Scholarly article on topic 'A Hybrid Separation Process for the Recovery of Carbon Dioxide From Flue Gases'

A Hybrid Separation Process for the Recovery of Carbon Dioxide From Flue Gases Academic research paper on "Chemical engineering"

CC BY-NC-ND
0
0
Share paper
Academic journal
Energy Procedia
OECD Field of science
Keywords
{"CO2 capture" / "flue gas" / "pressure swing adsorption" / "membrane separation" / "hybrid process."}

Abstract of research paper on Chemical engineering, author of scientific article — Krzysztof Warmuzinski, Marek Tanczyk, Manfred Jaschik, Aleksandra Janusz-Cygan

Abstract Due to the low CO2 concentrations in flue gases (below 20vol.%), the adequate recovery and purity of carbon dioxide in the product can only be achieved by using two-stage adsorptive or membrane systems. In these systems, the high recovery is commonly obtained by minimising CO2 content in the gas leaving stage 1 and recycling the CO2 that remains after stage 2 to the inlet of the installation. The hybrid technique is an obvious extension of the two -stage adsorptive or membrane process. However, there are a number of problems that have to be tackled before such a system becomes practically attractive. One of these problems is the sequence in which the two sections (adsorption and membrane separation) are incorporated into the complete installation. Preliminary studies show that, in the case of flue gas purification, the adsorptive section based on pressure swing adsorption (PSA) should come first. In the present paper results of detailed numerical simulations are presented for the arrangement proposed. A demonstration installation processing around 10 m3(STP)/h of the flue gas is briefly described.

Academic research paper on topic "A Hybrid Separation Process for the Recovery of Carbon Dioxide From Flue Gases"

Available online at www.sciencedirect.com

SciVerse ScienceDirect

Energy Procedia 37 (2013) 2154- 2163

GHGT-11

A hybrid separation process for the recovery of carbon dioxide from flue gases

Krzysztof Warmuzinski*, Marek Tanczyk, Manfred Jaschik, Aleksandra Janusz-Cygan

Institute of Chemical Engineering, Polish Academy of Sciences, ul.Baltycka 5, 44-100 Gliwice, Poland

Abstract

Due to the low CO2 concentrations in flue gases (below 20vol.%), the adequate recovery and purity of carbon dioxide in the product can only be achieved by using two-stage adsorptive or membrane systems. In these systems, the high recovery is commonly obtained by minimising CO2 content in the gas leaving stage 1 and recycling the CO2 that remains after stage 2 to the inlet of the installation. The hybrid technique is an obvious extension of the two-stage adsorptive or membrane process. However, there are a number of problems that have to be tackled before such a system becomes practically attractive. One of these problems is the sequence in which the two sections (adsorption and membrane separation) are incorporated into the complete installation. Preliminary studies show that, in the case of flue gas purification, the adsorptive section based on pressure swing adsorption (PSA) should come first. In the present paper results of detailed numerical simulations are presented for the arrangement proposed. A demonstration installation processing around 10 m3(STP)/h of the flue gas is briefly described.

© 2013 The Authors. Published by Elsevier Ltd. Selection and/or peer-review under responsibility of GHGT

Keywords: CO2 capture; flue gas; pressure swing adsorption; membrane separation; hybrid process.

1. The principle of the hybrid separation of CO2 from flue gases

Due to the low CO2 concentration in flue gas streams (below 20vol%), the adequate purity and recovery of CO2 requires the use of two-stage adsorptive or membrane separation systems [1-6]. High

* Corresponding author. Tel.: +48-32-234-6915; fax: +48-32-231-0318. E-mail address: kwarmuz@iich.gliwice.pl.

1876-6102 © 2013 The Authors. Published by Elsevier Ltd. Selection and/or peer-review under responsibility of GHGT doi: 10. 1016/j .egypro .2013.06.094

recoveries are usually achieved by minimizing CO2 content in stage 1 and recycling the CO2 remaining after stage 2 to the inlet of the first stage.

The hybrid process is a natural extension of the two-stage system. An important question is, however, what the sequence of the individual components of the whole arrangement should be. If we model the removal of CO2 as the separation of CO2/N2 mixtures [6], then both membrane and adsorptive systems will generate two gaseous streams. If we aim at 100% recovery of CO2, stage 1 of the hybrid system should yield two streams: a stream of pure nitrogen and a gas enriched in CO2. It has been demonstrated in [6] that this is indeed feasible in PSA installations, where the concentration of CO2 in the enriched gas could be as high as 80vol.%. The complete recovery of CO2 is also, theoretically, possible via membrane separation. This, however, necessitates membrane units of very high surface areas, while the CO2 content in the enriched gas does not exceed 40-50vol.% [3]. In the case of PSA, the flue gas does not require compression - the columns can be fed at a pressure close to atmospheric [1, 6-8], and the regeneration is carried out under vacuum [1, 6, 8-11]. Gas permeation process does, however, require the initial compression of the feed gas [4]. Although there is an option of using ambient pressure on the feed side of the membrane unit and applying vacuum on the permeate side [3, 12-13], membrane manufacturers tend to caution against this option. Furthermore, the membrane process leads to lower energy consumption per tonne of CO2 avoided provided the inlet CO2 concentration exceeds 20vol.% and the moderate (below 80%) recovery can be accepted [14]. Consequently, a hybrid separation installation should be composed of an upstream PSA unit and a downstream membrane separation stage.

The scheme of the demonstration installation which currently undergoes performance tests in our laboratory is shown in Fig. 1.

Fig. 1. The hybrid installation for CO2 capture from flue gases

A1 -A4 - adsorbers, C - gas composition, MB - membrane module, O1-O2 - dryers, P - pressure, P1 - blower, P2 - fan, P3 -vacuum pump, P4 - compressor, T - temperature, V - gas flow rate, W1 and W3 - electric heaters, ZB1 - buffer tank, ZB2 -enriched gas tank, ZB3 - purified gas tank, ZB4 - buffer tank

Nomenclature

Ff feed flow rate, m3(STP)/h Fp permeate flow rate, m3(STP)/h p pressure, bar

q* equilibrium concentration in the adsorbed phase, mol/kg y mole fraction in the gas phase a selectivity coefficient

Ap pressure difference between the feed side and the permeate side, bar

2. PSA separation unit

2.1. Adsorbent selection

A key problem in the design of PSA processes for the separation of CO2 from flue gases is the choice of a suitable adsorbent. Such an adsorbent should possess, on the one hand, a high sorption capacity vis-avis CO2 and, on the other, an adequate selectivity relative to the other components of the gas stream. The adsorbents that fulfil these conditions include activated carbons (AC) and zeolite molecular sieves 13X (ZMS) [7, 15-18]. Therefore, the relevant laboratory studies carried out in our Institute were focused on commercial adsorbents representing these two groups. In Fig. 2 experimental CO2 adsorption isotherms at a temperature of 20°C are shown. As can be seen, the CO2 adsorption capacities of two samples of ZMS 13X are by far higher than those for activated carbons, especially over the pressure range 0-2 bar.

$ $$ « *

ZMS 13X - sample 1 XZMS 13X - sample 2 □ AC - sample 1 AC - sample 2 AC - sample 3

p, bar

Fig. 2. Experimental adsorption isotherms of CO2 at a temperature of 20°C

To assess the selective properties of the adsorbents studied, CO2/N2 selectivity coefficients were determined based on the following definition [19]:

aco2/N2 - ,

./CO^^N2 (1)

It is assumed that the concentration of carbon dioxide in CO2/N2 mixtures is 15vol.%. As has been shown in [20], selectivities depend on pressure for nonlinear adsorption isotherms. It is indeed seen in Fig. 2 that selectivities decrease with an increase in pressure for all the adsorbents studied, with the values obtained for ZMS 13X by far higher than those for activated carbons. This tendency is the most pronounced over the range of low pressures (0-2 bar) associated with the process of CO2 separation. From the standpoint of the process in question, ZMS 13X is thus a better adsorbent and will be used in columns A1-A4 of the hybrid installation (cf. Fig. 1).

500 -400 -

2 300 -

d 200 -100 -0 -

0 1 2 3 4 5 6

p, bar

Fig. 3. CO2/N2 selectivity coefficients at 20°C

2.2. The PSA cycle

Based on our own experimental and numerical studies [6] supported by a vast body of literature data [7-11] it is assumed that the relevant PSA cycle should include the following steps: feed with the flue gas, cocurrent depressurisation, purge with the CO2-rich stream, countercurrent depressurisation, vacuum regeneration with the purified gas and countercurrent repressurisation with a fraction of the nitrogen stream (or, alternatively, cocurrent repressurisation with the feed). The term "cocurrent" refers to flow in the same direction as that during the adsorption step. The PSA cycle used in the hybrid installation is shown in Table 1.

X ZMS 13X - sample 1

XZMS 13X - sample 2 □ AC - sample 1 AC - sample 2

AC - sample 3 * * * * *

H a_JE_M_M_M_M_m_

Table 1. The PSA cycle used in the hybrid demonstration installation

Step—> 1 2 3 4 5 6 7 8 J, Column ____

1 F| Dt PÎ PÎ RI R PR

2 RJ, PRJ. Ff Dî PLî PÎ R

3 PÎ DJ RJ, PRJ, Fî Dî PÎ

4 DÎ PÎ PÎ DJ RJ, PR Fî

Ff - feed, Df - cocurrent depressurisation, Pf - purge, DJ, - countercurrent depressurisation, RJ, -regeneration, PR,[, - countercurrent repressurisation

2.3. Performance of the PSA unit

The PSA section should yield a stream of pure nitrogen and a CO2-rich gas containing almost all of the carbon dioxide fed into the hybrid installation. The concentration of CO2 in the CO2-rich stream depends on the proper selection of flow rates during purge (Pf) and vacuum regeneration (RJ,) steps. In [6] simulation results were presented concerning the effect of these flow rates on CO2 content in the enriched gas. These results clearly show that, for the PSA process analysed, it is possible to obtain the enriched gas containing over 80vol.% of CO2, with the complete recovery of carbon dioxide.

The relevant model of the PSA separation of CO2 from flue gases, presented in [6], formed the quantitative basis for extensive simulations of the process occurring in the PSA section of the demonstration installation shown in Fig. 1. The results are quite promising as, for the feed flow rates below 7.5 m3(STP)/h, the PSA unit can produce an enriched stream containing over 70vol.% of CO2, with the recovery nearing 100% (Fig. 4). The flow rate of the enriched stream was 18-20% of the feed flow

o o c o o

concentration

recovery □

- 99.995

99.985 O

Feed flow rate, m3(STP)/h

Fig. 4. Carbon dioxide purity and recovery in the PSA unit

3. Membrane separation unit

The membrane unit of the demonstration installation is based on commercially available membrane modules. The mature technology employs polymeric solution-diffusion active layers [21-23]. In our laboratory a number of modules have been tested experimentally to assess their performance characteristics in the separation of CO2/N2 and CO2/N2/O2 mixtures. In these experiments the feed parameters were chosen to comply with the predicted outlet parameters from the PSA unit of the hybrid installation.

In Fig. 5 CO2 concentration in the permeate vs transmembrane pressure difference is shown for the feed containing 70vol.% of CO2 and 30vol.% of N2, at a flow rate of 1.2 m3(STP)/h.

v o c o o

94.5 -

o 93.5 -

92.5 -

X X X X

1.5 Ap, bar

Fig. 5. CO2 concentration in the permeate vs transmembrane pressure difference. Feed flow rate: 1.2 m3(STP)/h, CO2 content in the

feed: 70%

As can be seen, CO2 content initially increases with Ap, to start decreasing upon reaching a maximum value. The same tendency could be observed for the other sets of the operating parameters. It is worth noting that the maximum value of CO2 concentration (>95vol.%) is attained for a relatively low transmembrane pressure difference of around 1.5 bar. Since the permeate is withdrawn at atmospheric pressure, this means that the enriched feed produced by the PSA unit should only be compressed to 2.5 bar.

As shown in Fig. 6, the concentration of CO2 in the retentate drops monotonically with an increase in Ap.

70 60 50 40

O > 30 -

20 -\ 10 0

1.5 Ap, bar

Fig. 6. CO2 concentration in the retentate vs transmembrane pressure difference. Feed parameters the same as in Fig. 5

In Fig. 7 the cut rate is shown. This parameter is of economical importance in commercial membrane processes, as it directly influences the product purity and yield.

0.7 0.6 0.5

"K "-0 4 i£0.3 0.2 0.1 0

1.5 Ap, bar

Fig. 7. Cut rate. Feed gas parameters: 1.2 m3(STP)/h, 70% CO2

The rat rate increases with a rise in Ap. For Ap = 1.5 bar and for the feed flow rate assumed, the cut rate is about 0.42. As, in our process, the retentate from the membrane section is recycled to the inlet to the hybrid installation, it is advantageous to maintain both the CO2 concentration in the retentate and the retentate's flow rate at the lowest level possible. It is thus worth noting that an increase in Ap to 2 bar leads to a drop in CO2 retentate concentration by about 10vol.% and an increase in the permeate flow rate by around 25%, at the expense of only slight drop in the concentration of CO2 in the permeate (see Fig.

5). In such a case, the flow rate of the stream recycled to the inlet of the hybrid installation would be only 8.9%-9.4% of the total feed flow rate.

4. Summary and conclusions

The hybrid technique is an obvious extension of the two-stage adsorptive or membrane process. However, there are a number of problems that have to be tackled before such a system becomes practically attractive. One of these problems is the sequence in which the two sections (adsorption and membrane separation) are incorporated into the complete installation. Preliminary studies clearly show that, in the case of flue gas purification, the adsorptive section based on pressure swing adsorption (PSA) should come first. In such a configuration, the PSA unit can be fed with the gas at a pressure slightly above the atmospheric, with the regeneration carried out under vacuum. Moreover, the membrane separation has a lower energy consumption per tonne of CO2 removed, provided the concentration of carbon dioxide at the inlet to the membrane unit exceeds 20vol.% (which is indeed the case in the arrangement proposed).

The other problem is the selection of a suitable adsorbent that, on the one hand, would offer a high sorption capacity for CO2 and, on the other, would safeguard a reasonable selectivity vis-a-vis the components of the flue gas relative to CO2. The initial screening revealed that the most promising candidates include activated carbons and zeolite molecular sieves 13X. Further equilibrium and kinetic measurements were thus focused upon these two adsorbent types leading, among others, to the reliable values of selectivity coefficients over a range of pressures and temperatures.

Finally, based on simulations and literature data a PSA cycle was designed that yields favourable values of CO2 purity and recovery without excessive complexity which might compromise the operating flexibility of the whole installation. The PSA cycle proposed includes feed with the flue gas, cocurrent depressurisation, purge with the CO2-rich stream, countercurrent depressurisation, vacuum regeneration with the purified feed and, finally, cocurrent pressurisation with the feed gas (or, alternatively, countercurrent pressurisation with the purified stream).

The membrane section of the hybrid installation is based on commercially available modules. Simulations and experimental studies show that the most suitable modules are those using polymeric membranes with solution-diffusion mechanism of mass transport. Extensive separation experiments done in our laboratory provided a host of useful data for a range of commercially available membranes. The feed parameters at the inlet to the test module corresponded to those predicted at the outlet from the PSA section. The systems studied were CO2/N2 and CO2/N2/O2 mixtures, while the main operating parameters varied in the permeation experiments were transmembrane pressure difference, initial CO2 concentration and the feed flow rate. One of the principal conclusions is that the concentration of CO2 in the permeate reaches its maximum for relatively modest transmembrane pressure differences (around 1.5 bar). As to the retentate, the content of CO2 decreases monotonically with transmembrane pressure drop; conversely, the cut rate increases with a rise in this parameter. Since the retentate from the membrane section is recycled to the inlet of the hybrid system, the most favourable situation is that in which both the CO2 retentate concentration and the retentate flow rate are kept at as low a level as possible.

The approach proposed in the present study - the combination of pressure swing adsorption with membrane separation should offset the limitations of the two techniques employed separately, i.e. high energy consumption in the case of PSA and excessive capital costs of membrane separation. The configuration analysed includes a four-column PSA unit and a capillary membrane unit. Based on preliminary simulations and experiment it is found that, in the PSA section, it is possible to recover nearly

all of carbon dioxide, with a CO2 outlet concentration of 50-80vol.%. In the membrane section, this stream can be further enriched to yield a permeate stream containing 90 -99vol.% of CO2. The recycle of the retentate, with a content of 20-60vol.% of CO2, to the inlet of the hybrid system will lead to an almost complete recovery of carbon dioxide. Experiments in a demonstration installation processing around 10 m3(STP)/h of flue gas are currently underway, supported by the relevant optimisation studies.

Acknowledgements

This study was done as part of the R&D Project NR14 0113-10/2010, financed by the Polish National Centre for Research and Development (NCBR).

References

[1] Kim J-N, Park JH, Beum HT, Han S-S, Cho S-H. PSA process for recovery of carbon dioxide. Sixth International Conference on Greenhouse Gas Control Technologies, Kyoto, Japan; 2002.

[2] Ishibashi M, Ota H, Akutsu N, Umeda S, Tajika M, Izumi J, Yasutake A, Kabata T, Kageyama Y. Technology for removing carbon dioxide from power plant flue gases by physical adsorption method. En Conv Mgmt 1996;37:929-33.

[3] Zhao L, Riensche E, Menzer R, Blum L, Stolten D. A parametric study of CO2/N2 gas separation membrane processes for post-combustion capture. J Mem Sci 2008;325:28^94.

[4] Zhao L, Menzer R, Riensche E, Blum L, Stolten D. Concepts and investment cost analyses of multi-stage membrane systems used in post-combustion processes. Energy Procedia 2009;1:269-78.

[5] Park J-H, Beum H-T, Kim J-N, Cho S-H. Numerical analysis on the power consumption of the PSA process for recovering CO2 from flue gas. IndEng Chem Res 2002;41:4122-31.

[6] Tanczyk M, Warmuzinski K, Jaschik M, Wojdyla A, Gielzak K. Separation of carbon dioxide from flue gases by pressure swing adsorption. ChemProcEng 2010;31:359-72.

[7] Chue KT, Kim JN, Yoo YJ, Cho SH, Yang RT. Comparison of activated carbon and zeolite 13X for CO2 recovery from flue gas by pressure swing adsorption. Ind Eng Chem Res 1995;34:591-98.

[8] Chou C, Chen C. Carbon dioxide recovery by vacuum swing adsorption. Sep Pur Tech 2004;39:51-65.

[9] Na B-K, Ko K-K, Eum H-M, Lee H, Song HK. CO2 recovery from flue gas by PSA process using activated carbon. Korean J Chem Eng 2001;18:220-27.

[10] Kikkinides ES, Yang RT, Cho SH. Concentration and recovery of CO2 from flue gas by pressure swing adsorption. Ind Eng Chem Res 1993;32:2714-20.

[11] Chang D, Min J, Moon K, Park Y-K, Jeon J-K, Ihm S-K. Robust numerical simulation of pressure swing adsorption process with strong adsorbate CO2. Chem Eng Sci 2004;59:2715-25.

[12] Ho MT, Allinson GW, Wiley DE. Reducing the cost of CO2 capture from flue gases using membrane technology. Ind Eng Chem Res 2008;47:1562-68.

[13] Favre E. Carbon dioxide recovery from post-combustion processes: Can gas permeation membranes compete with absorption? J Mem Sci 2007;294:5^9.

[14] Bounaceur R, Lape N, Roizard D, Vallieres C, Favre E. Membrane processes for post-combustion carbon dioxide capture: A parametric study. Energy 2006;31:222^34.

[15] Metz B, Davidson O, de Coninck H, Loos M, Meyer L. (Eds.) IPCC Special Report. Carbon Dioxide Capture and Storage. Cambridge: Cambridge University Press; 2005.

[16] Siriwardane RV, Shen M-S, Fisher EP, Poston JA. Adsorption of CO2 on molecular sieves and activated carbon. Energy and Fuels 2001;15:279-84.

[17] Hoffman JS, Fauth DJ, Pennline HW. Development of novel dry regenerable sorbents for CO2 capture. 19th Annual International Pittsburgh Coal Conference, Pittsburgh, USA; 2002.

[18] Tanczyk M, Jaschik M, Warmuzinski K, Janusz-Cygan A, Jaschik J. Separation properties of adsorbents for the removal of CO2 from flue gas streams. Inzynieria i Aparatura Chemiczna 2010;49:82-3 (in Polish).

[19] Ruthven DM. Principles of Adsorption and Adsorption Processes. New York: John Wiley; 1994.

[20] Warmuzinski K, Sodzawiczny W. Effect of adsorption pressure on methane purity during PSA separations of CH4/N2 mixtures. Chem Eng Proc 1999;38:55-60.

[21] Powell CE, Qiao GG. Polymeric CO2/N2 gas separation membranes for the capture of carbon dioxide from power plant flue gases. J Mem Sci 2006;279:1-49.

[22] Scholes CA, Kentish SE, Stevens GW. Carbon dioxide separation through polymeric membrane systems for flue gas applications. RecPat Chem Eng 2008;1:52-66.

[23] Bernardo P, Drioli E, Golemme G. Membrane Gas Separation: A Review/State of the Art. Ind Eng Chem Res 2009;48:4638-63.