Intensification of low temperature thermomorphic biphasic amine solvent regeneration for CO2 capture

chemical engineering research and design 9 0 ( 2 0 1 2 ) 743–749

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Research and Design journal homepage: www.elsevier.com/locate/cherd

Intensification of low temperature thermomorphic biphasic amine solvent regeneration for CO2 capture Jiafei Zhang ∗ , Yu Qiao, David W. Agar Institute of Reaction Engineering (TCB), Department of Biochemical and Chemical Engineering, Technische Universität Dortmund, Emil-Figge-Straße 66, D-44227 Dortmund, Germany

a b s t r a c t High-energy requirements for solvent regeneration represent one of the main challenges in the conventional postcombustion capture (PCC) process. Thermomorphic biphasic solvent (TBS), comprising lipophilic amines as the active components, exhibit a liquid–liquid phase separation (LLPS) upon heating, giving rise to extractive behaviour, and thus enhancing desorption at temperatures well below the solvent boiling point. The low regeneration temperature of less than 90 ◦ C together with the high cyclic CO2 loading capacity, 3–4 mol/kg, of such TBS system permits the use of low temperature and even waste heat for desorption purposes. In order to improve the solvent regeneration process and reduce the commensurate energy demand still further, desorption experiments with various techniques for enhancing CO2 release in place of gas stripping, such as nucleation, agitation, ultrasonic method, etc., were studied at temperatures in the range of 75–85 ◦ C. Nucleation and agitation both accelerate CO2 desorption, but regenerability by nucleation only achieves 70–85%, while by agitation attains 80–95%. Ultrasonic desorption also intensifies the solvent regeneration and superior to conventional stripping process. The energy consumption for TBS system with those intensification techniques is only half of that for alkanolamine-based process with steam stripping. Extractive regeneration is another potential method to substitute for stripping and reduce the exergy demands. An extraction process using inert solvent was developed for improving the regeneration efficiency and elevated pressures were applied for reducing the significant volatile solvent loss. © 2012 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. Keywords: CO2 capture; Amine absorbent; Solvent regeneration; Process intensification; Hybrid technology

1.

Introduction

Solvent regeneration with its high energy consumption at temperatures of 120–150 ◦ C represents one of the most significant weaknesses in alkanolamine-based CO2 capture, contributing to more than half of the total processing costs. Development of new solvents and the integration of process heat are the most promising means of reducing the energy consumption (Wang et al., 2011). Many researchers have devoted a lot of work to identify new absorbents and techniques for reducing the energy requirement and thus increasing the capture efficiency (Chowdhury et al., 2009, 2011) have screened various aqueous tertiary amine solvents to mitigate the energy requirements by lowering the heat of absorption, with only modest deterioration in CO2 loading capacity and reactivity. Some undisclosed solvents, which can

∗

allegedly cut the regeneration energy by ≈30% compared to the benchmark monoethanolamine (MEA), have been investigated in the CASTOR project (Knudsen et al., 2007) and also by other researchers (Puxty et al., 2009; Goto et al., 2009; Kim et al., 2011), but the desorption still needs to be carried out with steam stripping at 120 ◦ C or above. A novel TBS system has been proposed by Agar et al. (2008) to ameliorate this problem, since they offer excellent reactivities and high loading capacities but also very deep regenerabilities at temperatures of only 80–90 ◦ C, due to the partially miscible properties and thermally induced LLPS behaviour (Zhang et al., 2011a). Initial development work focussed on the miscibility of organic and aqueous phases and their temperature dependent phase transition behaviour. The concept of the phase transition CO2 capture process is primarily determined by two processes:

Corresponding author. Tel.: +49 231 755 2582. E-mail address: [email protected] (J. Zhang). Received 12 July 2011; Received in revised form 5 February 2012; Accepted 27 March 2012 0263-8762/$ – see front matter © 2012 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cherd.2012.03.016

744

chemical engineering research and design 9 0 ( 2 0 1 2 ) 743–749

Lean Solvent

40 °C

Treated Gas CO2

Org. 80 °C

Flue Gas

Aq.

Absorber

Rich Solvent

Regenerator

Fig. 1 – Basic process flow diagram of TBS system. • The homogeneous loaded lipophilic amine solution separates into two phases upon heating at temperature of around 80 ◦ C during regeneration; the organic phase formed mainly contains the regenerated lipophilic amine, while the aqueous phase comprises water, carbamate, bicarbonate and protonated amine species. • The regenerated lean biphasic solvent reverts to a single phase upon cooling to the lower critical solution temperature (LCST) of ca. 40 ◦ C, i.e., the absorbent becomes homogeneous again at the operating temperature of absorber. Such a biphasic system potentially offers considerable advantages (Svendsen et al., 2011). Since the absorbent is resolved into one phase highly concentrated in CO2 and another phase low in CO2 , only the concentrated phase need be sent to the stripper. This is equivalent to a system operated with an extremely high CO2 capacity. In addition, the concentrated phase can be loaded up to a very high level, enabling desorber operation at elevated pressure, reduced temperature or both. Furthermore, there is no deterioration of wetting in the absorber operation, even when a very concentrated solution is circulated to the absorber. This could facilitate the use of absorber with a greater cross-sectional area and thus a lower pressure drop. Recognising the advantages of integrating LLPS in PCC, a new CO2 capture technology – the DMXTM process – has been developed by IFP Energies nouvelles to reduce energy consumption with a demixing unit prior to thermal regeneration: the lean organic phase is returned to the absorber and while the CO2 rich aqueous phase is sent to a steam stripping column (Raynal et al., 2011a). The activating solvent used in such a process has a LCST of 50 ◦ C, however, the LLPS temperature is higher than 90 ◦ C, being elevated by the solubilisation effect of the ionised CO2 . In the TBS process on the other hand, the LLPS unit is designed to achieve “deep” solvent regeneration (>90%) with simultaneous CO2 release (see Fig. 1). By removing the regenerated CO2 from the LLPS unit, the required LLPS temperature of the partially protonated amine solution is reduced and equilibrium is driven in the direction of desorption. Therefore, a subsequent stripping process is not required for such regenerable TBS system. In order to accelerate the CO2 desorption at 80 ◦ C, it is essential to explore alternative technologies for enhancing the

low temperature regeneration process in place of steam stripping. Intensified methods such as agitation, ultrasound and nucleation have been employed in this work. Speed of agitation was found to have influence on regeneration (Zhang et al., 2011b), but the fundamental mechanisms are still requested to be disclosed. Its application for TBS system is expected to be more suitable, economic and easier to achieve deep regeneration. No publication has indicated nucleation can be used for CO2 regeneration, the performance should be evaluated and principles must be studied. Ultrasonic desorption was proven to be a promising method, since it significantly reduces the regeneration temperature and solvent degradation compared to conventional stripping process (Gantert and Möller, 2012). In addition, an extractive regeneration with an additional hydrophobic inert solvent, for instance, pentane, hexane or cyclohexane, was studied to reduce the operating temperature for CO2 desorption still further (Zhang et al., 2011c). However, significant volatile inert solvent loss is a major shortcoming and operating parameters must be optimised for minimising such loss.

2.

Experiment

The loaded amine standard solutions for regeneration experiments with intensified methods such as extraction, agitation, nucleation, ultrasound, etc., were prepared in a 500 mL glass bubble column with c. 320 mL aqueous amine solutions and a gas comprising 15 mol% CO2 balanced with N2 at 30 ◦ C. The feed gas was saturated with water vapour to prevent losses in the bubble column and the flow rate was regulated by mass flow controllers so as to be constant in the absorption and desorption tests. A chilled-water condenser was employed to minimise the vaporisation loss of volatile lipophilic amines. The detailed experimental setup for absorption is found in the previous publication (Zhang et al., 2011b). Lipophilic amines with concentrations of 3–5 M were employed over contact times of 3–5 h, to ensure that equilibrium was achieved. Regeneration was carried out by nitrogen gas stripping or other intensive means in a 100 mL vessel with 50 mL loaded solution at 70–85 ◦ C. For the intensification techniques, nitrogen was only used as a carrier gas and as a reference for on-line gas chromatography (GC, Agilent HP6890 with a GS-GasPro capillary column 30.0 m × 320 ␮m at an oven temperature

745

chemical engineering research and design 9 0 ( 2 0 1 2 ) 743–749

N

N NH

NH

DPA (II)

DsBA (II)

H2N

HA (I)

DMCA (III)

EPD (III)

Fig. 2 – Examples of lipophilic amines comprising hydrophobic alkyl group(s) and hydrophilic amino group. 35 ◦ C) monitoring of the outlet gas after desorption. Agitation was carried out by a hotplate stirrer (RCT basic IKAMAG safety control) with magnetic stirring bar, laid at the bottom of the cylindrical glass reactor to agitate the loaded amine solution for accelerating the CO2 desorption. Nucleation was performed also in the cylindrical glass reactor with different porous materials and various sizes, for instance, silica beads (0.2–0.5 mm, from Merck) with diameter of pore openings ˚ Al2 O3 spheres (≈3 mm, from Degussa), active (DPO) 60 A, carbon spheres (≈2.5 mm and ≈5 mm, from Merck), zeolite chips (≈4 mm, from Merck), polytetrafluoroethylene (PTFE) boiling stone (≈4 mm, from Bola), ceramic Raschig rings (5 mm), ceramic Berl-Saddle (5 mm), SYLOBEAD molecular sieves (grade 562c, 3.2 mm and grade 564, 1.6 mm) with DPO ˚ cotton fibre and wood fibre. In ultrasonic desorption, an 3 A, Economic Ultrasonic Cleaner (Elmasonic E) with constant wave frequency 37 kHz was employed and the electronic energy consumption was measured by an ammeter. After the ab-/desorption had been completed, the CO2 loading was ascertained by the barium chloride method, total amine concentration was determined by acid–base back-titration and the blended amine compositions were determined by GC analysis (CP-Volamine capillary column 60.0 m × 320 ␮m with programmed oven temperatures from 80 to 240 ◦ C) of the liquid samples. The inert solvent regeneration was carried out in a 150 mL double-wall glass reactor and the temperature of the extraction was controlled using an external thermostat. During the extraction process, the solution was agitated by an externally driven magnetic stirrer operated at constant speed. 50 mL loaded amine solutions were extracted with the same volume of inert solvent for 2–4 h. The extractor was connected to a reflux condenser operated at 15 ◦ C to minimise volatility losses of the inert solvent. The components in the organic and aqueous phases were determined by analysis of the liquid phase using GC analysis of liquid samples.

to 10–60 ◦ C, since the aqueous soluble species carbamate, carbonate and bicarbonate play the role of solubilisers in dissolving all the non-protonated amine in water, which significantly improves the technical feasibility by providing more degrees of freedom for regulating the phase transition behaviour in the absorption process. Based on the results from the screening tests, lipophilic amines may be classified into two categories: an absorption activator, with rapid reaction kinetics, and a regeneration promoter, to enhance regenerability. Examples of each class are given in Table 1. DPA was first selected as an alternative absorbent, but unfavourable precipitation of bicarbonate and protonated DPA species was observed at high concentrations (≥3 M). The newly developed secondary lipophilic amine “A1” has exhibited an outstanding performance for CO2 capture, for instance, 40% faster absorption rate and 120% higher net CO2 loading capacity (see Fig. 3) than the values for the conventional 30 wt% MEA solvent. According to the phase change tests, several of the selected alternative lipophilic amine absorbents studied, such as HA, DPA and A1 exhibit critical solution behaviour around 40 ◦ C, but the required LLPS temperature is too high (≈90 ◦ C) to permit utilisation of waste heat in regeneration. However, for other less soluble lipophilic amines, such as DMCA and DsBA, the LLPS can be reduced to temperatures of lower than 80 ◦ C, but the critical solution temperature is even lower – below 20 ◦ C, which makes it possible to exploit low value heat for regeneration but renders a homogeneous solution for absorption infeasible. Blended lipophilic amine solutions were therefore used in the TBS system to enable a compromise overcoming these drawbacks. Temperature-induced LLPS is a distinctive and very beneficial phenomenon in the regeneration process using lipophilic

500 400

Results and discussion

3.1.

Lipophilic amine for CO2 absorption

The lipophilic amines and water are not completely miscible in all proportions at certain temperatures, since lipophilic amine is a hybrid molecule with hydrophilic and hydrophobic functional groups (see Fig. 2). Due to the restricted miscibility, the characteristics of LCST behaviour were found in mixtures of DPA and water (Davison, 1968; Zhang, 2007). The aqueous solubility of lipophilic amine decreases with increasing temperature and thus exhibits a phase separation upon heating. For example, the LCST of the selected aqueous fresh amine solutions is between −10 and 30 ◦ C at concentrations of 30–60 wt%, while in the regenerated lean solutions it rises

PCO2 (mbar)

300

3.

200

B1=DsBA 100 90 80 70

A1, 5M, 40°C A1+B1, 4+1M, 30°C A1+B1, 4+1M, 80°C DMX-1, 40°C MEA 30wt%, 40°C MEA 30wt%, 120°C

60 50 40

0

1

2

3

4

Loading (mol-CO2 /kg-sol.)

Fig. 3 – Vapour–liquid equilibrium of blended amine solutions.

5

746

chemical engineering research and design 9 0 ( 2 0 1 2 ) 743–749

Table 1 – LCST and LLPS temperatures of lipophilic amine solutions. Amine

Type

LCST (◦ C)

LLPS Temp. of 3 M solution (◦ C)

Regenerability at 80 ◦ C (%)

Absorption activator

Hexylamine Cycloheptylamine Dipropylamine A1

I I II II

20 ≈15 −5 10

90 90 90 90

≈40 ≈45 ≈50 ≈50

Regeneration promoter

Di-sec-butylamine (B1) N,N-dimethylcyclohexylamine N-ethylpiperidine

II III III

−20 −15 10

60 70 80

>95 >90 >80

amine absorbents. It was observed for some of the primary (I) and secondary (II) amines as well as most tertiary (III) amines in the screening experiments conducted at 60–90 ◦ C (see Table 1). In the DMXTM process proposed by IFP, the CO2 lean phase is split off from the rich phase by using LLPS and the lean phase recycled directly to the absorber to save energy in the stripping step (Raynal et al., 2011b). However, in the TBS process, we have developed the regeneration technology still further to the point where steam stripping becomes superfluous. Due to the excellent performance characteristics of the lipophilic amines selected, deep regeneration was achieved without steam stripping that is to say only by LLPS and the extractive regeneration behaviour of blended lipophilic amine system. As can be seen from Table 1, the temperature difference between the LCST and LLPS is at least 70 ◦ C, which means that steam stripping is still required for the absorption activator, in order to attain deep regeneration, and the regeneration promoter remains as a two phase system in the absorber feed. Therefore, an important objective of solvent formulation is to reconcile these differences. Nevertheless, regeneration of A1 still required temperatures higher than 90 ◦ C. Blending a less water soluble lipophilic amine, e.g. DMCA or EPD, as a regeneration promoter can ameliorate this problem. In the blended system, for example, 2 M DMCA + 1 M A1 solution, deep CO2 regeneration (>90%) was attained at only 80 ◦ C, but the regenerated solvent must still be cooled to <20 ◦ C to obtain a homogeneous solution. Conversely, in the 3 M A1 + 1 M DMCA solution, a single phase solvent was obtained at 40 ◦ C, but less than 70% regenerability was achieved at 80 ◦ C and the rate of CO2 release was very slow if no advanced method was adopted. Therefore, techniques for intensifying solvent regeneration were investigated.

surface roughness, fluid properties and operating conditions (Maruyama and Kimura, 2000; Thome, 2010). Various porous materials were used to enhance nucleate bubble formation in the desorption experiments. Since they provide extensive nucleation sites, CO2 regeneration can hence be intensified. Significantly rapid CO2 desorption was observed by using particles such as zeolite chips, Al2 O3 spheres and PTFE boiling stones. Those porous materials with cavities on the surface, which are poorly wetted by the liquid, have the greatest tendency to entrap gases and show the positive influence on bubble nucleation according to classical nucleation theory (Cole, 1974). The effective surface energy is lower at such preferential sites, thus diminishing the free energy barrier and facilitating bubble nucleation. However, amine degradation was found with ceramic materials and Sylobeads, decolourisation was detected with active carbons and gel formation was observed with fibres and silica beads. In nucleate desorption, the CO2 concentration in the loaded solution is quite high after absorption. By raising the temperature, the equilibrium of dissolved CO2 in the aqueous phase is displaced towards the dissociation of the carbamate and bicarbonate species and the solution attains a supersaturated state. The additional porous particles thus provide active nucleation sites for reducing the free energy barrier of bubble nucleation and initiating the formation of “excess” CO2 bubbles. Thus, nucleation promotes CO2 releasing from the loaded solutions. However, the particles only influence the desorption rate but not the reaction equilibrium, since they are inert in terms of the absorption. Fig. 4 shows the good experimental results achieved by using porous particles, Al2 O3 spheres, at a level of only 1.25–2.5 wt% in the solution, to accelerate CO2 evolution from the loaded solution. By combining agitation and

2,0

Nucleation

Lipophilic amine solvent regeneration with thermomorphic LLPS has been shown to be an effective means for cutting the energy demands and reducing the quality of heat source needed for absorbent regeneration. However, the slow desorption rate becomes a challenge at lower temperatures if steam stripping is not used. In order to accelerate the release of CO2 from loaded solutions, measures for intensifying solvent regeneration without gas stripping were studied. Nucleation has been studied as a method for accelerating CO2 regeneration in loaded solutions. Nucleation usually occurs at nucleation sites on surfaces contacting the liquid or gas. Nucleation sites are generally provided by suspended particles or minute bubbles, so-called heterogeneous nucleation, which occurs much more common than homogeneous nucleation that takes place without preferential nucleation sites. Bubbles form at nucleation sites depending on the

Solution: A1+B1, 2+1M at 75 °C 1,6

Desorbed CO2 (mol/L)

3.2.

1,2

1/40 wt. 1/60 wt. 1/80 wt. N2 stripping

0,8

w/o any method 0,4

0,0 0

10

20

30

40

Time (min) Fig. 4 – Solvent regeneration by nucleation with Al2 O3 spheres.

747

chemical engineering research and design 9 0 ( 2 0 1 2 ) 743–749

4

2,0

Solution: A1+B1, 2+1M at 75 °C

Nucleation with PTFE Agitation at 800 rpm

1,6

1,2

750 rpm 500 rpm 250 rpm N2 stripping

0,8

w/o any method

Desorbed CO2 (mol/L)

Desorbed CO2 (mol/L)

3

2

1

0,4

0,0 0

10

20

30

0

40

Time (min)

CA+A 4M DM

) ) ) ) .5) 1 (1:1 1 (2:1 1 (1:2 1 (1:3 1 (1:3 CA+A CA+A CA+A CA+A 4M DM 4M DM 4M DM 4M DM

Solvent

Fig. 5 – Influence of agitation speed on CO2 evolution.

3.3.

Agitation

Bubbles formation can be observed in a stirred vessel and is typically influence by speed of agitation and surface tension of solvent. The low surface tension promotes the bubble breakage and reformation, leading to more bubbles regenerated with smaller sizes (Laakkonen et al., 2005). Agitation also breaks the surface tension of the solution to increase the cavitation level, bubbles are thus more likely to be formed. Agitated regeneration was hence applied in a continuous stirred tank reactor (CSTR) as a substitute for steam stripping. Fig. 5 illustrates that the desorption rate with agitation is comparable to that for gas stripping in the blended A1 + B1 solution and that the agitation speed also has a positive influence on CO2 desorption, with more CO2 being desorbed at higher agitation rates. Because cavitation is formed by agitation and it is enhanced by increase of agitation speed. By immediate implosion of cavities in the liquid solution, CO2 is liberated from liquid phase to gas phase, reaction is therefore driven towards CO2 desorption. The higher speed of agitation creates more cavitation bubbles and thus intensifies CO2 releasing. In addition, agitation also promotes the mass transfer in the reactor and accelerates the reaction towards CO2 releasing. According to the calculation method described by Geuzebroek et al. (2009), Nienow (2011) and Oexmann and Kather (2010), the total energy consumption in agitated desorption for TBS system is estimated to be only less than 2.0 MJ/kg-CO2 , which mainly includes 1.2 MJ/kg-CO2 of reaction enthalpy, 0.14 MJ/kg-CO2 of agitation energy, 0.4 MJ/kg-CO2 of sensible heat and 0.2 MJ/kg-CO2 of heat loss. Compared to the MEA-based “state-of-the-art” technology with steam stripping (4.0 MJ/kg-CO2 ), it significantly limits the consumption of latent heat and saves half of the required desorption energy. A comparison study between nucleation and agitation indicates that solvent regeneration has been intensified more deeply by agitation (see Fig. 6), since a more effective mechanical force is involved for cavities formation. Nucleation is more suitable for solvents with more regeneration promoter, since they are more easily to be regenerated; while agitation is preferred for solvents with more activators, due to their higher CO2 loading capacity.

Fig. 6 – Comparison of nucleation and agitation for solvent regeneration at 80 ◦ C.

3.4.

Ultrasonic regeneration

Ultrasonic desorption has also been investigated for enhancing the solvent regeneration, using high frequency sound waves to agitate the aqueous solution and encourage bubble formation, since the compression waves in the liquid tears the liquid apart, leaving behind many millions of microscopic “voids” or “partial vacuum bubbles”–cavitations (Reidenbach, 1994). The collapse of these microscopic cavities facilitates CO2 bubble formation. The positive influence of ultrasound on the CO2 desorption rate is illustrated in Fig. 7. Although the rich solvent, for example A1 + DMCA, cannot be regenerated by nucleation or ultrasound as deeply as with gas stripping – typically 5–20% less – the regeneration rate is comparable or even faster than for N2 stripping. Since the frequency of ultrasonic wave was found to have minor influence on desorption (Gantert and Möller, 2012), a constant frequency was adopted in this study. The total energy consumption in ultrasonic desorption for TBS system is also around 0.2 MJ/kg-CO2 , where only 0.15 MJ/kg-CO2 of ultrasonic energy is consumed.

2,0

Desorbed CO2 (mol/L)

nucleation, the regeneration rate has been enhanced synergistically by 10–20%, compared to employing the measures individually.

1,5

Ultrasound Al2O3 spheres

1,0

PTFE chips w/o any method

0,5

Solution: 3M A1+DMCA (3:1) at 80 °C 0,0 0

10

20

30

40

50

60

Time (min)

Fig. 7 – Intensification of CO2 release with various methods.

748

3.5.

chemical engineering research and design 9 0 ( 2 0 1 2 ) 743–749

Distillation

Extractive regeneration with immiscible amine

Inert sol.

The novel solvent system introduced in this paper was investigated as a means of avoiding high exergy demands by low temperature solvent regeneration through thermomorphic LLPS and extractive regeneration behaviour. Lipophilic amine A1 with excellent performance with respect to reaction kinetics and loading capacity was chosen as an alternative absorbent in place of MEA. However, the temperature required for deep regeneration is higher than 90 ◦ C, which makes it difficult to use waste heat from the energy recovery network. An effective regeneration promoter is thus required to reduce the regeneration temperature and enhance CO2 release. Based on the comparison of regenerability in the screening experiments, DsBA, a water-insoluble lipophilic amine and also regarded as extracting agent, yielded the best performance and became the preferred component for regeneration promotion. The LLPS was observed at temperatures lower than 70 ◦ C in such a solvent. The promoted A1 (p-A1) solution exhibits biphasic behaviour at the desired concentrations and temperatures. By addition of a regeneration promoter such as DsBA, the LLPS temperature of loaded A1-based solution has been shifted from 90 ◦ C down to 75 ◦ C and the regenerability is improved from 85% to 97%, which permits the use of low temperature or even waste heat for regeneration purposes and cuts the energy consumption. But the reaction kinetics of pA1 in CO2 absorption is depressed by DsBA, the proportion of DsBA and A1 is thus limited to not higher than 1:2.

3.6.

Lipo. amine Org. phase Homo. TBS

Heating

(lipo. amine) LLPS

Aq. Phase Org. phase

(water + ions)

(inert (inert ++ amine) amine)

Absorption

Pre-regeneration

Aq. Phase (water)

Extraction

Fig. 8 – Concept of LLSP with inert solvent extractive regeneration. 50% of the amines were recovered and an organic lean phase was formed. The organic phase was recycled to the absorber. The CO2 is highly concentrated in the aqueous rich phase. As is shown in Fig. 8, this was passed to the extraction unit with an inert solvent. Compared to the previous extraction method (Zhang et al., 2011c), the developed process reduces both inert solvent requirement and energy consumption in desorption. But vaporisation loss of inert solvent was found to be a significant drawback. In order to reduce the volatile losses of the inert solvent, the system pressure was increased to up to 3 bar at 50–70 ◦ C. Due to the different miscibility properties of the solvent system, lipophilic amines were preferably dissolved in the organic phase, the residual amine was thus extracted by the inert solvent with simultaneous CO2 release. The desorbed CO2 from pre-regenerator and extractor can be compressed for storage; the lean aqueous phase from the extractor was recycled to absorber and organic phase containing inert solvent and lipophilic amines was sent to a low temperature distillation unit. Since the boiling point of the inert solvent is lower than 70 ◦ C and for lipophilic amine is higher than 130 ◦ C, the thermal separation is very easy. Following separation, the recovered lipophilic amines can be recycled to absorber and the inert solvent can be recycled to the extraction unit. One of the key tasks in this extractive regeneration is the selection of appropriate inert solvent. Pentane with an excellent partition coefficient was initially employed, but the significant volatile losses represent a major weakness. Other organic solvents such as hexane, cyclohexane, iso-hexane were further evaluated. Using less volatile hexane can effectively reduce the vaporisation loss, but the partition coefficient is inferior. This shortcoming can be overcome by increasing the extraction temperature and the number of stages. Since the carbon capture technology is being developed to mitigate global climate change, a negative environmental impact of the CO2 capture process should be avoided. Therefore, a high

Extractive regeneration with inert solvent

For achieving fast absorption kinetics, lipophilic activator is preferred to be comprised as the main component in most of the TBS solutions, but they are only partially regenerated without steam stripping at 80 ◦ C. Therefore, a regeneration technique with inert solvent has been developed to improve performance. An extraction unit with inert solvent such as pentane, hexane or cyclohexane was employed for achieving an extensive regeneration at temperatures well below 70 ◦ C, giving more degrees of freedom for integrating waste heat sources from industrial energy recovery networks in the solvent regeneration process. However, large amount of inert solvent involved in the regeneration process may cause the problems such as negative environmental impact and additional heat demand. The extraction process and operating parameters must be optimised. In the system examined, an aqueous A1 + DMCA solution (4 M, 3:1) was used as the activated solvent. A high proportion of A1 is present in order to increase the absorption rate and capacity; a small amount of DMCA is incorporated to decrease the LLPS temperature to 80 ◦ C. The loaded solution was separated into two phases in the pre-regeneration stage. More than

Table 2 – Performance of inert solvent for extractive regeneration of loaded 4 M DMCA + A1 (3:1) solution. Inert solvent

Pentane Hexane Cyclohexane Iso-hexane a

Atmospheric pressure

Elevated pressure (≈3 bar)

Partition coefficient (mol/mol)

Volatile loss (%)

Required stagesa

3.2 3.5 3.6 3.0

45 4 2 10

4.5 4.3 4.2 4.6

T

(◦ C) 40 60 60 60

Partition coefficient (mol/mol)

Volatile loss (%)

Required stagesa

T (◦ C)

2.1 1.8 1.9 –

20 <1 <1 –

6.0 7.0 6.6 –

60 70 70 –

Calculated by the Kremser formula (Müller et al., 2005) for counter-current extraction with 96% amine recovery and Vinert = Vamine .

chemical engineering research and design 9 0 ( 2 0 1 2 ) 743–749

boiling point hydrophobic solvent scrubbing unit is employed to remove the low boiling point inert extracting agent vapour. Table 2 demonstrates the practicability of using inert solvent extraction for CO2 regeneration.

4.

Conclusions

CO2 capture using TBS system potentially permits an extensive regenerability of >90% without gas stripping at 75–85 ◦ C, when other process intensification measures are adopted. Agitation, nucleation and ultrasonic methods are the most promising techniques for accelerating CO2 desorption in place of steam stripping in the low temperature regeneration process. The required temperature and energy consumption for TBS system with those intensified means are much lower than conventional alkanolamine-based technologies, thus leading to low solvent degradability and permitting a more flexible and expedient thermal integration of CO2 regeneration system into the low-value heat recovery network. Extractive regeneration using inert hydrophobic solvents can reduce the desorption temperature down to 40–70 ◦ C and intensifies absorbent recovery, releasing more than 90% of CO2 by multiple-stage extraction. This measure provides further degrees of freedom for reducing the exergy demand and thus improving the technical feasibility of using waste heat for CO2 desorption. Elevation of extraction pressure and use of less volatile inert solvent reduce vaporisation loss but depress the partition coefficient. However, this drawback can be overcome by increasing the extraction temperature and the number of stages. Those intensified TBS processes offer significant advantages for improving capture efficiency and reducing process costs, due to the rapid reaction kinetics, high loading capacity, low regeneration temperature, excellent regenerability and low energy consumption.

Acknowledgement This work was supported by Shell Global Solutions International B.V.

References Agar, D.W., Tan, Y., Zhang, X., 2008. CO2 removal processes by means of absorption using thermomorphic biphasic aqueous amine solutions, Patent WO/2008/015217. Chowdhury, F.A., Okabe, H., Shimizu, S., Onoda, M., Fujioka, Y., 2009. Development of novel tertiary amine absorbents for CO2 capture. Energy Procedia 1, 1241–1248. Chowdhury, F.A., Okabe, H., Yamada, H., Onoda, M., Fujioka, Y., 2011. Synthesis and selection of hindered new amine absorbents for CO2 capture. Energy Procedia 4, 201–208. Cole, R., 1974. Boiling nucleation. Adv. Heat Transfer 10, 85–166. Davison, R.R., 1968. Vapor-liquid equilibria of water-diisopropylamine and water-di-n-propylamine. J. Chem. Eng. Data 13 (3), 348–351. Gantert, S., Möller, D., 2012. Ultrasonic desorption of CO2 – a new technology to save energy and prevent solvent degradation. Chem. Eng. Technol. 35 (3), 576–578.

749

Geuzebroek, F., Schneider, A., Last, T., Zhang, X., 2009. Benchmark study on the energy consumption of novel solvents for CO2 capture. In: 5th Trondheim CCS Conference, Trondheim, Norway, June. Goto, K., Okabea, H., Shimizua, S., Onodaa, M., Fujioka, Y., 2009. Evaluation method of novel absorbents for CO2 capture. Energy Procedia 1, 1083–1089. Kim, J.H., Lee, J.H., Lee, I.Y., Jang, K.R., Shim, J.G., 2011. Performance evaluation of newly developed absorbents for CO2 capture. Energy Procedia 4, 81–84. Knudsen, J.N., Vilhelmsen, P.J., Jensen, J.N., Biede, O., 2007. First year operating experience with a 1 t/h CO2 absorption pilot plant at Esbjerg coal-fired power plant. In: Proceedings of the 6th European Congress of Chemical Engineering, Copenhagen, 16–20 September. Laakkonen, M., Moilanen, P., Aittamaa, J., 2005. Local bubble size distributions in agitated vessels. Chem. Eng. J. 106, 133–143. Maruyama, S., Kimura, T., 2000. A molecular dynamics simulation of a bubble nucleation on solid surface. Int. J. Heat Technol. 18, 69–74. Müller, E., Berger, R., Blass, E., et al., 2005. Liquid–Liquid Extraction, Ullmann’s Encyclopedia of Industrial Chemistry, 7th ed. Wiley-VCH, Weinheim. Nienow, A.W., 2011. Stirred Tank Reactors, Ullmann’s Encyclopedia of Industrial Chemistry, 7th ed. Wiley-VCH, Weinheim. Oexmann, J., Kather, A., 2010. Minimising the regeneration heat duty of post-combustion CO2 capture by wet chemical absorption: The misguided focus on low heat of absorption solvents. Int. J. Greenhouse Gas Control 4 (1), 36–43. Puxty, G., Allport, A., Attalla, M., 2009. Vapour liquid equilibria data for a range of new carbon dioxide absorbents. Energy Procedia 1, 941–947. Raynal, L., Bouillon, P., Gomez, A., Broutin, P., 2011a. From MEA to demixing solvents and future steps, a roadmap for lowering the cost of post-combustion carbon capture. Chem. Eng. J. 171 (3), 742–752. Raynal, L., Alix, P., Bouilon, P., et al., 2011b. The DMXTM process: an original solution for lowering the cost of post-combustion carbon capture. Energy Procedia 4, 779–786. Reidenbach, F., 1994. ASM Handbook, Vol. 5, Surface Engineering, 10th ed. ASM International, Materials Park, OH. Svendsen, H.F., Hessen, E.T., Mejdell, T., 2011. Carbon dioxide capture by absorption; challenges and possibilities. Chem. Eng. J. 171 (3), 718–724. Thome, J.R., 2010. Wolverine Engineering Data Book III, Chapter 9. Boiling Heat Transfer on External Surface, http://www.wlv.com/products/databook/db3/DataBookIII.pdf (accessed February 2012). Wang, M., Lawal, A., Stephenson, P., Sidders, J., Ramshaw, C., 2011. Post-combustion CO2 capture with chemical absorption: a state-of-the-art review. Chem. Eng. Res. Des. 89 (9), 1609–1624. Zhang, J., Agar, D.W., Zhang, X., Geuzebroek, F., 2011a. CO2 absorption in biphasic solvents with enhanced low temperature solvent regeneration. Energy Procedia 4, 67–74. Zhang, J., Nwani, O., Tan, Y., Agar, D.W., 2011b. Carbon dioxide absorption into biphasic amine solvent with solvent loss reduction. Chem. Eng. Res. Des. 89 (8), 1190–1196. Zhang, J., Misch, R., Tan, Y., Agar, D.W., 2011c. Novel biphasic amine solvent for CO2 absorption and low temperature extractive regeneration. Chem. Eng. Technol. 34 (9), 1481–1489. Zhang, X., 2007. Studies on multiphase CO2 capture systems. PhD dissertation, VDI Verlag, Düsseldorf.