TBAB solution leads to the decrease of the induction time at the fixed CP/TBAB ratio.

C O

C A P T U R E

150 ml TBAB Fresh 150 ml TBAB Memory 160 ml TBAB Fresh 160 ml TBAB Memory 170 ml TBAB Fresh 170 ml TBAB Memory 180 ml TBAB Fresh 180 ml TBAB Memory 200 ml TBAB Fresh 200 ml TBAB Memory

45 40

Induction time (sec)

35 30 25 20 15 10 5 0 5.0

10.0

15.0

20.0

CP/TBAB ratio (vol%)

Figure 23 Induction time change for hydrate formation from the systems with the fresh and the memory TBAB solution vs. CP/TBAB ratio with different volumes of TBAB solution at 274.65 K and 4.0 MPa.

Table 5 Experimental conditions along with measured induction time at 274.65 K and 4.0 MPa CP/TBAB ratio (vol%) 3 5 8 15 20

Sample

Pexp (MPa)

Fresh Memory Fresh Memory Fresh Memory Fresh Memory Fresh Memory

4.0 4.0 4.0 4.0 4.0 4.0 4.0 4.0 4.0 4.0

150ml TBAB 33 18 26 12 20 12 10 3 9 <1

Induction time (s) 160ml 170ml TBAB TBAB 39 42 20 22 35 37 17 18 23 25 14 16 12 13 4 6 10 10 <1 <1

180ml TBAB 45 23 38 19 26 16 13 6 11 <1

200ml TBAB 46 24 39 21 27 17 14 7 12 <1

In addition, it is interestingly noted that the induction time with memory water is even lower than 1 second when the CP/TBAB ratio is 20 vol%. It means that the gas hydrate containing the TBAB can be formed instantaneously with the CP/TBAB ratio of 20 vol%. However, it is at the cost of the increase of the CO2 concentration in the gas phase, as shown in Figure 17 and Figure 19. The comparisons of the induction time at the moderate conditions obtained from the different studies are showed in Table 4 [16, 17, 32, 40]. As given in Table 6, the induction time in this work is much smaller than those in other work [16, 17, 32, 40]. For example, compared to the previous results [16, 17, 32], the induction time from 87

C O

C A P T U R E

the system of either the fresh or memory 0.29 mol% TBAB solution with the CP/TBAB ratio of 5 vol% is reduced by at least 10 times at the temperature ranges from 273.7 K to 279.6 K and pressure ranges from 3.14 MPa to 8.5 MPa. Accordingly, it is expected that the system with the TBAB solution of 0.29 mol% as well as the CP/TBAB ratio of 5 vol% can be the one potential choice for the hydratebased CO2 separation from IGCC plants.

Table 6 Comparison of measured induction time in this work with that in other work

System

Temperature (K)

CO2/H2/H2O

273.7K

CO2/H2/C3H8/H2O Gas phase: 3.2 mol% C3H8

273.7K

CO2/H2/THF/H2O Solution phase: 1.0 mol% THF

279.6K

CO2/H2/ TBAB/H2O Solution phase: 0.29 mol% TBAB

278.15 K

CO2/H2/TBAB/CP/H2O Solution phase: 0.29 mol% TBAB CP/TBAB ratio: 5 vol%

276.65K

Induction time (min)

Ref.

8.5 8.5 7.5 7.5 4.8 4.8 3.8 3.8 3.14 3.14 4.12 4.12

5.0 3.7 7.7 5.3 11.0 5.0 31.0 28.3 25 6.6 10 3.3

[32]

Fresh

2.5

13.4

Memory

2.5

5.8

Fresh Memory Fresh Memory

4.0 4.0 3.5 3.5

0.87 0. 27 0.92 0.35

Sample Fresh Memory Fresh Memory Fresh Memory Fresh Memory Fresh Memory Fresh Memory

Driving force (MPa) 3.4 3.4 2.4 2.4 2.7 2.7 1.7 1.7 0.89 0.89 1.87 1.87

Pexp (MPa)

[16]

[17]

[40]

This work

3-3-5- Hydrate formation rate Figure 24 and Figure 25 show the changes of the gas uptakes in the processes of the hydrate formation at the different temperatures and different pressures, respectively. As shown in Figure 24, the amount of gas consumed and the gas uptake rate increase with the drop of the temperature. It is because the lower temperature leads to the higher driving force for the hydrate formation. The same behavior can be also found by Zhong et al.[21]. As seen in Figure 25, the amount of gas consumed and the gas uptake rate also increase with the pressure. This is also attributed to the fact that the higher pressure results in the higher driving force for the hydrate formation. However, it is found from Figure 24 and Figure 25 that the degree of the effect of the pressure on the amount of gas consumed and the gas uptake rate is not significant, compared with the temperature. In addition, it can be also seen from Figure 24 and Figure 25 that more than 80 % of the total amount of gas consumed are obtained within 400 seconds. For example, the 84 % of the total amount of gas consumed are obtained within 400 seconds at 274.65 K and 4.0 MPa. It means that most of the gas uptakes can be achieved within a relative short time. Figure 26 gives also a typical curve showing the 88

C O

C A P T U R E

change of the CO2 concentration in the gas phase in the process of the hydrate formation at 274.65 K and 4.0 MPa. As shown in Figure 26, the CO2 concentration decreases rapidly in the period from 0 to 115 seconds, then it decreases slowly. After 400 seconds, the decrease is quite small. It illustrates that the main process of hydrate formation occurs before 400 seconds, and this also means that the CO2 separation process can be completed within a shorter time. Furthermore, this work lays a foundation for the CO2 capture from the IGCC syngas in industry in the future.

0.35

273.65 K 274.65 K 275.65 K 276.65 K 277.65 K

0.30

Gas uptake (mol)

0.25

0.20

0.15

0.10

0.05

400 sec

0.00 0

500

1000

1500

2000

Time (s)

Figure 24 Gas uptake changes with time in 180 ml TBAB solution with CP/TBAB ratio of 5 vol% at 4.0 MPa and different temperatures.

C O

C A P T U R E

0.26

4.5 MPa 4.0 MPa 3.5 MPa 3.0 MPa 2.5 MPa

0.24 0.22

Gas uptake (mol)

0.20 0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02

400 sec

0.00 0

500

1000

1500

2000

Time (s)

Figure 25 Gas uptake changes with time in 180 ml TBAB solution with CP/TBAB ratio of 5 vol% at 275.65 K and different pressures.

CO2 concentration in gas phase (mol%)

40.0 Temperature: 274.65 K Pressure: 4.0 MPa CP/TBAB ratio: 5 vol% TBAB solution: 180 ml

35.0

30.0

25.0

20.0

15.0

10.0 0

100

200

300

400

500

600

700

800

Time (s)

Figure 26 Change of CO2 concentration in the gas phase with time at 274.65 K and 4.0 MPa.

900

1000

C O

C A P T U R E

4- A Hybrid Hydrate/Membrane Process. The above results indicate that following a one-stage hydrate formation/decomposition process for the CO2/H2 gas mixture in 0.29 mol% TBAB at 3.0 MPa, a CO2-rich gas containing approximately 97.28 mol% CO2 which accords with the emission criterion obtained, whereas the residual gas (H2-rich) contains approximately 82 mol% H2. It is worthy noted that the CO2 composition of the feed gas phase and the hydrate slurry vary from 39.2 to 97.28 mol%. It means that approximately 65% CO2 can be captured by the hydrate process. However, as the objective is to obtain a high purity H2, a further stage is required to treat the residual gas (H2-rich) stream. However, the operation pressure of hydrate formation increases with the H2 composition in the mixture gas, which means the high energy for the compression work. Thus, it is necessary to find other approach for the further separation of the residual gas. Membrane separation is believed to be one of the promising methods to separate CO2/H2 mixture gas, but currently this process is not suitable for such a huge load that comes from a commercial power generation station. A hybrid gas separation process, combining with the advantage of high selectivity (hydrate crystallization) and small size (membranes), is a worthwhile alternative, especially for IGCC power station. In the current conceptual process, the membrane separation only takes the load of the residual gas phase from hydrate process when it in conjunction with hydrate separation process is used. Also in this case, the gas mixture is available at the high pressure and with the high H2 concentration, a polymeric membrane would give better selectivity and higher flux for industrial application, which will be very efficient and economical to operate [41]. Therefore, a hybrid hydrate-membrane process is proposed for CO2 recovery from fuel gas as shown in Figure 27. The fuel gas which contains about 39.2 mol% CO2 was blown into a crystallizer to form the hydrate with the 0.29 mol% TBAB aqueous solution, and approximately 67% CO2 in the hydrate slurry was encaged. Then the residual gas containing approximately 82 mol% H2 was further treated by a membrane process to obtain 99 mol% H2 and 99 mol% CO2. While the hydrate slurry was pumped to a high temperature vessel to decompose, the decomposition gas which containing approximately 97 mol% CO2 was compressed for disposal. It is worthy noted that this flow chart only has one stage hydrate process. Moreover, the operating pressure (3.0 MPa) happens to be the outlet pressure of the fuel gas, which means no need to compress, and just to cold the mixture gas from 288.15 K to 278.15 K. Compared with the two-stages hybrid conceptual flow sheet operated at 3.8 and 3.5 MPa by Klara et al.[42] and operated at 7.5 and 3.8 MPa by Kumar et al.[16] respectively, it not only simplifies the process flow, but also remarkably reduces the equipment and operating cost. Compared with the integrated gasification combined cycle with chilled methanol CO2 recovery, whose power penalty was found to be 124.6 MW accounting for approximately 44.64% of the power output and mainly contributed to cooling the solvent and the feed gas to 216.48 K and compressing CO2 from 330.37 K and 14.70.096 MPa to 407.04 K and 0.65 MPa (which is far away from the feasible sequestration condition), our conceptual flow only cools the solution and the syngas to 278.15 K (the temperature difference 91

C O

C A P T U R E

between formation and decomposition is just 10 K) and compresses CO2 from 288.15 K and 3.0 MPa to the sequestration condition (313.15 K and 15.0 MPa), the power penalty will reduce about 20%. This process will be more competitive in application for hydrate-based capture from fuel gas and can be also applied to the separation and recovery of other target gases from gas mixtures without changing the basic concept.

H2O MEMBRANE

DRIER COMPRESS

99% CO2

99% H2

313.15 K 15.0 MPa

18.43% CO2

96.85% CO2

REACTOR2 REACTOR1

Feed Gas 39.2% CO2 313.15 K 3.0 MPa

EXHEAT1

COOLER

EXHEAT2

288.15 K 3.0 MPa

278.15 K 3.0 MPa 278.15 K 3.0 MPa TBAB Solution Hydrate Slurry

PUMP

Figure 27 Block flow diagram of a hybrid hydrate-membrane process for CO2 recovery and H2 purification from fuel gas in the presence of TBAB.

Conclusion The hydrate separation method for gas mixture is a novel gas capture technology. Based on the selective partition of the CO2 component between the hydrate phase and gaseous phase, CO2 can be captured from the flue and fuel gases. The main obstacles for hydrate-based CO2 capture are to moderate the operating condition, to accelerate the hydrate formation, to increase the gas uptake, and to improve the CO2 separation efficiency. TBAB as a promoter is effective on removing the obstacles. The TBAB of 0.29 mol% is found to be optimum for lowering the equilibrium hydrate formation pressure at the same temperature for caputuring CO2 from fuel gas. Furthermore, quite small quantities of adjuvant additives added into the TBAB of 0.29 mol% can further improve the hydrate-based CO2 capture. The addition of the 0.028 mol% DTAC makes the large amount of the gas hydrates form instantaneously at as low as 0.66 MPa and 277.15 K for capturing CO2 from the flue gas. At the condition, the induction time is shortened considerably, and CO2 is purified from 17.0 mol% to 99.4 % with the two-stage hydrate capture process. The addition of the CP enhances the CO2 capture efficiency remarkably and shortens the induction time for separating CO2 92

C O

C A P T U R E

from the fuel gas. The induction time in the system with the optimum CP/TBAB solution volume ratio of 5 vol% is shortened to 0.2 minutes at 4.0 MPa/274.65 K, the gas uptakes of more than 80 mol% are obtained within 6 ~ 7 minutes, and a 96.0 mol% CO2-rich and a 86.5 mol% H2-rich gas can be obtained from one-stage precombustion separation. The results provide the data and the supports for the developments of the processes for the CO2 captures from flue and fuel gases. In addition, a hybrid hydrate/membrane process is proposed as a conceptual process which consists of one-stage hydrate process in conjunction with a membrane separation stage to capture CO2 and H2. It is a big merit for simplifying process and reducing investment and operating cost.

C O

C A P T U R E

Reference [1] J. Happel, M.A. Hnatow, H. Meyer, The Study of Separation of Nitrogen from Methane by Hydrate Formation Using a Novel Apparatus, International Conference on Natural Gas Hydrates, 715 (1994) 412-424. [2] N.H. Duc, F. Chauvy, J.M. Herri, CO2 capture by hydrate crystallization - A potential solution for gas emission of steelmaking industry, Energ Convers Manage, 48 (2007) 1313-1322. [3] E.D. Sloan, Clathrate hydrates: Relations between micro-, meso-, and macroscopic properties, Fundamentals of Advanced Materials for Energy Conversion, (2002) 209-226. [4] T.A. Strobel, C.A. Koh, E.D. Sloan, Hydrogen storage properties of clathrate hydrate materials, Fluid Phase Equilibr, 261 (2007) 382-389. [5] X.S. Li, C.G. Xu, Z.Y. Chen, H.J. Wu, Hydrate-based pre-combustion carbon dioxide capture process in the system with tetra-n-butyl ammonium bromide solution in the presence of cyclopentane, Energy, 36 (2011) 10. [6] J. Lipowski, V. Komarov, T.V. Rodionov, Y.A. Dyadin, L.S. Aladko, The Structure of Tetrabutylammonium Bromide Hydrate (C4H9)4NBr. 2 1/3 H2O, Journal of Supramolecular Chemistry, 2 (2003) 5. [7] M. Arjmandi, A. Chapoy, B. Tohidi, Equilibrium data of hydrogen, methane, nitrogen, carbon dioxide, and natural gas in semi-clathrate hydrates of tetrabutyl ammonium bromide, J Chem Eng Data, 52 (2007) 2153-2158. [8] E.G. Hammerschmidt, Formation of gas hydrates in natural gas transmission lines, Ind Eng Chem, 26 (1934) 851-855. [9] D.F. Spencer, Methods of selectively separating CO2 from a multicomponent gaseous stream, in, US, 1997. [10] D. Aaron, C. Tsouris, Separation of CO2 from flue gas: A review, Separ Sci Technol, 40 (2005) 321-348. [11] J.S. Zhang, P. Yedlapalli, J.W. Lee, Thermodynamic analysis of hydrate-based pre-combustion capture of CO2, Chem Eng Sci, 64 (2009) 4732-4736. [12] S.P. Kang, H. Lee, C.S. Lee, W.M. Sung, Hydrate phase equilibria of the guest mixtures containing CO2, N-2 and tetrahydrofuran, Fluid Phase Equilibr, 185 (2001) 101-109. [13] P. Linga, R. Kumar, P. Englezos, The clathrate hydrate process for post and precombustion capture of carbon dioxide, J Hazard Mater, 149 (2007) 625-629. 94

C O

C A P T U R E

[14] T. Sugahara, S. Murayama, S. Hashimoto, K. Ohgaki, Phase equilibria for H2+CO2+H(2)Osystem containing gas hydrates, Fluid Phase Equilibr, 233 (2005) 190193. [15] R. Kumar, H.J. Wu, P. Englezos, Incipient hydrate phase equilibrium for gas mixtures containing hydrogen, carbon dioxide and propane, Fluid Phase Equilibr, 244 (2006) 167-171. [16] R. Kumar, P. Linga, J.A. Ripmeester, P. Englezos, Two-Stage Clathrate Hydrate/Membrane Process for Precombustion Capture of Carbon Dioxide and Hydrogen, J Environ Eng-Asce, 135 (2009) 411-417. [17] H.J. Lee, J.D. Lee, P. Linga, P. Englezos, Y.S. Kim, M.S. Lee, Y.D. Kim, Gas hydrate formation process for pre-combustion capture of carbon dioxide, Energy, 35 (2010) 2729-2733. [18] V. Natarajan, P.R. Bishnoi, N. Kalogerakis, Induction Phenomena in Gas Hydrate Nucleation, Chem Eng Sci, 49 (1994) 2075-2087. [19] J. Clint, Surfactant Aggregation, New York: Chapman and Hall, (1992). [20] D. Myers, Surfactant Science and Technology 2nd ed, (1992). [21] Y. Zhong, R.E. Rogers, Surfactant effects on gas hydrate formation, Chem Eng Sci, 55 (2000) 4175-4187. [22] K. Watanabe, S. Imai, Y.H. Mori, Surfactant effects on hydrate formation in an unstirred gas/liquid system: An experimental study using HFC-32 and sodium dodecyl sulfate, Chemical Engineering Science, 60 (2005) 4846-4857. [23] J. Deschamps, D. Dalmazzone, Dissociation enthalpies and phase equilibrium for TBAB semi-clathrate hydrates of N2, CO2, N2 + CO2 and CH4 + CO2, J Therm Anal Calorim, 98 (2009) 113-118. [24] Y.T. Seo, H. Lee, Structure and guest distribution of the mixed carbon dioxide and nitrogen hydrates as revealed by X-ray diffraction and C-13 NMR spectroscopy, Journal of Physical Chemistry B, 108 (2004) 530-534. [25] S.S. Fan, S.F. Li, J.Q. Wang, X.M. Lang, Y.H. Wang, Efficient Capture of CO2 from Simulated Flue Gas by Formation of TBAB or TBAF Semiclathrate Hydrates, Energ Fuel, 23 (2009) 4202-4208. [26] T. Daimaru, A. Yamasaki, Y. Yanagisawa, Effect of surfactant carbon chain length on hydrate formation kinetics, Journal of Petroleum Science and Engineering, 56 (2007) 89-96.

C O

C A P T U R E

[27] N. Kalogerakis, Jamaluddin, Dholabhai, Bishnoi, Effect of Surfactants on Hydrate Formation Kinetics, Society of Petroleum Engineers, 10.2118/25188-MS (1993) 25188-MS. [28] U. Karaaslan, M. Parlaktuna, Promotion effect of polymers and surfactants on hydrate formation rate, Energy & Fuels, 16 (2002) 1413-1416. [29] K. Okutani, Y. Kuwabara, Y.H. Mori, Surfactant effects on hydrate formation in an unstirred gas/liquid system: An experimental study using methane and sodium alkyl sulfates, Chemical Engineering Science, 63 (2008) 183-194. [30] R. Rogers, G. Zhang, J. Dearman, C. Woods, Investigations into surfactant/gas hydrate relationship, Journal of Petroleum Science and Engineering, 56 (2007) 82-88. [31] P. Linga, A. Adeyemo, P. Englezos, Medium-pressure clathrate hydrate/membrane hybrid process for postcombustion capture of carbon dioxide, Environmental Science & Technology, 42 (2008) 315-320. [32] P. Linga, R.N. Kumar, P. Englezos, Gas hydrate formation from hydrogen/carbon dioxide and nitrogen/carbon dioxide gas mixtures, Chemical Engineering Science, 62 (2007) 4268-4276. [33] J.S. Zhang, J.W. Lee, Inhibition Effect of Surfactants on CO2 Enclathration with Cyclopentane in an Unstirred Batch Reactor, Industrial & Engineering Chemistry Research, 48 (2009) 4703-4709. [34] Y. Kamata, H. Oyama, W. Shimada, T. Ebinuma, S. Takeya, T. Uchida, J. Nagao, H. Narita, Gas separation method using tetra-n-butyl ammonium bromide semi-clathrate hydrate, Japanese Journal of Applied Physics Part 1-Regular Papers Short Notes & Review Papers, 43 (2004) 362-365. [35] X.S. Li, C.G. Xu, Z.Y. Chen, H.J. Wu, Tetra-n-butyl ammonium bromide semiclathrate hydrate process for post-combustion capture of carbon dioxide in the presence of dodecyl trimethyl ammonium chloride, Energy, 35 (2010) 3902-3908. [36] R.D. Whitaker, An Historical Note on the Conservation of Mass, Journal of Chemical Education, 52, 10, 658-659 (1975) 1. [37] V.S.K. A.M. Rosen, Theory of scaling up and hydrohynamics modeling of industrial mass transfer equipment, Chemical Engineering Journal, 7 (1974) 13. [38] J.D. Lee, R. Susilo, P. Englezos, Kinetics of structure H gas hydrate, Energy & Fuels, 19 (2005) 1008-1015. [39] T. Uchida, T. Ebinuma, H. Narita, Observations of CO2-hydrate decomposition and reformation processes, Journal of Crystal Growth, 217 (2000) 189-200. 96

C O

C A P T U R E

[40] X.S. Li, Z.M. Xia, Z.Y. Chen, K.F. Yan, G. Li, H.J. Wu, Gas Hydrate Formation Process for Capture of Carbon Dioxide from Fuel Gas Mixture, Ind Eng Chem Res, 49 (2010) 11614-11619. [41] S.P. Kaldis, G. Skodras, G.P. Sakellaropoulos, Energy and capital cost analysis of CO2 capture in coal lGCC processes via gas separation membranes, Fuel Process Technol, 85 (2004) 337-346. [42] S.M. Klara, R.D. Srivastava, US DOE integrated collaborative technology development program for CO2 separation and capture, Environ Prog, 21 (2002) 247253.

C O

C A P T U R E

C O

C A P T U R E

Chapter

4 Capture of Carbon Dioxide by Amine Functioned Mesoporous Silicas Yao Shi a, Yamin Liu a,b a Institute of Industrial Ecology and Environment, Department of Chemical and Biological Engineering, Yuquan Campus, Zhejiang University, Hangzhou, China, 310027 b Department of Environment and Equipment, Fujian University of Technology, Fuzhou, China, 350108 Phone: +86-571-88273591; Fax: +86-571-88273693 E-mail :

shiyao@zju.edu.cn (Y. Shi) mingjing2000@126.com (Y. Liu)

n extensive attention has been attracted to the reduction of greenhouse gas emission from fossil fuel combustion due to its great impact on climate changes. Carbon dioxide capture and sequestration (CCS) technologies have been considered as an effective option to reduce CO2 emission. In this chapter, several newly developed solid amine sorbents for carbon dioxide capture were prepared by mesoporous silica modified by amines. There adsorption behaviors of carbon dioxide from simulated flue gas were investigated using an adsorption column. CO2 adsorption isotherms of KIT-6/TEPA exhibit a type Ι at various temperatures. The average isosteric heat of adsorption is 43.8 kJ·mol-1. The adsorption capacity increases with the rise of relative humidity (RH) of the simulated flue gas. The adsorption capacity is slightly inhibited by the presence of sulfur dioxide when the concentration of sulfur dioxide is lower than 300 ppm. As to other pollutant in the flue gas, the adsorption capacity is nearly uninfluenced by nitric oxide during a range of concentration up to 400 ppm. The solid amine sorbent, KIT-6(TEPA), behaves excellent performance for CO2 capture and seperation from flue gas.

Instruction During recent years, intensive research has been focused on the development of technologies for adsorption of CO2 on solid amine sorbent surfaces (1-15). These sorbents can be obtained by chemically bonding an amine to a support or by immobilizing liquid amines within the pores of a support. To develop efficient aminefunctionalized sorbents, support materials require well-developed structural properties, a high surface area and an appropriate pore volume. Mesoporous silica materials, including MCM-series, SBA-series and KIT-series, have been considered as ideal solid supports for their large and uniform pores, tunable pore sizes, high surface-area and the large number of highly dispersed active sites (the hydroxyl groups) on their surface. Using amine-modified mesoporous silica materials for CO2 adsorption has been extensively investigated (for a review see Choi et al.(16)). Recently, more researchers have begun to study the amine-loaded mesoporous materials obtained by impregnation. (1-3, 17) Compared with grafting, impregnation has the advantage of convenient preparation.

4-1- Character of amine functionalized mesoporous silicas Mesoporous silica materials, including MCM-series, SBA-series and KIT-series, have been considered as ideal solid supports. And using amine-modified mesoporous silica materials for CO2 adsorption has been extensively investigated.

4-1-1- Low angle XRD patterns for some adsorbents Fig. 1.1 provides the low angle XRD patterns for as-synthesized MCM-41(denote as M) and amine-impregnated M. The major peaks of M were almost replicated by those of the amine-impregnated M, although the intensities of peaks decreased slightly and the peaks shifted to a little higher 2Î¸ angles. These changes were possibly caused by the pore filling effect of the support channels and the amine coating on the outer surface of the support. Fig. 1.2 exhibits the XRD patterns of SBA-16 support and PEI-impregnated SBA-16. All locations of characteristic Bragg diffraction peaks of every sample are nearly identical, indicating that the mesoporous structure of KIT-6 is preserved after loading PEI. However, the intensity of the diffraction patterns of SBA-16 decreases substantially with increasing PEI loadings while pores of SBA-16 may be filled with PEI. The maximum PEI loading in the pore channel is about 40 wt.%. The XRD patterns of KIT-6 support and KIT-6 with different amounts (10-60 wt.%) of loaded TEPA are shown in Fig. 1.3. All locations of characteristic Bragg diffraction peaks of every sample are nearly identical, indicating that the mesoporous structure of 100

KIT-6 is preserved after loading TEPA. However, the he intensity of the diffraction patterns of KIT KIT-66 decreas decreases substantially bstantially with increasing TEPA loadings while pore pores of KIT-6 may be filled with TEPA. The he degree of the Bragg diffraction angle of 211 plane slightly increases increases from 0.93 for KIT KIT-66 support to 0.98 0.98ďź?1.14 1.14 for amine loaded KIT-6. The he maximum TEPA loading in the pore channel is about 50 wt wt.%.

Fig. 1.1 low angle XRD patterns for M and various amine modified M.. (M: as-synthesized synthesized MCM-41)

Fig. 1.2 Low angle X-ray X ray diffraction patterns of the SBA SBA-16 and the SP-nn (SP: PEI PEI-impregnated impregnated SBA-16; 16; n: 10, 20, 30 and 40)

101

Fig.1.3.. XRD patterns of the KIT KIT-66 and KT KT-n. n. (KT: TEPA TEPA-impregnated impregnated KIT-6; 6; n:10,20,30,40,50,60)

4-1-2 The he thermal stability of some adsorbents TGA and DTG data of M before and after impregnati impregnation on by various amines are shown in Fig. 1.4.. For M, DTG profile shows two mass loss peaks: the firs first peak with a mass loss of 2% below 100°C can be attributed to the moisture evolving, and the second peak with a mass loss of 38% near 280 °C is due to the re removal moval of the template. After impregnation by DETA, TETA or AMP, the first peaks intensities increased compared with that of M, because of the easy affinity of moisture and CO2 impurities on them. The order of the first peaks intensities is M-TETA M TETA > M M-DETA > M M-AMP AMP > M, because the base groups of amines decrease conversely conversely. After the encapsulation of the various amines, the second mass loss peaks with shoulders are observed, which is due to the simultaneous decomposition of the amines and the template. Fig. 1.5 shows the TGA mass loss curve curves of the PEI-impregnated impregnated SBA SBA-16 16 and TEPA TEPAimpregnated KIT-6. The ssignificant ignificant mass loss occurred when samples are heated from a temperature range of 333 333－523 523 K K. The he sample samples display a mass loss about 10% from 303 K to 373 K K. It can be attributed to desorption of moisture and CO2. From 373 K to 423 K,, none obviously mass loss can be found. Above the temperature of 423 K, these samples show different mass loss due to the decomposition of amines (PEI or TEPA) TEPA with different percent percentage ages loaded in the supports (SBA-16 16 or KIT KIT-6) 6). These hese results indicate that the thermal stability temperature of samples in air can be determined as 423 K.

102

Fig. 1.4 TGA and DTG data of M before and after various amine impregnation impregnation.

103

Fig. 1.5 TGA mass loss profiles of SP-n SP and KT-n

4-1-3- The he textural parameters of adsorbents The surface areas, total pore volumes and pore ore diameter of adsorbents were calculated. The results are listed in Table 11.1. The he surface area and pore volume decrease significantly wi with th increasing increas amines loadings.

104

Table 1.1 The textural parameters of the adsorbents Samples Surface area(BET) (m2/g) MCM-41 669.7 M 27.0 M-DETA 8.1 M-TETA 3.4 M-AMP 2.2 SBA-16 547 SP-10 287 SP-20 162 SP-30 78 SP-40 KIT-6 943 KT-10 452 KT-30 249 KT-50 89 KT-60 - Cannot be obtained

Pore volume (cm3/g) 0.60 0.10 0.012 0.0048 0.0029 0.57 0.39 0.31 0.14 1.0 0.62 0.44 0.14 -

Pore diameter(BJH) (nm)

4.8 4.4 4.2 3.9 6.0 5.6 5.1 4.8 -

4-2- CO2 adsorption performances of adsorbents 4-2-1- CO2 breakthrough curves of adsorbents Typical CO2 breakthrough curves of amine-impregnated M are given in Fig. 2.1. All the amine-impregnated M followed the similar adsorption pattern. Breakthrough time of CO2 are in the following order: M-AMP < M-DETA < M-TETA. Fig. 2.2 shows the breakthrough curves of 10% CO2 adsorption on KT-50 at different temperatures (293, 303, 313, 323, 333 and 353 K). It can be seen that at 333 K, the optimum temperature for CO2 adsorption capacity is achieved. The breakthrough time (the time at which C reaches 10% allowable breakthrough concentration) follows the order of 333 > 353 > 323 > 313 > 303 > 293 K.

105

Fig. 2.1 Experimental breakthrough curves of amine amine-impregnated impregnated as-synthesized as synthesized MCM MCM-41 41 at 41 mL/min (a) and 70 mL/min (b).

106

Fig. 2.2. Breakthrough curves of 10% CO2 adsorption on KT KT-50 at different temperatures

4-2-2- CO2 adsorption capacity of adsorbents The adsorption capacity of amine amine--impregnated impregnated M was calculated and the results are shown in T Table 2.1. The order of the adsorption capacities is M-TETA TETA > M M-DETA DETA > M-AMP, AMP, the highest adsorption capacity is 2.22 mmol/g for M M-TETA. Fig. 2.3 shows tthe he CO2 adsorption capacity of the PEI PEI-impregnated impregnated SBA-16 16 and TEPA-impregnated impregnated KIT KIT-6 at various temperatures. temperatures. The CO2 adsorption capacity increases ases with the increasing of loaded amine. The CO2 adsorption capacity rises with the raising temperature. The maximum adsorption capacity is 3.1mmol 3.1mmol路g-1 at 343K for KT-50 50 and KT KT-60. The he adsorption capacity of adsorbents are shown in table 2.1 2.1.

107

Fig. 2.3 CO2 adsorption capacity for SP SP-n and KT-nn at various temperatures

Table 2.1 The he CO2 adsorption capacity of the adsorbents Adsorption capacity (mmol (mmol路g-1) Adsorbents 303K 313K M-DETA M-TETA M-AMP SP-10 0.7 0.9 SP-20 0.9 1.1 SP-30 1 1.2 SP-40 1 1.3 KT-10 1 1 KT-30 1.3 1.6 KT-50 2.2 2.4 KT-60 2.2 2.5

323K 1.1 1.4 1.6 1.6 1.1 1.8 2.7 2.8

108

333K 1.87 2.22 1.14 1.3 1.5 1.7 1.8 1.6 2.1 3 3.1

343K 1.4 1.6 1.9 1.9 1.7 2.2 3.1 3.1

353K 1.2 1.4 1.7 1.8 1.5 2 2.9 2.9

4-2-3- CO2 adsorption dsorption isotherms of adsorbents The isotherms for CO2 adsorption by KT KT-50 50 at various temperatures (293, 303, 313, 323, 333 and 353 K) are presented in Fig. 2.4. Itt is apparent that the adsorption capacities ((qa) rises with increasing C0 and temperature over o a temperature range of 293 to 333 K. K Itt decreases decreases, however however, with increasing temperature from 333 to 353 K. The isosteric adsorption heat ( Qst) of KT-50 50 at 293-333 333 K for constant qa of 1.6 1.6-2.4 mmol g-1 (in 0.2 mmol g-1 increments) are given in Table 22.2. The values of Qst are negative between 293 and 333 K, implying that the entropy is reduced in the adsorption process. And this is consistent with the exothermic character of the adsorption process. The average isosteric heat of adsorption is 43.8 kJ mo mol-1.

Fig. 2.4. CO2 adsorption isotherms of KT KT-50 50 at various temperatures

Table 2.2 Isosteric Heats of 10% CO2 Adsorption on KT KT-50 qa (mmol g-1) -Qst (kJ mol-1)a 293-333K 1.6 51.5 1.8 49.8 2.0 42.9 2.2 39.2 2.4 38.8 Average 43.8

r2 0.978 0.994 0.996 0.994 0.995

a: The “-”” in front of Qst means that the adsorption process is exothermic.

109

4-2-4- Effect of Moisture, NO and SO2 The flue lue gas gases of power plant plants often contain various amounts of water, especially after wet FGD, and can become almost saturat saturated with water vapor vapor.. Fig. 2.5A depicts the relationship between qa and relative humidity (RH). Results show that qa increases by about 10% with increasing RH. However, owever, the increase of qa is less pronounced when RH is greater than 37%. The increase of qa could uld be explained by the partial production of bicarbonate during humid CO2 adsorption on KT KT-50.

Fig. 2.5 Effects of moisture, NO and SO2 on 10% CO2 adsorption on KIT KIT-6-50 at 333K

The effect of NO on qa is shown in Fig. 2.5 2.5B B.. Results show that qa barely changes when the concentration of NO rises from 0 to 400 ppm. When hen the concentration of NO increases from 0 to 400 ppm ppm, the values for qa are 2.85 mmol g-1 and 3.20 mmol g-1, respectively, for both dry and moist gas (RN: 100%).

110

The adsorption stability of KIT-6-50, in the presence of SO2, is illustrated in Fig. 2.5C. The value of qa decreases slightly as the concentration of SO2 increases from 0 to 300 ppm. However, qa drops significantly when the concentration of SO2 reaches 400 ppm. SO2 is a much stronger acid than CO2 and can probably react with TEPA on the solid adsorbent to form heat-stable salts. These salts may block accessibility to some adsorbent pores at concentrations above 300 ppm. Consequently, CO2 will be hindered from reacting with the TEPA impregnated within the adsorbent pores and the CO2 adsorption capacity of KIT-6-50 will be reduced.

4-3- Modeling of CO2 adsorption breakthrough curves The adsorption of CO2 on the amine amine-functionalized sorbents is more likely a chemisorption or gas-solid non-catalytic reaction than a simple physisorption, which involves a number of steps, such as CO2 diffusion into the surface and pores, reaction with the active sites, and formation of a product layer over the surface or pore walls. With the adsorption going on, changes in the pore structure and in the active surface area are expected. Additionally, formation of the product layer also creates penetration resistance for the transport of CO2 to the active sites of the adsorbents. Therefore, all these factors seem to result in the gradual deactivation of the adsorbents. In the deactivation model, an activity term a is introduced into the rate expression to represent the deactivation due to the decrease of the active site concentration, the textural changes and so forth. With the assumption of the pseudo steady state and neglecting of the axial dispersionterm (18, 19), the species conservation equation for CO2 concentration in the dynamic packed column and the rate equation for the activity change of the solid reactant with time are shown as following Eqs. (1) and (2): −Q

−

dC A − k0C A a = 0 dW

(1)

da = kd C An a m dt

(2)

Where Q is volumetric flow rate (cm3 min-1), CA is the CO2 concentration in the effluence, W is the adsorbent mass (g), k0 is the initial adsorption rate constant (cm3 min-1 g-1), a is the activity of the amine impregnated M, t is time (min) and kd is the deactivation rate constant (min-1). Generally, the deactivation rate is expected to depend on the first order of CO2 concentration (n=1) and also on the activity itself (m=1), so the solution of this model gives the following equation for the breakthrough curves:

111

1 − exp ( k0W / Q ) (1 − exp ( − kd t ) )   CA   exp − k t  = exp  ( d) C A0 1 − exp ( − kd t )  

(3)

Through the non-linear regression analysis of the experimental breakthrough data from the adsorption CO2 on the various amines amine-functionalized sorbents, the two parameters, k0 and kd, were obtained.

4-3-1- Analysis of CO2 adsorption on amine-impregnated M Nonlinear regression analysis was used to determine parameters of the deactivation model. The parameters are shown in Table 3.1. Comparison of the experimental results with deactivation model is given in Fig. 2.1. The prediction of the deactivation model shows good agreement with the experimental data since the correlation coefficient R2 is higher than 0.98 in all cases.

Table 3.1 Initial adsorption rate constant and deactivation rate constant of CO2 on the amine-impregnated as-synthesized MCM-41 at 41 and 70 mL/min Sample 41 mL min-1 M-DETA M-TETA M-AMP 70 mL min-1 M-DETA M-TETA M-AMP

k0 (mL min-1 g-1)

kd (min-1)

309.3 332.4 121.9

1.09 0.98 0.75

0.9856 0.9928 0.9886

312.1 328.9 122.1

0.94 0.97 0.97

0.9961 0.9915 0.9719

4-3-2- Analysis of CO2 adsorption on TEPA-impregnated KT-50 The parameters of the deactivation model were obtained by nonlinear regression analysis, and the parameters are list in Table 3.2. Comparison of the experimental results with deactivation model is given in Fig. 3.1. As seen from the results, the deactivation model tested gives rational fits as far as R2 is concerned for carbon dioxide adsorption in this study. The good fit of the experimental data with the deactivation model predictions indicates activity of adsorbent deceases significantly with the extent of adsorption. Changes in the activity of the adsorbent may be due to the changes in pore structure, in the active surface area and in the active site distribution of the adsorbent. Parameters of the deactivation model give reasonable values, in which changes are insignificant with modification in carbon dioxide concentration. It indicates that the deactivation model is an appropriate model for analysis of breakthrough curves of carbon dioxide adsorption on KT-50.

112

Table 3.2. Parameters of the deactivation model for carbon dioxide adsorption on KT KT-50 50 with different carbon dioxide inlet concentrations at 333K C0, CO2 (vol.%) k0M/Q k0 (mL min-1 g-1) kd (min-1) R2 5 23.947 1.197 0.855 0.997 10 12.053 0.603 0.997 0.996 1.845 20 0.092 1.326 0.982 Q (cm3 min-1) 13.224 0.529 0.998 80 0.821 11.913 0.596 0.997 100 0.957 7.831 0.470 0.997 120 1.13

Fig. 3.1 Comparison of deactivation model predictions with experimental results for ca carbon rbon dioxide adsorption on KT KT-50 50 ( A: different carbon dioxide inlet concentrations at 333K, B: different flow rates of inlet gas at 333K)

113

4-4- Industrial cconcep onceptual tual process aanalyses A conceived onceived schematic of multipollutant control control, including a CO2 adsorption unit for a power plant plant, is illustrate illustratedd in Fig. 4.1. If we assume that the removal efficiency of NOx and SO2 in a 600 MW coal coal-fired fired power plant burning low sulfur (<1.0%), are 85 85–90% 90% and 90–95% 95%,, respectively, the NOx and SO2 concentration concentrations in the effluent gas may be lower than 400 and 100 ppm. The he flue gas temperature temperatures range from 313 to 353 K. These conditions of flue gas meet the requirement for the CO2 adsorption process. Therefore, a CO2 adsorption unit can be directly installed install behind the FGD unit unit, without further treatments, such as cooling and extra extra-desulfurization. desulfurization. The consumption energy wass approximately 0.39 MWh ton-1CO2. The flue lue ggas conditions required for the KT-50 50 adsorption unit are show shown in Table able 4.1. For comparison comparison, the gas conditions for a liquid iquid amine absorption unit are also displayed in Table able 4.1.

Fig. 4.1. Conceived schematic of multipollutant control including CO2 adsorption unit in a power plant

Table 4.1. Comparison omparisonss of gas conditions between liquid amine absorption and KT KT-50 50 adsor adsorption

Liquid amine KIT-6-TEPA TEPA adsorption

Temperature (K) ≤313 313~353

RH (%) ≤100 ≤100

114

SO2 (ppm) ≤10 ≤100

NOx (ppm) ≤400 ≤400

Ref. (20)

Conclusion Extensive efforts have been devoted in CO2 capture and storage, especially for CO2 separation from flue gas. Adsorption is considered as one of the most promising methods to CO2 capture. And new adsorption separation technologies are actively pursued to develop a competitive alternative for the removal of CO2 from flue gas application. In this chapter, a series of amine-impregnated mesoporous silica were prepared by impregnating the new mesoporous silica with the poly-amine. Characteristics of amine-impregnated mesoporous silica and dynamic adsorption/desorption performance were investigated. The main conclusions are drawn: 1. The mesoporous structure of mesoporous silica are preserved after loading amines. Surface area, pore size and pore volume of adsorbents decrease with the increasing of amine loading. And the CO2 adsorption capacity increases with the increasing of loaded amine and the raising temperature. The maximum adsorption capacity is 3.2mmol·g-1 at 343K for KIT-6/TEPA. 2. The dynamic adsorption capacity increases with the temperature increases from 293K to 343K. However, the dynamic adsorption capacity reduces at 353K. CO2 adsorption isotherms of KIT-6/TEPA exhibit a type Ι at various temperatures. The average isosteric heat of adsorption is 43.8 kJ·mol-1. 3 The adsorption capacity is improved by the presence of water vapor in the simulated flue gas. The adsorption capacity increases from 2.9mmol·g-1 to 3.2mmol·g-1 when the relative humidity rise from 0% to 37% for KT-60.The adsorption capacity is slightly inhibited by the presence of sulfur dioxide when the concentration of sulfur dioxide ranges from 0 ppm to 300 ppm. As to other pollutant in the flue gas, the adsorption capacity is nearly uninfluenced by nitric oxide during a range of concentration up to 400 ppm. 4. The KT-60 is found to be attractive adsorbents for the CO2 capture. The deactivation model was used to simulate the CO2 breakthrough curves of the dynamic column absorber and showed an excellent representation.

115

References (1) Wei, J.; Liao L.; XiaoY.; Zhang P.; Shi Y. Capture of carbon dioxide by amineimpregnated as-synthesized MCM-41. Journal of Environmental Sciences 2010, 22(10) 1558–1563. (2) Liu, Y. M.; Ye Q.; Shen M.; Shi, J. J.; Chen, J.; Pan, H.; Shi, Y. Carbon Dioxide Capture by Functionalized Solid Amine Sorbents with Simulated Flue Gas Conditions. Environ. Sci. Technol., 2011, 45 (13), 5710–5716. (3) Liu, Y. M.; Shi, J. J.; Chen, J.; Ye, Q.; Pan, H.; Shao, Z. H.; Shi, Y. Dynamic performance of CO2 adsorption with tetraethylenepentamine-loaded KIT-6. Micropor. Mesopor. Mater. 2010, 134, 16-21. (4) Wei, J.; Shi, J.; Pan, H.; Zhao, W.; Ye, Q.; Shi, Y. Adsorption of carbon dioxide on organically functionalized SBA-16. Micropor. Mesopor. Mater. 2008, 116，394-399. (5) Wei, J.; Shi, J.; Pan, H.; Su, Q.; Zhu, J.; Shi, Y. Thermal and hydrothermal stability of amino-functionalized SBA-16 and promotion of hydrophobicity by silylation. Micropor Mesopor Mat. 2009, 117，596-602. (6) Shigemoto, N.; Yanagihara, T.; Sugiyama, S.; Hayashi, H. Bench-scale CO2 recovery from moist flue gases by various alkali carbonates supported on activated carbon. J. Chem. Eng. Jpn. 2005, 38, 711-717. (7) Dillon, E. P.; Crouse, C. A.; Barron, A. R. Synthesis, Characterization, and carbon dioxide adsorption of covalently attached polyethyleneimine-functionalized singlewall carbon nanotubes. ACS NANO. 2008, 2, 156-164. (8) Su, F.; Lu, C.; Kuo, S.; Zeng, W. Adsorption of CO2 on an amine-functionalized Y-type zeolite. Energy fuels. 2010, 24, 1441-1448. (9) Dadwhal, M.; Kim, T. W.; Sahimi, M.; Tsotsis, T. T. Study of CO2 diffusion and adsorption on calcined layered double hydroxides: the effect of particle size. Ind. Eng. Chem. Res. 2008, 47, 6150-6157. (10) Bastin, L.; Barcia, P. S.; Hurtado, E. J.; Silva, J. A.; Rodrigues, A. E.; Chen, B. A microporous metal-organic framework for separation of CO2/N2 and CO2/CH4 by fixed-bed adsorption. J. Phys. Chem. C. 2008, 112, 1575-1581. (11) Leal, O.; Bolivar, C.; Ovalles, C.; Garcia, J. J.; Espidel, Y. Reversible adsorption of carbon dioxide on amine surface-bonded silica gel. Inorg. Chim. Acta. 1995, 240, 183-189. (12) Kim, S.; Ida, J.; Guliants, V. V.; Lin, J. Y. Tailoring pore properties of MCM-48 silica for selective adsorption of CO2. J. Phys. Chem. B. 2005, 109, 6287-6293. 116

(13) Yue, M. B.; Sun, L. B.; Cao, Y.; Wang, Z. J.; Wang, Y.; Yu, Q.; Zhu, J. H. Promoting the CO2 adsorption in the amine-containing SBA-15 by hydroxyl group. Micropor. Mesopor. Mater. 2008, 114, 74-81. (14) Fadhel, B.; Hearn, M.; Chaffee, A. CO2 adsorption by pamam dendrimers: significant effect of impregnation into SBA-15. Micropor. Mesopor. Mater. 2009, 123, 140-149. (15) Lee, S.; Filburn, T. P.; Gray, M.; Park, J. W.; Song, H. J. Screening test of solid amine sorbents for CO2 capture. Ind. Eng. Chem. Res. 2008, 47, 7419-7423. (16) Choi, S.; Drese, J. H.; Jones, C. W. Adsorbent materials for carbon dioxide capture from large anthropogenic point sources. ChemSusChem. 2009, 2, 796~854. (17) Son, W. J.; Choi, J. S.; Ahn, W. S. Adsorptive removal of carbon dioxide using polyethyleneimine-loaded mesoporous silica materials. Micropor. Mesopor. Mater. 2008, 113, 31-40.. (18) Wang X P, Yu J J, Cheng J, Hao Z P, Xu Z P,. High temperature adsorption of carbon dioxide on mixed oxides derived from hydrotalcite-like compounds. Environmental Science & Technology, 2008, 42(2): 614–618. (19) Yasyerli, S.; Dogu, G.; Ar, I.; Dogu, T. Activities of Copper Oxide and Cu-V and Cu-Mo Mixed Oxides for H2S Removal in the Presence and Absence of Hydrogen and Predictions of a Deactivation Model. Ind. Eng. Chem. Res. 2001, 40，5206-5214. (20) Rao, A. B.; Rubin, E. S. A Technical, Economic, and Environmental Assessment of Amine-based CO2 Capture Technology for Power Plant Greenhouse Gas Control. Environ. Sci. Technol. 2002, 36，4467-4475.

117