N2 gas separation membrane processes for post-combustion capture

Journal of Membrane Science 325 (2008) 284–294

Contents lists available at ScienceDirect

Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci

A parametric study of CO2 /N2 gas separation membrane processes for post-combustion capture Li Zhao ∗ , Ernst Riensche, Reinhard Menzer, Ludger Blum, Detlef Stolten Institute of Energy Research (IEF-3), Forschungszentrum Jülich, D-52425 Jülich, Germany

a r t i c l e

i n f o

Article history: Received 9 January 2008 Received in revised form 11 July 2008 Accepted 23 July 2008 Available online 20 August 2008 Keywords: CO2 Post-combustion Gas separation membrane Permeability Selectivity

a b s t r a c t Capture of CO2 from flue gases produced by the combustion of fossil fuels and biomass in air is referred to as post-combustion capture. Chemisorbent processes are considered to be the most feasible method and are already at an advanced stage of development, but gas separation membranes are attracting more and more attention as a possible alternative. This paper describes a detailed parametric study of mass and energy balances for a simulated single membrane process. Typical operating conditions (CO2 concentration in the flue gas, pressure and temperature, etc.) together with the influence of the membrane quality (permeability, selectivity) and membrane area on membrane performance (CO2 separation degree and CO2 purity) are simulated over a wide range of parameters. © 2008 Elsevier B.V. All rights reserved.

1. Introduction Carbon dioxide (CO2 ) capture and storage (CCS) is a process consisting of the separation of CO2 from industrial and energy-related sources, transport to the storage location and long-term isolation from the atmosphere [1,2]. There are several different types of CO2 capture systems: post-combustion, pre-combustion, oxyfuel combustion and other industrial separation techniques [3]. The so-called post-combustion process captures CO2 from flue gases produced by the combustion of fossil fuels and biomass [2]. Power plants emit more than one-third of all CO2 emissions worldwide. Because the CO2 concentrations in the flue gas are low – typically, 3–5 mol% in gas plants and 13–15 mol% in coal plants – the cost of capture would be significant [1,4]. Separation processes such as absorption, membranes, adsorption and cryogenic fractionation are listed as potential candidates for CO2 capture in the postcombustion process. Furthermore, the ammonia scrubbing process based on chemical absorption with a monoethanolamine (MEA) solvent plays a dominant role [5], at least in the near future. Many companies are currently involved in building full-scale demonstration power plants with MEA absorption process, e.g. RWE 500 MWe coal-fired power plant (PCC) in Tilbury, UK; Statoil/Shell 860 MWe natural gas power plant (NGCC) in Tjeldbergodden, Norway; Statoil/Dong 280 MWe natural gas power plant (NGCC) in Mongstad,

∗ Corresponding author. Tel.: +49 2461 614064; fax: +49 2461 616695. E-mail address: [email protected] (L. Zhao). 0376-7388/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2008.07.058

Norway [6,7]. MEA was developed more than 60 years ago as a general, nonselective solvent to remove acid gases, such as CO2 and H2 S, from natural gas streams [8]. However, the chemical processes associated with the degradation of the MEA are still not completely understood, which leads to increased material costs, waste disposal costs, and energy demands for the CO2 capture process [8]. Of course, in order to extend the solvent life the high demands made by the SOx (<10 ppm) and NO2 (<20 ppm) [9] content in the flue gas leads to extraordinarily high expenditure for the pre-capture process. Furthermore, almost 50% of the low-pressure steam from the intermediate/low-pressure (IP/LP) steam turbine should be employed for regeneration of the solvent used to release CO2 , which requires a considerable amount of thermal energy generated by the power plant [5]. A life cycle assessment (LCA) of MEA-based carbon capture shows that the reduction of CO2 emissions using this method is achieved at the expense of increasing other emissions and corresponding environment impacts and calls for a further investigation of membrane separation processes [10]. All above this is the motivation for investigating the most competitive alternative technique – gas separation membranes. In view of the application field for gas separation membranes, especially for CO2 flood EOR (enhanced oil recovery) and purification in pipeline natural gas projects, it is well known that since the early 1980s polymer gas separation membranes have achieved developmental and commercial success in comparison to conventional amine and cryogenic fractionation [11,12]. Several features distinguish the gas separation membranes used in postcombustion at power plants from the aforementioned application

L. Zhao et al. / Journal of Membrane Science 325 (2008) 284–294

285

Fig. 1. Schematic diagram of CO2 membrane position in a post-combustion flue gas line. The Arabic numerals 1, 2 and 3 show the proposed test positions for different membranes in the EnBW power plant.

fields: one is the different gas separation goal, which is CO2 /N2 , instead of CO2 /CH4 , H2 /CH4 ; the second is the low CO2 concentration in the flue gas – about 14 mol% (the gas recovered from the gas field by EOR obviously has a higher CO2 concentration); and the third is the pressure of the flue gas, which is in the range of ambient pressure rather than being pressurized. A feasibility study for the polymer gas separation membranes used in post-combustion by Van der Sluijs [13] claimed that gas separation membranes with a selectivity of at least 200 are required to make them a serious competitor in comparison to other separation techniques. This conclusion is still widely accepted and has been cited by many other researchers [14,15]. Gas separation membranes have recently been rejected again in the early stages of a process selection project [2,5]. CO2 /N2 gas separation membranes have developed remarkably in the past 15 years [16–18], especially polymer membranes; a PAMAM dendrimer membrane with a selectivity of 230 and 400 has been developed by RITE [19,20]. In addition, systems analysis and design help us to understand and arrange the membrane process effectively. Although several groups, e.g. Favre (French), Hägg (Norwegian) and Hendriks (Dutch), have done a great deal of work in this field [21–25], comprehensive simulation work on operating conditions (feed gas CO2 concentration, pressure and temperature, etc.) together with the influence of the membrane quality (permeability, selectivity) and membrane area on the membrane performance (CO2 separation degree and CO2 purity) has seldom been involved in a coal-fired post-combustion capture process. It is now absolutely necessary to investigate the separation process of the membranes for post-combustion so that a more competitive CO2 capture method can be developed. As part of the METPORE project [26], the first membrane test worldwide concerning mechanical stability in a real flue gas environment used for post-combustion CO2 capture was performed in a power plant of the EnBW company, Germany, in April 2008. The experiment was carried out both for a cermet-type membrane and a polymer membrane. A publication is being prepared about the experimental equipment and results. The present work is concerned with a detailed parametric study of mass and energy balance simulation for a single-stage CO2 /N2 membrane process.

2. A single-stage CO2 /N2 membrane process The flue gas line for a conventional pulverized coal combustion power plant is shown in Fig. 1. On the basis of the working conditions of coal-fired power plants several different type of membranes (e.g. sol–gel derived stainless-steel-based cermet membrane developed by IEF-1 Forschungszentrum Jülich, Germany; PEBAX polymer membrane (working temperature lower than 70 ◦ C) developed by GKSS, Germany) were proposed for testing in three different positions in the EnBW power plant. Position 1, analogous to amine stripping processes [5] after the deNOx (SCR), dust removal (E-filter) and desulphurization (FGD) processes, was chosen for the first test. This position has been widely accepted to be most suitable for the polymer membrane. The decision to apply the cermet-type membrane at the other positions was based on the fact that this membrane can work in a harsh environment. According to the purity requirements of CO2 transport (>95 mol% CO2 purity) [27,28] and re-injection, a purification process should be included (CO2 purity more than 99 vol%) after capture [29]. The flue gas information given in Table 1 was obtained from the industrial partners. A single-stage membrane process is shown in Fig. 2. All relevant parameters are listed here. Operating conditions (pressure, temperature, CO2 concentration and flow rate of the feed gas) act on a certain membrane (selectivity, permeability and area), then the performance of the membrane (CO2 purity and degree of CO2 separation) can be predicted. The degree of CO2 separation is also described as the CO2 recovery ratio in the literature [13,21–23]. Table 1 Flue gas data after FGD in power plants corresponding to data from the literature [30] Data

Unit

Power plant 1

Power plant 2

Coal Efficiency Power output Flue gas flow rate CO2 N2 Temperature

– % MWe m3 s−1 mol% mol% ◦C

Hard coal 45 1000 833 13.6 70.7 50

Hard coal 44 1000 850 14.8 70 80

286

L. Zhao et al. / Journal of Membrane Science 325 (2008) 284–294

Fig. 2. All relevant parameters of a single-stage membrane process.

Fig. 3. Asymmetric hollow-fiber membrane operating in the co-flow mode with feed flow outside the fiber [31].

To readily understand gas separation membranes in postcombustion processes our first simulation for the membrane process was simplified to a single-stage membrane with a binary feed gas system – CO2 and N2 . 3. Description of the simulation method and the reference case 3.1. Simulation method The membrane module in the PRO/II software (Simulation Sciences Inc.), which is composed of high-flux asymmetric gas separation membranes [31], was used for a series of parametric studies. There are different thermodynamic models for the energy balance calculation by PRO/II; the Soave–Redlich–Kwong equation of state was chosen for our case. A mathematical simulation for the permeation of the asymmetric membrane is based on the following assumptions: • The feed gas is on the skin side of the asymmetric membrane. • No mixing of permeate fluxes of different compositions occurs inside the porous supporting layer of the membrane (plug flow). • The porous supporting layer has negligible resistance to gas flow, and diffusion along the pore path is insignificant due to high permeate flux. • The membrane permeability is independent of pressure and concentration. • Feed gas pressure drop is negligible. • The permeate flow inside the fiber and the spiral-wound leaf is governed by the Hagen–Poiseuille equation.

• In the case of hollow-fiber membrane, the deformation of the fiber under pressure is assumed to be negligible. Fig. 3 illustrates a single asymmetric hollow fiber operating in the co-current mode with feed flow outside the fiber. Here it is important to note that the concentration of the permeate leaving the membrane skin surface y is generally different from that of the bulk permeate stream outside the porous layer y, except at the closed end of the fiber where they are identical. The mathematical formulation set up in terms of co-flow, counter-flow and cross-flow pattern can be found in Pan’s original paper [31] Table 2 Input and output parameters of the membrane module in PRO/II Input Membrane

• Permeability (CO2 , N2 ) • Area A

Operating conditions

Feed • Temperature T • Pressure pfeed • Flow rate V˙ feed • Gas mol fraction x Permeate • Pressure ppermeate

Membrane performance

Output

• CO2 partial pressure difference pCO2

Permeate (CO2 , N2 ) • Flow rate V˙ permeate • Gas mol fraction yP Retentate (CO2 , N2 ) • Flow rate V˙ retentate •

Gas mol fraction yR

L. Zhao et al. / Journal of Membrane Science 325 (2008) 284–294

287

Fig. 4. Schematic illustration of the reference case.

and some other common tools for membrane gas separation processes, e.g. Geankoplis’s book [32]. Two significant conclusions from Pan’s work are presented here: (1) for the asymmetric membrane the net effect of the permeate pressure build-up and the feed flow direction is that the feed–permeate flow pattern has little effect on the membrane performance and that the counterflow pattern is not necessarily always the preferred operating mode; (2) the mathematical model presented for calculating the performance of permeators with the high-flux asymmetric membrane has been verified by the field pilot-plant experiments for helium recovery from natural gas using large hollow-fiber modules. The model is applicable to both hollow-fiber and spiral-wound modules. Input and output parameters of the membrane module in PRO/II are listed in Table 2. On the basis of the membrane simulation platform offered by PRO/II, here the definitions of several parameters are introduced in order to understand the CO2 /N2 gas separation membrane process of our case. CO2 /N2 selectivity ˛ = process selectivity ˛∗ =

PCO2 PN2 yP,CO2 /yP,N2 xCO2 /xN2

(1)

(2)

pCO2 = pretentate,CO2 − ppermeate,CO2 = pretentate × yR,CO2 −ppermeate × yP,CO2 =

ppermeate pfeed

(4)

CO2 purity = yP,CO2 (CO2 mol fraction at permeate side) CO2 separation degree =

(3)

V˙ permeate,CO2 V˙ feed × xCO2

(5)

be CO2 permeability, membrane area or permeate side pressure. Through the definition of the controller, this parameter will be solved exactly. Here, this solution line is termed “Method 2”. This method is of advantage if the membrane performance (output data) has to be defined before the calculation, e.g., if the degree of separation of the membrane remains constant at 50%, through using the controller to define this parameter, then choosing one of the input parameters – “membrane area” – as a variable which can be worked out exactly under a certain operating condition. 3.2. Reference case A so-called cardo polyimide membrane developed by RITE [34] with CO2 /N2 selectivity of 40 and CO2 permeance (thickness/permeability = 1/permeance, here “thickness” points to the thickness of the membrane layer) of 7.5 × 10−9 m3 m−2 s−1 Pa−1 = 2.7 m3 m−2 h−1 bar−1 is taken as a benchmark for CO2 /N2 gas separation owing to the quite high CO2 permeability [35]. The CO2 permeance was fixed at 1.0 m3 m−2 h−1 bar−1 instead of 2.7 for the reference case before the parametric studies, which could be checked roughly and easily by mental arithmetic. For the reference case the basic membrane arrangement, typical operating conditions and simulation results are shown in Fig. 4. As previously mentioned, the driving force for CO2 permeation is the partial pressure difference between the feed and the permeate side. Fig. 5 shows the partial pressure distribution within a membrane process with a selectivity of 40 and a permeate pressure of 30 mbar. It can be seen how difficult it is under these given operating conditions (feed pressure: 1 bar, permeate pressure: 30 mbar, CO2 in feed 14 mol%) to separate CO2 from N2 completely. The driving force for CO2 decreases gradually along the membrane length (assuming that the membrane area is large enough), whereas the

(6)

One aspect to be noted here is the parameter of process selectivity ˛*, which is a “ratio of ratios”, and in the literature is also termed the “separation coefficient” [33] or “real selectivity” [22]. The term “process selectivity” indicates that this parameter strongly depends on the operating parameters of the process. It is already well known that the gas permeation driving force for the membrane process is the partial pressure between the permeate side and the feed side. According to Table 2, after defining a certain membrane and its operating conditions, CO2 purity and degree of CO2 separation for a membrane process can be calculated precisely by PRO/II; here this logical line is termed “Method 1”. Simultaneously, the partial pressure difference (driving force) can be solved as an output parameter by Eq. (3). In PRO/II there is a component “controller”. If one of the output data (see Table 2), e.g. CO2 partial pressure difference, is chosen as a controlled parameter, then this output parameter can be taken as an input parameter in the calculation, which leads to it becoming unnecessary to define one of the original input parameters exactly. This parameter could

Fig. 5. Partial pressure distribution along the membrane length.

288

L. Zhao et al. / Journal of Membrane Science 325 (2008) 284–294

Table 3 A plan for the parametric study and related remarks Parameters (m3

m−2

pCO2 ˛ A (m2 )  xCO2 (mol%)

h−1

bar−1 )

Reference case

Variation

Remarks

1 40 28 0.03 14

0.05, 0.1, 0.5, 2, 3 10, 20, 25, 30, 50, 55, 70, 80, 100, 200 1, 11, 21, 31, . . .. . ., 991 0.001, 0.01, 0.05, 0.1, 0.2 10, 20

Guide for membrane development

driving force for N2 is quite large at the beginning of the process and continues to increase with the permeation of CO2 . A study plan for the parameter variation and the corresponding remarks is given in Table 3. Based on “Method 1” and “Method 2”, two or three input parameters can be varied simultaneously. The influence of membrane quality and operating conditions on the membrane performance can be investigated comprehensively. 4. Parameter study results and discussion 4.1. Influence of permeability and selectivity At first, only one parameter was varied – CO2 permeability or CO2 /N2 selectivity, the others remained constant as in the reference case. Fig. 6a and b shows the results. It can be observed from the diagrams that at a given operating condition and with a certain membrane area the CO2 separation degree is increased with increasing membrane permeability pCO2 and CO2 purity is fully controlled by the membrane selectivity ˛. The degree of CO2 separation and CO2 purity reflects the influence of membrane quality on process performance. Of course, for a post-combustion process, a high degree of CO2 separation and CO2 purity are desirable for the membrane process. The target region is located at the top right-hand corner of the diagrams according to the challenge from the competing technology of MEA absorption. According to our above simulation results it can be concluded that from the reference case to the target corner both high selectivity and high permeability should be required. The current membranes applicable for gas separation in postcombustion processes can be mainly divided into inorganic ceramic membranes and organic polymeric membranes. So-called hybrid membranes are composed of inorganic molecular sieves and polymers (mixed matrix membranes) [36] or, vice versa, microporous polymer layers supported on porous ceramic substrates [37,38].

Related to investment cost Influence on energy consumption Basis for developing a multi-stage process

For inorganic ceramic membranes for CO2 separation from flue gas, the US Department of Energy (DOE) requires a CO2 permeance >3 × 10−7 mol m−2 s−1 Pa−1 (2.42 m3 m−2 h−1 bar−1 ) and a CO2 /N2 selectivity >100 [39]. This is supported by NEDO (New Energy and Industrial Technology Development Organization) in Japan. The primary goals of the NEDO group are to develop an inorganic ceramic membrane with the following properties: CO2 /N2 selectivity of 100 and permeance of CO2 at 3.4 × 10−7 mol m−2 s−1 Pa−1 (2.74 m3 m−2 h−1 bar−1 ) at 350 ◦ C, also with a durability of 500 h at 350 ◦ C [40]. The CO2 separation membrane could be either a single or composite structure made of e.g. silica, zeolite, zirconia, titania or alumina [30,40–42]. The state-of-the-art ceramic membrane has quite good CO2 permeance >2 × 10−7 mol m−2 s−1 Pa−1 (1.61 m3 m−2 h−1 bar−1 ) [41], but the selectivity is not above 10 [39,41]. With regard to polymeric membranes, Powell and Qiao [18] made a detailed review of different polymer structures developed for flue gas separation. The original discussion of the “Robeson plot” has become one of the most highly cited papers in the gas membrane separation literature [43]. A comparison to calculation results from Freeman [16] and a non-exhaustive collection of experimental selectivity/permeability data for CO2 /N2 separation cited from [23] are shown in Fig. 7. As mentioned at the beginning of this paper, the state-of-the-art polymeric membrane was developed by RITE [19,20], and is made of PAMAM dendrimer with quite high CO2 /N2 selectivity of 400, but quite low CO2 permeability 1.6 × 10−7 m3 m−2 s−1 kPa−1 (0.0576 m3 m−2 h−1 bar−1 ). Comparing the two previously mentioned types of membranes, it is known that ceramic membranes have quite a good permeability, but quite a low selectivity and can operated under harsh conditions: working temperatures higher than 350 ◦ C. Normally the working temperature of the polymeric membranes is below 200 ◦ C, though a membrane that can withstand high temperatures (above 300 ◦ C) has been developed by Los Alamos National Laboratory [17].

Fig. 6. (a) Influence of CO2 permeability PCO2 on membrane performance, (b) influence of CO2 /N2 selectivity ˛ on membrane performance.

L. Zhao et al. / Journal of Membrane Science 325 (2008) 284–294

289

Table 4 Membrane area requirement for power plant 1 (1000 MWe ) with RITE membranes under the operating condition of 30 mbar at the permeate side

Fig. 7. Robeson plot, trade-off curves for glassy (dotted line) and rubbery (bold dotted line) polymers are taken as the upper bound for polymer membrane separation ability; points represent a collection of experimental selectivity/permeability data for CO2 /N2 mixtures [23].

According to our calculation results (Fig. 6a and b), the membrane permeability and selectivity are both important for membrane performance. And according to the trade-off between these two parameters, it is not easy to reach the target (high degree of CO2 separation and high CO2 purity) with a one single-stage ceramic or polymeric membrane. A carefully designed multi-stage hybrid (ceramic + polymeric) system may help to reach this target. 4.2. Influence of the membrane area In addition to the permeability and selectivity, the membrane area is also an important membrane parameter. A curve family combining the variation of the membrane area A and selectivity ˛ in Fig. 8 shows how these two parameters influence membrane performance. The other parameters are kept the same as those shown in Fig. 4 of the reference case.

Fig. 8. A curve family of the variation of CO2 /N2 selectivity ˛ (10–200) and membrane area A (1, 11, . . ., 991 m2 ), feed gas flow rate 100 m3 h−1 , CO2 14 mol%, permeate vacuum 30 mbar.

˛

pCO2 (m3 m−2 h−1 bar−1 )

Separation degree (%)

CO2 purity (mol%)

A (×106 m2 )

230 230 400 400

0.1656 0.1656 0.0576 0.0576

50 80 50 80

95 91 97 94

17 49 49 151

It can be observed from Fig. 8 and also from Fig. 6b that the CO2 purity can be increased by increasing the membrane selectivity. The membrane area plays an important role in membrane performance. Increasing the area contributes to a higher degree of separation; but the CO2 purity deteriorates, especially at the tail of each curve. So a “practicable membrane area” should be considered for each application in order to avoid undesired N2 permeation. The most attractive region for industrial application should be the upper right-hand corner (CO2 purity >90 mol%, degree of CO2 separation >80%). Then it can be interpreted from the curve family that the selectivity of the single-stage membrane should approach 200, which coincides with the work done by the Dutch group [13,25]. Table 4 lists the one-stage membrane area requirement for power plant 1 to achieve 50% or 80% separation if RITE membranes are applied under the operating conditions of the permeate vacuum of 30 mbar. One aspect to be mentioned here is the enormous membrane area required. 4.3. Process selectivity As defined by Eq. (2), the process selectivity is a parameter composed of a ratio of ratios. If the CO2 purity at the permeate side is 80 or 90 mol% and the feed gas has a CO2 concentration of 14 mol%, the process selectivity should be 25 and 55, respectively. Based on the reference example, an extremely positive case is considered: the permeate pressure is 1 mbar and the feed pressure is 1 bar. Fig. 9 shows how under this operating condition the membrane performs with a selectivity of 25 and 55, respectively. In this scenario, we can see that the membrane selectivity and the process selectivity are only equal in the case of the permeate side pressure being near a high vacuum level (1 mbar) if the degree of separation is quite low (< 1%). On the basis of this calculation, it can be deduced that under the given operating condition (reference case) the membrane selectivity should be at least larger than 25 or

Fig. 9. Membrane performance with selectivities of 25 and 55 under a permeate vacuum of 1 mbar, process selectivity being kept constant at a certain CO2 purity shown in the right axis, membrane area A varying from 1, 11, 21 to 991 m2 .

290

L. Zhao et al. / Journal of Membrane Science 325 (2008) 284–294

Table 5 Process selectivities 25 and 55 are the baselines for CO2 purity requirements of 80 and 90 mol%, respectively; operating conditions: feed flow 100 m3 h−1 , feed CO2 molar fraction 14 mol%, permeate vacuum 30 mbar CO2 purity (mol%)

Selectivity

Process selectivity

Separation degree (%)

Membrane area (m2 )

80

20 50 80

33 86 230

27 43 77

25

90

20 50 80

28 90 264

77 101 197

55

55, respectively, to obtain 80 or 90 mol% CO2 purity and the corresponding degree of separation. From Table 5 it can be seen clearly that the deviation between the membrane selectivity and the process selectivity increases with increasing membrane selectivity and higher CO2 purity requirements.

Fig. 10. Influence of the level of the permeate vacuum on the membrane performance, ˛ = 40, membrane area A = 1, 11, 21, . . ., 991 m2 .

4.4. Influence of the pressure ratio The pressure ratio is a very important parameter for the operating conditions. On the basis of the reference case, the varying influence of the permeate vacuum on the membrane performance is illustrated in Fig. 10. A high degree of permeate vacuum contributes to a high CO2 purity and to a high degree of separation. The driving force is generated by applying a pressure difference across the membrane, which is usually achieved by maintaining a vacuum at the permeate side or a compression at the feed side. Both measures can also be applied in parallel. As described in the third section of this paper defining a controller in PRO/II, a so-called “Method 2” can be used to make a further calculation. Here the parameter pCO2 (at the end of membrane separation process) is controlled. For the reference case its value is 0.09 bar. If the area is chosen as the extra parameter, then the correlation among several parameters (selectivity, membrane area and pressure ratio) can be manifested in parallel in two diagrams. Figs. 11 and 12 show the basic interconnection of these parameters (feed gas compression and permeate vacuum pumping) separately in an extremely wide range (theoretical consideration). Considering the reference case again, pCO2 is calculated as 0.09 bar by Method 1; then using Method 2, CO2 purity (82.6 mol%)

and separation degree (20.5%) can be calculated in exactly the same way as in Method 1 shown in Fig. 11a. Simultaneously, the required area of 28 m2 is calculated and denoted in Fig. 11b. Using these diagrams, it is possible to make some comparisons between feed side compression (FC) and permeate side vacuum pumping (PVP). For PVP at a certain pressure ratio and with decreasing selectivity, more membrane area will be required and the incremental area is proportional to the reduction of the selectivity; but for FC, at a certain pressure ratio the required membrane area remains almost constant for different membrane selectivities. Generally, it can be deduced from the diagrams that FC leads to a higher degree of CO2 separation than that caused by PVP at a certain pressure ratio to reach the desired pCO2 . In their very influential article on a membrane feasibility study, Van der Sluijs et al. [13] used FC for all simulations. An argument for using PVP is that FC for the flue gas consumes too much energy for compressing the nitrogen in the feed gas. Two points should therefore be considered here: using FC, at the retentate side an expander should be coupled to recover part of the energy, which was also done by Van der Sluijs et al.; using PVP, vacuum pumps have to be available for industrial applications in order to meet the requirement of 30 or 60 mbar; an energy recovery of PVP cannot be carried

Fig. 11. (a) Influence of pressure ratio and selectivity ˛ on the membrane performance by using permeate vacuum pump at pCO2 = 0.09 bar at the end of the membrane separation process, (b) membrane area required by pCO2 = 0.09 bar at the end of the membrane separation process, a supplement to (a).

L. Zhao et al. / Journal of Membrane Science 325 (2008) 284–294

291

Fig. 12. (a) Influence of pressure ratio  and selectivity ˛ on membrane performance by using a feed gas compressor at pCO2 = 0.09 bar at the end of the membrane separation process, (b) membrane area required by pCO2 = 0.09 bar at the end of the membrane separation process, a supplement to (a).

out. The discussion will be continued in the next section concerning “energy consumption”.

4.6. Energy consumption of a single-stage membrane process

As described, the CO2 concentration of the flue gas is 14 mol% for most hard-coal-fired steam power plants. The low CO2 molar fraction and the ambient feed gas pressure are two fatal shortcomings for a membrane process used in post-combustion capture. Process engineering therefore has to increase the CO2 concentration in the feed gas. Valuable research about the influence on the membrane performance of varying the CO2 concentration in the feed gas has been performed by Favre’s group (software DASSL), who used 10 mbar permeate vacuum instead of 30 mbar (normally used by us as a reference pressure). Compared to the above-mentioned literature, the complete operating conditions and membrane quality data have been calculated and the results are shown in Fig. 13. Our calculation results (e.g. ˛ = 50, x = 20 mol%, A = 12 m2 , separation degree 40%) are in complete agreement with Favre’s work (both: CO2 purity ∼90 mol%).

In addition to the mass balance analyses, the energy consumption of a single-stage membrane process was investigated. It is well known that the energy demand for the membrane process is a crucial factor. In order to obtain a high driving force a high CO2 partial pressure should be created through using FC or PVP or both. For the energy consumption simulation, the efficiency of each compressor and vacuum pump is assumed to be 85%. As shown in Fig. 5, the initial attempt was started with PVP. Considering a membrane process with a selectivity of 40, the energy consumption for the process was calculated and the result is shown in Fig. 14. Since CO2 partial pressure decreases dramatically, the CO2 purity is degraded; owing to the greater permeation of N2 the electrical energy consumption is increased by the ratio of CO2 recovered per kilogram. One aspect (separation degree 40%, CO2 purity 80 mol%) of the process based on the reference case was chosen to make a further optimization for the vacuum pump arrangement. Fig. 15 reveals that two-stage vacuum pumps help to reduce energy consumption. Two pumps were placed at the permeate side; the outlet

Fig. 13. Influence of the CO2 concentration of feed gas and membrane selectivity ˛ on membrane performance. Operating conditions: permeate vacuum 10 mbar, feed gas flow rate 1 kmol h−1 = 22.4 m3 h−1 , area A = 1, 2, 3, . . ., 200 m2 .

Fig. 14. Schematic diagram of the variation of the energy consumption along the membrane process (˛ = 40) by PVP, operating conditions: feed flow 100 m3 h−1 , CO2 molar fraction 14 mol%, permeate vacuum 30 mbar.

4.5. Influence of CO2 concentration in the feed gas

292

L. Zhao et al. / Journal of Membrane Science 325 (2008) 284–294

Fig. 15. Energy consumption distribution of two-stage permeate vacuum pumps.

pressure of Pump 1 varied from 30 to 970 mbar with steps of 20 mbar. There is minimum compression energy of 0.11 kWh/kg separating CO2 when the outlet  pressure of Pump 1 amounts to 173 mbar, 30/1000 bar. A rule can therefore be set up which is equal to for the two-stage compressors, i.e. the square root of the desired pressure ratio can be chosen as a node to arrange the stage of the compressor or vacuum pump. The flow sheet of the two-stage vacuum pump system is shown in Fig. 16, where the pressure, temperature and work of each component are labeled. In a similar way, the stage concept and the minimum pressure node can be taken over for the compressor simulation used at the feed side. Aspects to be highlighted here are that due to the compression an equivalent amount of heat is generated and the gas is heated to quite a

high temperature, which could lead to damage to the machine for a one-stage vacuum pump system. In the case of a two-stage system, the temperature can be halved for each machine. This part of the heat can be reused by a suitable system design coupled with CO2 compression. When talking about energy consumption, a comparison between an MEA scrubbing process and membrane gas separation cannot be avoided. Some researchers have attempted to give a quantitative comparison between these two methods [13,22,23]in order to evaluate which is the more advantageous. It should be noted that for the MEA two parts of the energy balance should be considered, one is the thermal energy used for regenerating the MEA solvent and is obtained by extracting steam from the IP/LP steam turbine, which dominates the energy consumption of the process. The other small part is electrical energy to drive the machines. For membrane gas separation, all the energy required is electrical energy to generate the partial pressure difference. It is therefore not correct to make a direct comparison of the energy consumption between these two methods without any calculation exchange between electrical and thermal energy. The related efficiency penalty based on the data of power plant 1 (shown in Table 1) was calculated and listed in Table 6. The energy penalty of the current MEA technologies ranges from about 8–14%points for different types of power plants [1,5]. Furthermore, for an MEA scrubber it is usual to reach quite a high product CO2 purity (>95 mol%) and a high degree of CO2 separation (>90%), but for the membrane technique as simulated in this work, it is not easy to approach the target by a single-stage membrane process alone. The main CCS processes include capture, compression, transport and storage. In order to obtain the required CO2 purity for the future transport and injection requirements, multi-stage membrane arrangements coupled with a CO2 liquefaction process should be adopted for the system design with current levels of membrane selectivity.

Fig. 16. Pressure, temperature and work distribution of the two-stage vacuum pump system for a membrane process with 40% separation degree and 80 mol% CO2 purity.

Table 6 Requirement for single-stage membrane selectivity and the relative energy consumption (without liquefaction) and efficiency penalty at a given separation target based on power plant 1 shown in Table 1 membrane permeability assumed to be 1 m3 m−2 h−1 bar−1 CO2 purity (mol%)

Separation degree (%)

Selectivity

Membrane area (×106 m2 )

80.0 90.0 90.0

50.0 50.0 80.0

45 108 208

2.600 2.732 8.153

Energy consumption (kWh/kgCO2 separated ) 0.113 0.100 0.100

Efficiency penalty (%-point) 2.28 2.02 3.23

L. Zhao et al. / Journal of Membrane Science 325 (2008) 284–294

5. Summary and conclusions • The low CO2 molar fraction (14 mol%) and ambient pressure level of the flue gas make it difficult to use only a one-stage gas separation membrane in the post-combustion capture to meet the CO2 purity requirement (>95 mol%) for future transport and injection at the current level of development of the membrane technique. • Membrane permeability, selectivity and area are decisive parameters for membranes. Each of these properties should be taken into account by membrane developers. Because the capture target is high CO2 purity and a high degree of CO2 separation, the onestage membrane process alone cannot fulfill the task owing to the trade-off relation between membrane permeability and selectivity. Consideration should be given to developing a multi-stage gas separation membrane system, which combines the advantages of membranes with high permeability and membranes with high selectivity and may help to reach the target more easily. • Operating conditions (pressure ratio, CO2 molar fraction in the feed gas) dramatically influence the degree of separation and CO2 purity. Generally, at the same pressure ratio the compression of the feed gas facilitates the achievement of the same degree of separation with a relatively small membrane area in comparison to using a vacuum pump at the permeate side. • Energy consumption is a significant problem for gas-separation membrane processes. In principle, a multi-stage membrane system consumes more energy than a single-stage process; through optimization and integration of CO2 capture and liquefaction it can be expected that membrane separation will become more competitive with respect to energy consumption.

[4] [5] [6] [7] [8] [9]

[10]

[11] [12] [13]

[14] [15] [16] [17]

[18]

[19]

Acknowledgements This work is part of the METPORE project, which is funded by the Federal Ministry of Economics and Technology (BMWi), Germany, within the framework of the COORETEC program.

[20]

[21]

[22]

Nomenclature pfeed ppermeate pretentate p P V˙ feed V˙ permeate V˙ retentate x yP yR

pressure at the feed side (bar) pressure at the permeate side (bar) pressure at the retentate side (bar) partial pressure difference (bar) permeability (m3 m−2 h−1 bar−1 ) flow rate at the feed side (m3 h−1 ) flow rate at the permeate side (m3 h−1 ) flow rate at the retentate side (m3 h−1 ) gas mol fraction at the feed side (mol%) gas mol fraction at the permeate side (mol%) gas mol fraction at the retentate side (mol%)

[23]

[24] [25] [26]

[27]

[28] [29]

Greek letters ˛ selectivity ˛* process selectivity  pressure ratio

[30]

[31] [32]

References [1] H.J. Herzog, What future for carbon capture and sequestration? Environ. Sci. Technol. 35 (2001) 148–153. [2] B. Metz, O. Davidson, H. de Coninck, M. Loos, L. Meyer, IPCC Special Report on Carbon Dioxide Capture and Storage, Cambridge University Press, United Kingdom & New York, USA, available in full at www.ipcc.ch, 2005. [3] G. Göttlicher, The energetics of carbon dioxide capture in power plants, in: D.o.E. U.S., available in full at www.netl.doe.gov/publications/carbon

[33] [34]

[35]

293

seq/refshelf.html (Ed.), National Energy Technology Laboratory (NETL), 2004. An assessment of carbon capture technology and research opportunities, Standford University Global Climate & Energy Project (GCEP), Spring, 2005. CO2 capture ready plants, IEA Greenhouse Gas R&D Programme (IEA GHG), 2007/4, May 2007. J. Gibbins, CCS policy developments, in: TSEC Programme – UK Carbon Capture and Storage Consortium, Nottingham, 18–19th April, 2007. M. Krumbeck, Post combustion capture from coal, in: CSLF CCS Workshop Paris, March 27th, 2007. Scrubbing CO2 from Power Plant Flue Gas Using Monoethanolamine (MEA), http://www.netl.doe.gov. S. Santos, CO2 Capture, processing and transport, IEA Greenhouse Gas R&D Programme, Asia-Pacific Economic Coperation (APEC), Mexico City, 24th May, 2007. A. Schreiber, P. Zapp, W. Kuckshinrichs, Environmental impacts of coal-fired power generation with amine-based carbon capture – a life cycle approach, in: Proceedings of the 4th European Congress on Economics and Management of Energy in Industry, Porto, 27–30 November, 2007. A. Kohl, R. Nielsen, Gas Purification, fifth ed., Gulf Publishing Company, Houston, Texas, 1997. W.J. Koros, G.K. Fleming, Membrane based gas separations, J. Membr. Sci. 83 (1993) 1–80. J.P. Van der Sluijs, C.A. Hendriks, K. Blok, Feasibility of polymer membranes for carbon dioxide recovery from flue gases, Energy Convers. Manage. 33 (1992) 429–436. P.H.M. Feron, A.E. Jansen, R. Klaassen, Membrane technology in carbon dioxide removal, Energy Convers. Manage. 33 (1992) 421–428. A. Meisen, X. Shuai, Research and development issues in CO2 capture, Energy Convers. Manage. 38 (1997) 37–43. B.D. Freeman, Basis of permeability/selectivity tradeoff relations in polymeric gas separation membarnes, Macromolecules 32 (1999) 375–380. D.R. Pesiri, B. Jorgensen, R.C. Dye, Thermal optimization of polybenzimidazole meniscus membranes for the separation of hydrogen, methane, and carbon dioxide, J. Membr. Sci. 218 (2003) 11–18. C.E. Powell, G.G. Qiao, Polymeric CO2 /N2 gas separation membranes for the capture of carbon dioxide from power plant flue gases, J. Membr. Sci. 279 (2006) 1–49. S. Duan, T. Kouketsu, S. Kazama, K. Yamada, Development of PAMAM dendrimer composite membranes for CO2 separation, J. Membr. Sci. 283 (2006) 2–6. T. Kouketsu, S. Duan, T. Kai, S. Kazama, K. Yamada, PAMAM dendrimer composite membrane for CO2 separation: formation of a chitosan gutter layer, J. Membr. Sci. 287 (2007) 51–59. R. Bounaceur, N. Lape, R. Denis, C. Valliéres, E. Favre, G. Grévillot, Membrane process for post-combustion carbon dioxide capture: a parameteric study, GHGT8, Trondheim, 2006. R. Bounaceur, N. Lape, R. Denis, C. Vallieres, E. Favre, Membrane processes of post-combustion carbon dioxide capture: a parameteric study, Energy 31 (2006) 2220–2234. E. Farve, Carbon dioxide recovery from post-combustion processes: can gas permeation membranes compete with absorption? J. Membr. Sci. 294 (2007) 50–59. M.-B. Hägg, A. Lindbråthen, CO2 capture from natural gas fired power plants by using membrane technology, Ind. Eng. Chem. Res. 44 (2005) 7668–7675. C. Hendriks, Carbon Dioxide Removal from Coal-fired Power Plants, Kluwer Academic Publishers, Dordrecht/Boston/London, 1994. METPORE (Nano-structured Ceramic and Metal Supported Membranes for Gas Separation), together with the University of Queensland, Australia, funded by Federal Ministry of Economics and Technology (BMWi), Germany and industrial companies E.ON, EnBW, RWE, start January 1st, 2007. M. Conturie, Reduction of Carbon Dioxide Emissions by Capture and Reinjection, Renewable Energy Sources and Environment, Vrnjacka Banja, Serbia, 17th–24th October, 2006. S. Hagedoorn, Transportation of CO2 , International Interdisciplinary Summer School 2007 on CCS, Munich, Germany, 19th–24th, August, 2007. R. Ritter, B. Holling, H. Altmann, M. Biele, Konzepte und Ausblick für eine CO2 -Anlage eines Oxyfuel-Kraftwerkes am Beispiel Schwarze Pumpe, 39. Kraftwerkstechnisches Kolloquium, Dresden, Germany, 11th–12th, October, 2007. K. Aoki, K. Kusakabe, S. Morooka, Gas separation properties of A-Type zeolite membrane formed on porous substrate by hydrothermal synthesis, J. Membr. Sci. 141 (1998) 197–205. C.Y. Pan, Gas separation by permeators with high-flux asymmetric membranes, AICHE J. 29 (1983) 545–552. C.J. Geankoplis, Transport Processes and Separation Process Principles, fourth ed., Prentice Hall, New Jersey, 2003. W.J. Koros, Evolving beyond the thermal age of separation processes: membranes can lead the way, AICHE J. 50 (2004) 2326–2334. S. Kazama, S. Morimoto, S. Tanaka, H. Mano, T. Yashima, K. Yamada, K. Haraya, Cardo polyimide membranes for CO2 capture from flue gases, in: Proceedings of the 7th International Conference on Greenhouse Gas Control Technologies, Vancouver, Canada, September 5th, 2005. K.-V. Peinemann, Overview polymeric gas separation membranes, in: 9th Jülicher Werkstoffsymposium Gas Separation Mebranes for Zero-emission Fossil Power Plants, Jülich, Germany, November 15th–16th, 2007.

294

L. Zhao et al. / Journal of Membrane Science 325 (2008) 284–294

[36] S. Kulprathipanja, Mixed matrix membrane development, Ann. N. Y. Acad. Sci. 984 (2003) 361. [37] T.A. Centeno, A.B. Fuertes, Carbon molecular sieve membranes derived from a phenolic resin supported on porous ceramic tubes, Sep. Purif. Technol. 25 (2001) 379–384. [38] T.A. Centeno, J.L. Vilas, A.B. Fuertes, Effects of phenolic resin pyrolysis conditions on carbon membrane performance for gas separation, J. Membr. Sci. 270 (2004) 101–107. [39] J.Y.S. Lin, S. Chung, D. Li, J. Ida, J. Park, Dual-phase inorganic membrane for high temperature CO2 separation, http://www.netl.doe.gov/publications/ proceedings/04/UCR-HBCU/presentations/Lin P.pdf.

[40] Ceramic membrane to combat global warming, Membr. Technol. 1997 (1997) 11–12. [41] P. Kumar, J. Ida, V.V. Guliants, High flux mesoporous MCM-48 membranes: effects of support and synthesis conditions on membrane permeance and quality, Microporous Mesoporous Mater. 110 (2008) 595–599. [42] Q. Hu, E. Marand, S. Dhingra, D. Fritsch, W. Wen, G. Wilkes, Poly(amideimide)/TiO2 nano-composite gas separation membranes: fabrication and characterization, J. Membr. Sci. 135 (1997) 65–79. [43] L.M. Robeson, Correlation of separation factor versus permeability for polymeric membranes, J. Membr. Sci. 62 (1991) 165–185.