CO2-selective membranes for hydrogen productionand CO2 capture – Part I: Membrane development

Haiqing Lin a,n, Zhenjie He a, Zhen Sun a, Jimmy Vu a, Alvin Ng a, Moyeen Mohammed a,Jay Kniep a, Timothy C. Merkel a,nn, Tony Wub, Robert C. Lambrecht b

a Membrane Technology and Research, Inc., 39630 Eureka Drive, Newark, CA 94560, USAb National Carbon Capture Center, Southern Company, P.O. Box 1069, Wilsonville, AL 35186, USA

a r t i c l e i n f o

Article history:Received 1 October 2013Received in revised form7 January 2014Accepted 8 January 2014Available online 18 January 2014

Keywords:Polymeric membraneHydrogen purificationCO2 captureCoal-derived syngasLiquid CO2

a b s t r a c t

Syngas operations require separation of CO2 and hydrogen for hydrogen production at refineries andpetrochemical plants, and potentially for combined hydrogen production and CO2 capture at integratedgasification combined cycle (IGCC) power plants. Membranes that selectively permeate CO2 haveattracted significant interest, because the purified hydrogen is retained at high pressure, thus avoidingcostly downstream recompression. In this work, Polaris™ thin film composite membranes wereevaluated for CO2 removal from syngas: these membranes are highly permeable to CO2 and reject theother major components of syngas, including hydrogen, carbon monoxide and methane. Part I of thisstudy describes the three-stage development of Polaris membranes for CO2/H2 separation, including:laboratory parametric tests of membrane stamps using pure- and mixed-gas streams, pilot-scale tests ofmembrane modules using real syngas at a coal-fired gasification plant, and field demonstration of amembrane system producing liquid CO2. The gasification plant where the field tests were conducted is atthe National Carbon Capture Center (NCCC) in Wilsonville, AL. The membrane demonstration systemprocessed 227 kg/h of syngas (equivalent to a 0.15 MWe IGCC power plant) containing �9% CO2, andproduced a liquid CO2 stream containing 95þ% CO2. Part II of this study will explore the potentialopportunities for the use of CO2-selective membranes in commercial hydrogen purification and CO2

capture, including the results of techno-economic analyses for various process designs.& 2014 Elsevier B.V. All rights reserved.

1. Introduction

Hydrogen is an important chemical in petroleum refining,ammonia production and methanol synthesis. With annual pro-duction of 53 million metric tons worldwide, the hydrogen marketwas valued at $88 billion in 2010 [1]. Currently, hydrogen isproduced principally by steam reforming of hydrocarbons suchas methane, followed by the water-gas shift reaction:

CmHnþH2O-mCOþnþ2m2

H2 ð1Þ

COþH2O- CO2þH2 ð2ÞThe produced hydrogen inevitably contains impurities such as

CO2, which typically need to be removed for hydrogen to be usedfurther, as shown in Fig. 1 [2]. The CO2 is usually removed using

pressure swing adsorption (PSA) technology, which produceshigh-purity hydrogen at high pressures. However, the PSA systemstypically attain only 75–85% hydrogen recovery [3]. Consideringthe enormous amount of hydrogen made annually, any improve-ment in CO2/H2 separation efficiency would lead to significanthydrogen production savings.

Hydrogen has also attracted attention as a clean energy carrierfor power production, often starting from coal, an especiallyabundant source of fossil energy in the US and China [4]. In thisapplication (also shown in Fig. 1), coal is gasified to producesyngas. Ultimately the syngas is sent to a combustion turbine andthe waste heat from gasification and syngas combustion is used topower steam turbines. The process is called Integrated GasificationCombined Cycle (IGCC). The syngas may be shifted and treated toremove CO2 prior to combustion to avoid emission of this green-house gas. Because the CO2 in syngas is relatively concentratedand at high pressure, CO2 capture is easier in the pre-combustionIGCC approach than from post-combustion flue gas, which istypically at atmospheric pressure and contains only 4–13% CO2

[5]. When costs of carbon capture are included, IGCC powerproduction is believed to be less costly than power generated bydirect combustion of coal [5,6].

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/memsci

Journal of Membrane Science

0376-7388/$ - see front matter & 2014 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.memsci.2014.01.020

n Corresponding author. Present address: Department of Chemical and BiologicalEngineering, University at Buffalo, State University of New York, Buffalo, NY 14260,USA. Tel.: þ1 716 645 1856; fax: þ1 716 645 3822

nn Corresponding author. Tel.: þ1 650 328 2228; fax: þ1 650 328 6580.E-mail addresses: haiqingl@buffalo.edu (H. Lin),

tim.merkel@mtrinc.com (T.C. Merkel).

Journal of Membrane Science 457 (2014) 149–161

The current leading technologies for CO2 capture from shiftedsyngas in an IGCC system are physical absorption processes, suchas Selexol™ or Rectisols [2]. These processes have a number ofdrawbacks [2,4]. For example, the size of the absorption units isusually proportional to the amount of CO2 being removed: post-shift syngas contains CO2 at concentrations up to 45%, and there-fore would require large and complex absorption systems for CO2

removal. These absorption systems also consume large amounts ofenergy in the absorber-stripper operation, and as a result, moreefficient ways of separating CO2 from the shifted syngas are widelysought.

Membranes offer an alternative approach for CO2 capture andhydrogen purification at hydrogen production facilities and IGCCpower plants due to their simplicity of operation and potentiallyhigh energy efficiency [4,5]. Both H2-selective membranes [7–13]and CO2-selective membranes [14–26] have been explored. CO2-selective membranes are of interest for CO2/H2 separations insteam methane reforming – where it is desirable to leave thepurified hydrogen at high pressure by permeating CO2 from thesyngas [14–20], in air-blown gasification processes – where CO2

should be removed from syngas containing H2 and N2, and inoxygen-blown IGCC operations – where a combination of H2-selective membranes for bulk hydrogen recovery and CO2-selec-tive membranes for CO2 purification and liquefaction could be alow-cost CO2 capture process [5].

This paper describes the development of PolarisTM CO2-selec-tive membranes for hydrogen purification and CO2 capture,through three stages of membrane evaluation, as shown in Fig. 2.

(1) First, Polaris membranes with promising CO2/H2 separationproperties were tested in the laboratory to understand theeffect of operating parameters (including feed gas pressure,temperature and composition) on separation performance.Membrane stamps of 30 cm2 were mounted in permeation

cells (as shown in Fig. 2(a)) for testing with pure andmixed gases.

(2) Secondly, the membranes were fabricated into semi-commercial spiral-wound modules containing 1–4 m2 mem-brane area (as shown in Fig. 2(b)) and tested on a pilot-scalemembrane skid treating a real syngas feed from a coalgasification plant at the National Carbon Capture Center(NCCC), in Wilsonville, AL.

(3) Finally, a membrane-based demonstration system was con-structed to process 227 kg/h syngas (equivalent to a 0.15 MWe

IGCC power plant) and produce liquid CO2 at NCCC.The demonstration unit used commercial-scale membranemodules (20 cm diameter, each containing about 20 m2 ofmembrane area), as shown in Fig. 2(c).

Part II of this study will explore the potential commercialopportunities for the CO2-selective membrane in hydrogen pur-ification and CO2 capture, including the results of techno-economic analyses for various process designs [27].

2. Background on membranes

Industrial composite membranes for gas separations are oftencharacterized by their permeance and selectivity. Gas permeance,PA/l [cm3(STP)/cm2 s cmHg], can be expressed as follows [28]:

PA

l¼ NA

Amðp2;A�p1;AÞð3Þ

where NA [cm3(STP)/s] is the flux through the membrane, Am is themembrane area (cm2), and p2,A and p1,A are the partial pressure(cmHg) of component A in the feed and permeate, respectively.PA [cm3(STP) cm/cm2 s cmHg] is the permeability coefficient of gascomponent A in the selective material, and l is the thickness of the

Fig. 1. Simplified process flow diagram for hydrogen or energy production from gasified or reformed fossil fuels.

Fig. 2. Three stages of Polaris membrane development for treating syngas: (a) a laboratory test of membrane stamps (0.0030 m2) using permeation cells; (b) a field (pilot)test of bench-scale spiral-wound modules (1 m in length and 10 cm in diameter), each containing 1–4 m2 membrane area; (c) a CO2 liquefaction system demonstration, usingcommercial-scale spiral-wound modules (1 m in length and 20 cm in diameter), each containing about 20 m2 membrane area.

H. Lin et al. / Journal of Membrane Science 457 (2014) 149–161150

membrane selective layer (cm). Because l cannot be accuratelymeasured, the permeance, instead of permeability, is often used tocharacterize composite membranes.

If the diffusion process obeys Fick0s law and the downstreampressure is much less than the upstream pressure, the perme-ability coefficient is given by [28]:

PA ¼DA � SA ð4Þwhere DA is the average effective diffusivity, and SA is the solubilityof penetrant A in the polymer. The ideal selectivity of a membranefor gas A over gas B is the ratio of their pure-gas permeabilities orpermeances [28]:

αA=B ¼PA

PB¼ DA

DB

� �� SA

SB

� �ð5Þ

where DA/DB is the diffusivity selectivity, and SA/SB is the solubilityselectivity.

For CO2/H2 separation, solubility selectivity favors CO2 becauseCO2 (with a critical temperature of 304 K) is much more con-densable than hydrogen (with a critical temperature of 33 K) [29].However, diffusivity selectivity favors hydrogen, because hydrogen(with a kinetic diameter of 2.89 Å) is smaller than CO2 (with akinetic diameter of 3.3 Å) [29]. The favorable CO2/H2 solubilityselectivity is offset by the unfavorable diffusivity selectivity, somost polymeric materials do not show high CO2/H2 selectivity[15,16]. Optimized CO2-selective materials should have strongaffinity towards CO2 to enhance CO2/H2 solubility selectivity andweak size-sieving ability, to move CO2/H2 diffusivity selectivity asclose to 1 as possible [15,16].

3. Experimental

3.1. Membrane and module preparation

A schematic of industrial thin film composite membranes usedfor CO2/H2 separation is shown in Fig. 3 [30]. The nonwoven paperlayer (made from polyester) provides composite membranes withmechanical strength, and a substrate on which to cast subsequentmembrane layers. On top of this nonwoven paper, which has arelatively rough surface, a microporous ultrafiltration supportlayer is formed by wet-phase separation casting [31]. The proprie-tary gutter layer and selective layer were formed by a dip-coatingprocess using an industrial-scale coater [32]. The smooth surfaceof the gutter layer allows a very thin dense selective layer withhigh CO2/H2 separation performance to be deposited [33]. Themicroporous support and gutter layer contribute essentially noresistance to gas flow and therefore, the resistance to gas transport

in the thin film composite membranes lies mainly in the selectivelayer. Consequently, the gas selectivity is determined primarily bythe selective layer, in this case made of a polar rubbery polymer. Asshown in Eq. (3), gas permeance is inversely proportional to theselective layer thickness, so the selective layer is typically made asthin as possible (50–200 nm).

Fig. 4 shows an exploded view of a conventional cross-flowspiral-wound module for gas separation with three ports (feed,residue and permeate) [32]. Two flat sheets with permeate spacersin between the sheets are glued along three sides to form anenvelope with an open end facing the permeate pipe. Many of

Fig. 3. Schematic illustration of thin-film composite Polaris membranes [32]. The selective layer has high intrinsic CO2/H2 separation performance, while the gutter layer,microporous support and paper layer show negligible gas transport resistance.

Fig. 4. Exploded view of a conventional spiral-wound membrane module (in cross-flow mode) for gas separation [32,34].

Fig. 5. Schematic of the pure- and mixed-gas test system used to characterize flatsheet composite membranes and spiral-wound modules. A custom-built permea-tion cell (with an active membrane area of 30 cm2) was used for membrane testing.FM: bubble flow meters; MFM: mass flow meters.

H. Lin et al. / Journal of Membrane Science 457 (2014) 149–161 151

these envelopes are wound around a perforated central collectionpipe with a feed spacer between each set of envelopes. Feed gaspasses down the module parallel to the collection pipe in thechannel created by the feed spacer. A portion of the feed gaspermeates the membrane, enters the permeate channel, and exitsthrough the collection pipe.

3.2. Determination of pure- and mixed-gas permeances in thelaboratory

Pure- and mixed-gas permeances in thin film compositemembranes were determined using a constant-pressure/variable-volume apparatus, as shown in Fig. 5 [35]. Before each measure-ment, the entire apparatus is purged using the feed gas orvacuumed to remove any impurities from the system. The perme-ate gas flow rate is measured using bubble flow meters (AlltechAssociates, Inc. Deerfield, IL, USA) or mass flow meters (SierraInstruments, Monterey, CA). The steady state pure-gas permeancefor gas A, PA/l, can be calculated by [35]

PA

l¼ 1Amðp2�p1Þ

273T

dVdt

� �ð6Þ

where the downstream pressure, p1, is atmospheric pressure inthis case, T is the absolute temperature of the gas (K), and dV/dt isthe permeate steady state volumetric flow rate (cm3/s). Gaspermeance is usually expressed in units of gas permeation units(gpu), where 1 gpu¼10�6 cm3(STP)/(cm2 s cmHg), equivalent to7.501 x 10-12 m3(STP)/m2 s Pa. The gpu is the the industry standardunit for reporting membrane permeance values, and is usedthroughout this paper.

The mixed-gas permeance of component A in the membranes,PA/l, can be calculated using the following equation, assuming thepermeate in the membrane downstream is well mixed [36]:

PA

l¼ NPx1;AAmðp2x2;A�p1x1;A Þ

ð7Þ

where NP is the total permeate flow rate, x1,A is the mole fraction ofcomponent A in the permeate and x2;A is the average concentrationof component A in the membrane upstream side. While theparameters of NP, x1,A, Am, p2 and p1 can be measured directly

using the instruments on the membrane skid, the parameter x2;A isestimated as the logarithmic mean of component A concentrationon the membrane upstream side.

x2;A ¼xF ;A�xR;A

lnðxF ;A=xR;AÞð8Þ

where xF,A and xR,A are the mole fraction of component A in thefeed entering the membrane or module and the residue exitingthe membrane or module, respectively. For the membrane stamptest, the stage-cut (defined as the percentage of the feed gaspermeating across the membrane) was less than 1%. For themodule test, the stage-cut was less than 20%. The gas permeancevalues obtained were also confirmed using a commercial processsimulation package, CHEMCAD 6.3 (from Chemstations Inc., Hous-ton, TX), combined with an enhanced membrane process simu-lator. Similar methods for membrane process calculations havealso been shown elsewhere [37,38].

Ultra-high purity hydrogen with a purity of 99.9% and chemical-grade carbon dioxide with a purity of 99% were used as receivedfrom Praxair Inc. (Hayward, CA). Certified gas mixtures containingcarbon dioxide, hydrogen and nitrogen were also received fromPraxair Inc.

3.3. Field test at the National Carbon Capture Center (NCCC)

3.3.1. Syngas productionThe National Carbon Capture Center (Wilsonville, AL) is finan-

cially supported by the US DOE as a leading test center fordevelopment of advanced technologies that can achieve near-zero emissions in coal-based power plants, including CO2 capturetechnologies [39,40]. NCCC operates a coal gasifier based on theTransport Integrated Gasification (TRIG™) process. The TRIGprocess is expected to have lower capital and operating coststhan other gasification processes because of its lower operatingtemperature, resulting in less expensive construction materials,and the capability of using less expensive low rank coals [41].The gasifier at NCCC is usually air blown and often fed withPowder River Basin (PRB) coal or lignite, producing as much as9072 kg/h syngas, equivalent to the output of a 6 MWe IGCC powerplant [40].

90 kg/h

227 kg/h

T.O./stack

Future upgrade1.5 MWe

2270 kg/h syngas

681 kg/h syngas

454 kg/h syngasPilot testunit #2

0.45 MWe

538-I F6

Fig. 6. Layout of the facility and the MTR membrane units at the National Carbon Capture Center (NCCC) for the evaluation of technologies at various levels of processmaturity [42]. The syngas feed produced from the gasifier may be treated to remove sulfur and tar components before entering the membrane skid. T.O.: thermal oxidizer.

H. Lin et al. / Journal of Membrane Science 457 (2014) 149–161152

Fig. 6 shows the layout of the NCCC facility for the evaluation ofsyngas conditioning and CO2 capture technologies at various levelsof process maturity. Typical syngas conditions from PRB coal arerecorded in Table 1. The raw syngas contains about 10% CO2 and10% hydrogen, with N2 accounting for the bulk of the remainderdue to the use of air in the gasifier. The syngas also contains high-boiling-point aromatic hydrocarbons (tars). The NCCC facility isable to provide a range of syngas flow rates (from 4.5 kg/h to 454kg/h) 454 kg/has needed for the evaluation of various technolo-gies. The raw syngas can be sent to a water-gas shift reactor toconvert most of the carbon monoxide to hydrogen and CO2. Thetypical composition of post-shifted syngas is included in Table 1.NCCC is equipped with the necessary analytical instruments forreal-time analysis and monitoring [40].

3.3.2. Pilot-scale membrane skidA pilot-scale membrane skid was constructed and installed at

NCCC to test the separation performance of Polaris membranemodules for CO2 removal from the real coal-derived syngas.Fig. 7(a) shows a photo of the skid, which can accommodate oneor two 10-cm-diameter membrane modules, each containing 1–4 m2

membrane area. Fig. 7(b) shows the simplified process flow diagramfor the membrane skid. After pretreatment (including syngas coolingand tar scrubbing), the syngas at near ambient temperature entersthe membrane skid. After separation, the CO2-lean residue gas andCO2-rich permeate gas streams are measured, re-combined, and sentas one stream to the thermal oxidizer (T.O.) at NCCC.

For the pilot-scale field test, the syngas flowed through themembrane skid at rates between 4.5 kg/h and 22.7 kg/h at feedpressures up to 13.4 bar. The temperature, pressure, and flow rateof the feed and permeate gas streams were measured using gaugeson the skid. The residue stream flow rate was calculated based ona mass balance. Gas compositions were determined using Siemensgas chromatography at NCCC. Mixed-gas permeances were calcu-lated using Eq. (7), and were also confirmed using the processsimulation package, ChemCAD 6.3.

3.3.3. Membrane demonstration system for liquid CO2 productionAfter the successful pilot-scale testing of Polaris membrane

modules, a small commercial-scale membrane demonstrationsystem was designed and built. This system was designed to testa membrane-assisted CO2 liquefaction process. Fig. 8 shows aprocess flow diagram of the membrane demonstration system,which was designed to treat 227 kg/h of coal-derived syngas,equivalent to the syngas output of a 0.15 MWe IGCC power plant.Before entering the membrane system, the syngas was pretreatedto remove sulfur compounds (including hydrogen sulfide (H2S)and carbonyl sulfide (COS)) to less than 1 ppm, sent to a water

quench tank and scrubber to remove tars, and then cooled to 10 1C.Two ZnO beds were used to remove H2S and COS, avoiding theirco-liquefaction in the liquid CO2 condenser of the membranesystem, and the tars were removed to prevent tar build-up andblocking of the valves or pipes in the system.

As shown in Fig. 8, the membrane demonstration systemconsists of a two-stage membrane separation and condensationunit for CO2 liquefaction [5]. The first membrane stage canaccommodate one or two 20-cm-diameter modules with a totalmembrane area of 20–40 m2, and the second membrane stageholds one or two 10-cm-diameter modules with a total membranearea of 3–6 m2. The syngas at 12.7 bar is heated to 20 1C andpasses to the first stage membrane, which concentrates CO2 from10% in the feed to 30–40% CO2 in the permeate. The hydrogen-enriched residue, which would typically be the product in acommercial system, is sent to the thermal oxidizer at NCCC.The Polaris membrane also preferably permeates polar compo-nents such as H2S and water, compared to hydrogen. The CO2-richpermeate stream is treated to remove H2S and water vapor,compressed to 30 bar, and cooled to �30 1C to condense liquid CO2.The overhead stream from the condenser at 30 bar is heated toabout 10 oC and sent to the second-stage membrane as the feed.The CO2-lean gas (containing 10% CO2) from the second-stagemembrane is sent to the thermal oxidizer, and the CO2-richgas (containing 50% CO2) from the second stage is recycled tothe front of the compressor. The overall system captures up to 70%of the CO2 in the feed as high-pressure high-density fluid ready forsequestration.

The membrane demonstration system has three skids: amembrane skid, a compressor skid and a chiller skid. Fig. 9 showsthe general arrangement (GA) drawing of the entire membranedemonstration system. The skid size is 25 feet (length) by 25 feet(width) by 12.5 feet (height).

Fig. 10 shows photos of the three skids that make up themembrane demonstration system, before and after installation at

Table 1Typical syngas stream conditions at the NCCC gasification plant.

Operating conditions Raw syngas Shifted syngas

Flow rate (lb/h) 20,000 1000Pressure (bar) 13.4 13.4Temperature (1C) 230 230Water Saturated SaturatedTars (aromatic hydrocarbons) Saturated Saturated

Gas composition (dry vol%)H2 7.3 15.4N2 73.0 65.8CH4 1.2 1.2CO 9.3 1.1CO2 9.2 16.5H2S 100–400 ppm 100–400 ppm

ResidueFeedPermeate

Fig. 7. (a) Photograph and (b) process flow diagram of the pilot-scale membraneskid installed at NCCC that can treat up to 22.7 kg/h syngas. The system canaccommodate up to two 10 cm modules with a total membrane area of 2–8 m2. FM:flow meter; P: pressure gauge.

H. Lin et al. / Journal of Membrane Science 457 (2014) 149–161 153

NCCC. Fig. 10(a) and (b) shows photos of the membrane skid thatenriches CO2 from the raw syngas. The membrane skid containstwo module housings, one sulfur removal vessel and two dryervessels. Fig. 10(c) shows the compressor skid, which can compressthe syngas from about 1 bar to as high as 30 bar. Fig. 10(d) showsthe chiller package, which includes the chiller that cools thesyngas to �30 1C and the condenser that is used to collect liquidCO2. Fig. 10(e) shows the NCCC site with foundation prior toinstallation, and Fig. 10(f) shows the complete membrane demon-stration system installed at the site.

4. Results and discussion

4.1. Laboratory testing of Polaris membranes

4.1.1. Pure-gas permeation propertiesPolaris membranes, comprised of a proprietary polymer as the

selective layer, have been designed for optimized CO2/H2 and CO2/N2 separation properties. Fig. 11 shows pure-gas permeances andCO2/H2 and CO2/N2 selectivities of a Polaris membrane stamptested at temperatures from �10 1C to 35 1C. In general, astemperature decreased, pure-gas permeances decreased, andpure-gas CO2/H2 and CO2/N2 selectivities increased, which are

typical behavior for solubility selective polymer materials [43,44].Decreasing temperature makes polymer chains more rigid, leadingto the decrease in gas diffusivity and thus gas permeance. On theother hand, the separation is governed by the solubility selectivity,which usually increases with decreasing temperatures [43].

Optimal membranes for CO2/H2 separation should have high CO2

permeance to reduce the required membrane area and capital cost,and high CO2/H2 selectivity to increase the product CO2 purity.Fig. 12 presents a permeance/selectivity map for CO2/H2 separationin polymeric membranes at 25 1C, where gas transport follows thesolution-diffusion mechanism [29,45,46]. This type of plot waspopularized by Robeson [45,46]. Each point represents the selectedseparation properties for one particular polymer. The upper boundline in the figure gives a rough estimate of the highest selectivitypossible for a given permeability in polymer-based materials [29]. Incontrast to separations which require strong size-sieving ability toachieve high selectivity, the upper bound for CO2/H2 separationshows a positive slope; that is, materials with higher CO2 permeanceoften have higher CO2/H2 selectivity as well [15].

Fig. 12 assumes that all polymers other than Polaris can bemade into stable asymmetric membranes with a selective layerthickness of 1 μm, without further discussions on the feasibility ofmembrane fabrication and physical aging behavior. For Polaris, thedata are actual thin film composite results for an optimized mem-brane. These Polaris membranes show CO2/H2 separation perfor-mance close to the upper bound (CO2 permeance of 2000 gpu, andCO2/H2 selectivity of 10 at 25 1C). Decreasing temperature decreasesCO2 permeance and increases CO2/H2 selectivity of the Polarismembranes, providing great flexibility in the combinations of CO2

permeance and CO2/H2 selectivity that can be produced.Fig. 12 also compares Polaris membranes with other membrane

materials developed for CO2/H2 separation, which contain mainlypoly(ethylene oxide) components, such as Pebax/PEG [47], PEGDA/PEGMEA [48], PEO/PDMS [49], and PEGMEAþparticles [50]. ThePolaris membranes show good combinations of CO2 permeanceand CO2/H2 selectivity, compared to these other membranematerials. Additionally, the Polaris membranes have been pro-duced at commercial scale for the field tests.

4.1.2. Mixed-gas separation propertiesVarious Polaris membranes were tested with gas mixtures

to understand the effect of operating conditions on mixed-gas

Fig. 8. A simplified process flow diagram of the membrane demonstration system constructed by MTR for capturing CO2 from 227 kg/h coal-derived syngas at NCCC.

H2S removal unitDryer

Membranevessel

Fig. 9. A general arrangement (GA) drawing of the 227 kg/h syngas membranedemonstration system installed at NCCC.

H. Lin et al. / Journal of Membrane Science 457 (2014) 149–161154

CO2/H2 separation properties. The results of one stamp test arepresented and discussed below. This membrane stamp showedpure-gas CO2 permeance of 1300 gpu and CO2/H2 selectivity of 12at 4.4 bar and 20 1C.

Fig. 13 shows the mixed-gas permeances and selectivities atvarious feed gas compositions, temperatures and pressures.Decreasing temperature decreased mixed-gas CO2 permeanceand increased mixed-gas CO2/H2 selectivity, which is similar tothe pure-gas permeation properties (as shown in Fig. 11), andtypical behavior for solubility selective polymers. The effect ofpressure on separation properties is more complicated, andreflects an interplay between CO2 plasticization, feed gas compres-sion and non-ideal behavior [15,19]. The CO2 in the feed sorbs intoand swells the polymer, increasing the polymer chain flexibility,and thereby gas diffusivity and permeance [19]. On the other hand,feed gas compression decreases polymer free volume, leading to

lower diffusivity and permeance [43,51]. High feed pressure alsoincreases the degree of non-ideal, that is, the fugacity coefficientdecreases with increasing feed pressure, especially at lower tem-peratures. Therefore, the use of Eq. (7) for permeance overestimatesthe driving force for CO2, leading to lower CO2 permeances. Thecombination of these factors results in there being no clear trend forthe effect of feed pressure on the mixed-gas separation properties.For example, mixed-gas CO2 permeance at �20 1C with the mixturecontaining 80% CO2 and 20% hydrogen decreased with increasingfeed pressure, which was different from the results at other tem-peratures. Because the process simulations (discussed in more detailin the Part II study [27]) are based on the pressure, which can bedirectly measured, the gas permeance is calculated based on pres-sure, rather than fugacity.

Despite the somewhat complex behavior shown in Fig. 13, it isclear that at 0 1C and �20 1C, the membrane stamp showed

Membrane skidCompressor

Chiller

Membrane skidCompressor

Chiller

Fig. 10. Photos of the 227 kg/h syngas membrane demonstration system: (a) membrane skid; (b) membrane module housings on the membrane skid; (c) compressor;(d) chiller and CO2 condenser; (e) site preparation for the demonstration system at NCCC; and (f) installed membrane demonstration system at NCCC.

H. Lin et al. / Journal of Membrane Science 457 (2014) 149–161 155

mixed-gas CO2/H2 selectivity of more than 10, which is a targetidentified by process simulations and economic analyses (dis-cussed in Part II of this study). In addition, the presence of nitrogenin the feed gas (at 0%, 20%, or 40%) seems to have minimal effect ofmixed-gas CO2/H2 separation properties.

4.2. Pilot-scale test of Polaris membrane modules

The gasifier plant at NCCC operates on a campaign basis with atotal operating time of approximately 2000 h per year. There are2–4 campaigns per year with each campaign lasting 500–1000 h.The Polaris membrane modules were tested for CO2 removal fromsyngas using the pilot-scale skid (shown in Fig. 7) in several runs.

4.2.1. CO2 removalDuring the gasifier R03 run (an NCCC designation for the

gasifier run from November 17, 2009 to December 10, 2009), theraw syngas (with a flow rate of 4.54 kg/h was sent to a water-gas

shift reactor and then passed through a desulfurization unit toremove sulfur components (H2S and COS) down to less than10 ppm. The shifted syngas was then fed to a water cooler/condenser to remove tars, and cooled to near ambient tempera-ture. The typical shifted syngas stream conditions entering thepilot-scale test unit were provided earlier in Table 1.

Fig. 14 shows the Polaris membrane module performance forCO2 removal from syngas during approximately 20 days ofcontinuous operation with a single pilot-scale module. The feedtemperature fluctuated between �5 1C and 25 1C during thefield test, mainly due to changes in the ambient temperature.Fig. 14(a) shows that the Polaris membrane module can enrichCO2 content, from 10–12% in the feed to 40–60% in the permeate.The performance was stable during the entire period of operation,indicating that there was no aging or performance deterioration ofthe membrane module after exposure to the shifted desulfurizedsyngas.

Fig. 14(b) and (c) shows mixed-gas permeances and mixed-gas CO2/gas selectivities for the module as a function of time.In general, the Polaris membrane module shows CO2 permeanceof 100–300 gpu, CO2/H2 selectivity of 6–10, CO2/CO selectivityof 10–20, and CO2/N2 selectivity of 20–50. It should be notedthat a relatively low permeance Polaris membrane modulewas prepared for this test to control the stage cut in theexperiment to less than 15%. If a high permeance Polaris modulewas used with the available syngas feed flow rate, most of thefeed gas would permeate the module. Overall, the importantresult from this test is that the module performance was stablewith time.

4.2.2. H2S removalIn another gasifier campaign (R06 run from July 11 to August 20,

2011), the sulfur components were not removed before enteringthe membrane pilot skid. The syngas compositions were similarto that in Table 1, but with a H2S content of 320 ppm. Fig. 15shows the H2S/CO2 selectivity in the module at 32 1C as a functionof time. The Polaris membranes showed H2S/CO2 selectivityof around 3, consistent with rubbery polymers where separationis determined by the solubility selectivity. H2S with a criticaltemperature of 373 K is more condensable than CO2 with acritical temperature of 304 K, and therefore, H2S has highersolubility and permeability than CO2. The field test with the

1

10

100

1,000

3.0 3.2 3.4 3.6 3.8 4.0

Pure

-gas

per

mea

nce

(gpu

)

1,000/T (1/K)

CO2

H2

CH4

N2

35 20 10 0 -10 oC

1

10

100

3.0 3.2 3.4 3.6 3.8 4.0

Pure

-gas

sel

ectiv

ity

1,000/T (1/K)

CO2/H

2

35 20 10 0 -10 oC

CO2/N

2

Fig. 11. (a) Pure-gas permeances and (b) pure-gas CO2/H2 and CO2/N2 selectivities of a Polaris membrane stamp at a feed pressure of 4.4 bar and varying temperatures.1 gpu¼10�6 cm3(STP)/cm2 s cmHg.

1

10

100 101 102 103 104

Pure

-gas

CO

2/H

2 se

lect

ivity

CO2 permeance (gpu)

Upperbound

PolarisTM

25oCPEGDA/PEGMEA

Pebax/PEGblends

PEGMEA/particlesPEO/PDMS

-10oC20

Fig. 12. Comparison of an optimized Polaris membrane with other polymericmembranes in the literature (with assumed 1 mm-thick selective layer) usedfor CO2/H2 separation. Pebax/PEG [47]; PEGDA/PEGMEA [48]; PEO/PDMS [49];PEGMEAþparticles [50].

H. Lin et al. / Journal of Membrane Science 457 (2014) 149–161156

syngas containing H2S shows that the membrane module is stablein the presence of H2S.

Thus, the Polaris membranes provide a way to co-capture H2Sand CO2, if co-sequestration of H2S and CO2 is possible [5]. SinceH2S is more permeable than CO2 through Polaris membranes, if90% CO2 is removed from the syngas, even greater percentages ofH2S will be removed at no additional cost.

4.3. Membrane system demonstrating CO2 liquefaction

Based on successful pilot-scale module testing, a larger mem-brane demonstration system for treating 227 kg/h of syngaswas designed, built, and installed at NCCC. This system wouldtest not only the membrane components, but the overallmembrane process for syngas purification and CO2 liquefaction.

100

1,000

0 20 40 60 80

Mix

ed-g

as C

O2 p

erm

eanc

e (g

pu)

Feed pressure (bar)

20°C

0°C

-20°C

50 0

10

20

30

0 20 40 60 80

Mix

ed-g

as C

O2/H

2 sel

ectiv

ity

Feed pressure (bar)

20°C

0°C

-20°C

100

1,000

0 20 40 60 80

Mix

ed-g

as C

O2 p

erm

eanc

e (g

pu)

Feed pressure (bar)

20°C

0°C

50 0

10

20

30

0 20 40 60 80

Mix

ed-g

as C

O2/H

2 sel

ectiv

ity

Feed pressure (bar)

20°C

0°C

100

1,000

0 20 40 60 80

Mix

ed-g

as C

O2 p

erm

eanc

e (g

pu)

Feed pressure (bar)

20°C

0°C

-20°C

50 0

10

20

30

0 20 40 60 80

Mix

ed-g

as C

O2/H

2 sel

ectiv

ity

Feed pressure (bar)

20°C

0°C

-20°C

Fig. 13. Effect of temperature and pressure on the mixed-gas separation properties of a Polaris membrane stamp. (a) Mixed-gas CO2 permeance and (b) CO2/H2 selectivitywith a 20:40:40 (CO2:H2:N2) feed; (c) mixed-gas CO2 permeance and (d) CO2/H2 selectivity with a 60:20:20 (CO2:H2:N2) feed; (e) mixed-gas CO2 permeance and (f) CO2/H2

selectivity with an 80:20 (CO2:H2) feed.

H. Lin et al. / Journal of Membrane Science 457 (2014) 149–161 157

Before entering the membrane system, the syngas was treated toremove sulfur compounds and tars; typical stream conditionswere given in Table 1. The removal of tars (mainly heavy hydro-carbons) is critical to avoid liquid condensation and fouling in themembrane system. During the field tests, the tars were removedusing a water scrubber operating at 10 1C, and then the syngas washeated to 20 1C before being sent to the membrane system toavoid the condensation of water and tars in the membranemodules.

The liquefaction of CO2 is a critical step for CO2 capture andsequestration from syngas using membrane technology [5,52].For sequestration, the CO2 needs to be pressurized to 150 bar forinjection underground. A low energy means of doing this is toliquefy the CO2 at lower pressure and then pump the liquid toinjection pressure [5,52]. More details on the process design willbe shown in Part II of this study [27].

A simplified process flow diagram of the membrane systemwas shown in Fig. 8. During the NCCC R10 gasifier campaign(between March 21 and April 21, 2013), the first-stage membraneenriched CO2 from 9% to 33% in the permeate. The membranesystem successfully produced a liquid CO2 stream in the condenserat �30 1C and 30 bar, containing 95þ% CO2, starting from thecoal-derived syngas feed containing �9% CO2. Fig. 16(a) showsthe CO2 content in the liquid CO2 stream as a function of time.

10

100

0 5 10 15 20 25

CO

2 con

cent

ratio

n (%

)

Operating time (days)

Permeate

Feed

20

40

60

1

10

100

1000

0 5 10 15 20 25

Mix

ed-g

as p

erm

eanc

e (g

pu)

Operating time (days)

CO2

H2

N2

CO

1

10

100

0 5 10 15 20 25

Mix

ed-g

as s

elec

tivity

Operating time (days)

CO2/N

2

CO2/H

2

CO2/CO

Fig. 14. Operating the pilot-scale membrane module test skid: Polaris membrane module performance as a function of time during a test period of about 20 days. (a) CO2

concentration in the feed and permeate streams; (b) mixed-gas permeances for CO2, H2, CO and N2 and (c) mixed-gas CO2/gas selectivities. Feed gas composition is similar tothat shown in Table 1.

0

2

4

6

0 20 40 60

Mix

ed-g

as s

elec

tivity

Operating time (days)

H2S/CO

2

Fig. 15. Mixed-gas H2S/CO2 selectivity in Polaris membrane modules as a functionof time during the R05 test period from July 11 (Day 0) to August 20, 2011 (Day 40).The feed gas contained 320 ppm. H2S at a temperature of around 32 1C. Otheroperating conditions and the typical feed gas composition are shown in Table 1.

H. Lin et al. / Journal of Membrane Science 457 (2014) 149–161158

The fluctuation in the CO2 content (or the reduction of CO2 contentto less than 80%) is presumably due to sporadic brief shutdowns ofthe chiller that caused the CO2-lean gas to flow into the liquid CO2

stream. Fig. 16(b) shows the production of liquid CO2 in the sightglass of the condenser. The liquid CO2 production rate was 4.5-6.8kg/h, which corresponds to about 30–45% of the feed CO2. Theseresults are close to the values expected from process simulations,based on measured membrane properties.

Table 2 records several samples of measured liquid CO2 streamcompositions after the system was running at steady state forseveral hours, and compares these results with the simulatedcompositions from a commercial CHEMCAD 6.3 process simulator.The liquid streams were at �30 1C and 30 bar. In the simulation,the H2S content in the feed gas entering the condenser wasassumed to be 5 ppmv, and the Soave–Redlich–Kwong thermo-dynamic equation of state was used to describe the phase behaviorof the gas mixtures. As shown in Table 2, the measured composi-tions are fairly close to the simulated one.

Following the testing described above, minor modificationswere made to the demonstration system, including the addition ofimproved Polaris membrane modules. The system was thenoperated during the NCCC R11 gasifier campaign (between August6 and September 2, 2013). The first-stage membrane enriched CO2

from 11.5% in the feed to 40% in the permeate, which was higherthan the permeate CO2 content in the R10 gasifier campaign(about 33%), due to the better Polaris membrane modules. Themembrane system continuously produced a liquid CO2 stream inthe condenser containing 95þ% CO2 at �33 1C and 27 bar. Theliquid CO2 production rate improved significantly in R11 to 15.9-18.1 kg/h, which corresponds to about 60–70% of the feed CO2. The

run time of the demonstration skid was over 400 h during R11with no degradation in membrane performance.

The operation of the 227 kg/h syngas demonstration unitshows that an integrated Polaris membrane-refrigeration systemcan reliably produce liquid CO2 from coal-derived syngas. Thesuccessful operation of this system provides a baseline for futureoptimization and improvement. As we will discuss in Part II [27],we believe such a membrane system, possibly used in combinationwith hydrogen-selective membranes, has potential for use in CO2

capture from IGCC and hydrogen production processes.

5. Conclusions

This paper describes the development of CO2-selective mem-branes (Polaris) for CO2 capture from syngas, which can be dividedinto three phases: laboratory testing of membrane stamps, pilot-scale testing of small membrane modules at a syngas productionplant, and operation of commercial-sized modules on a demon-stration system that produced liquid CO2 from raw syngas. Theresults from this work are summarized below.

(1) Laboratory testing of membrane stamps (with a membranearea of 30 cm2) provided a database of CO2/H2 separationproperties for Polaris membranes at various temperatures,pressures, and gas compositions. The membranes showedmixed gas CO2/H2 selectivity above 10 at temperatures below0 1C, and pressures as high as 56 bar.

(2) Pilot-scale Polaris membrane modules were tested with coal-derived syngas at the National Carbon Capture Center (NCCC).These semi-commercial modules (containing 1–4 m2 mem-brane area) showed CO2/H2 separation properties similar tothose obtained in membrane stamp tests. The test alsodemonstrated the long-term stability of the Polaris modulesin a real syngas environment containing up to 320 ppm of H2S.

(3) A membrane demonstration system processed 227 kg/h syn-gas (equivalent to the syngas production of a 0.15 MWe IGCCpower plant) containing �9% CO2, and produced a liquid CO2

stream containing 95þ% CO2. Commercial-scale membranemodules (containing 20 m2 of membrane area) were used inthese tests.

Part II of this study will explore the potential opportunities forPolaris membrane in commercial hydrogen purification and CO2

capture, including the results of techno-economic analyses forvarious process designs.

0

20

40

60

80

100

0 10 20 30 40

CO

2 co

nten

t (%

)

Operating days

Syngas

LiquidCO2product

Liquid CO2

Fig. 16. (a) CO2 content in the raw syngas and the liquid CO2 stream leaving the condenser during the NCCC R10 gasifier campaign; (b) a photo of liquid CO2 visible in thecondenser sight glass. The lines in the figure are to guide the eye.

Table 2Comparison of several liquid CO2 stream compositions measured after the systemwas running at steady state for several hours, and the CHEMCAD simulatedcomposition. The liquid streams were at �30 1C and 30 bar.

Component Measured compositions of the liquid CO2

stream (mol%)Simulated composition(mol%)

1 2 3

CO2 94.36 97.43 99.00 98.52N2 3.67 1.34 0.32 1.26CO 0.54 0.20 0.04 0.18CH4 0.09 0.06 0.01 0.03H2 0.56 0.19 0.01 0.01H2S 0.78 0.78 0.62 0.20

H. Lin et al. / Journal of Membrane Science 457 (2014) 149–161 159

Acknowledgments

We gratefully acknowledge the partial financial support of thiswork by the United States Department of Energy, Office of FossilEnergy, Phase III Xlerator program (Award no.: DE-FE0006138). Thisresearch was also partially supported by the US Department ofEnergy, National Energy Technology Laboratory under CooperativeAgreement no. DE-FE0005795. The authors also acknowledge theNational Carbon Capture Center operated by Southern CompanyServices, and supported and funded by the US Department ofEnergy through Cooperative Agreement DE-NT0000749 and byother industrial cofounders. One of the authors, H. Lin, alsoacknowledges the funds provided by University at Buffalo, The StateUniversity of New York, during the preparation of the manuscript.

Nomenclature

Am membrane area (cm2)dV/dt permeate steady state volumetric flow rate (cm3/s)DA diffusivity of gas A (cm2/s)DA/DB diffusivity selectivity of gas A over gas Bgpu gas permeation unit [10�6 cm3(STP)/(cm2 s cmHg)]l membrane selective layer thickness (cm)MMscfd million standard cubic feet per dayNA steady state flux of gas A across the membrane

[cm3(STP)/s]NP total steady state flux across the membrane

[cm3(STP)/s]p1 downstream pressure (bar)p1,A downstream partial pressure of gas A (bar)p2 upstream pressure (bar)p2,A upstream partial pressure of gas A (bar)PA permeability of a polymer to gas A (Barrer)PA/l permeance of a composite membrane to gas

A (gpu)SA solubility of gas A in a polymer [cm3(STP)/

(cm3 polymer atm)]SA/SB solubility selectivity of gas A over gas BT temperature (K)x1,A mole fraction of component A in the permeatex2;A average concentration of component A in the

membrane upstreamxF,A mole fraction of component A in the feedxR,A mole fraction of component A in the residue

Greek letter

αA/B permeability or permeance selectivity of gas Aover gas B

Subscript

1 membrane downstream side2 membrane upstream sideA component AB component B

References

[1] Hydrogen Generation Market – By Merchant & Captive Type, Distributed &Centralized Generation, Application & Technology – Trends & Global Forecasts

(2011–2016), ⟨http://www.marketsandmarkets.com/⟩ (accessed in August2013).

[2] A.L. Kohl, R. Nielson, Gas Purification, 5th ed., Gulf Publishing, Houston, TX,1997.

[3] S. Sircar, W.E. Waldron, M.B. Rao, M. Anand, Hydrogen production by hybridSMR-PSA-SSF membrane system, Sep. Purif. Technol. 17 (1999) 11–20.

[4] J.D. Figueroa, T. Fout, S. Plasynski, H. McIlvried, R.D. Srivastava, Advances inCO2 capture technology – the US Department of Energy0s Carbon Sequestra-tion Program, Int. J. Greenh. Gas Control 2 (2008) 9–20.

[5] T.C. Merkel, M. Zhou, R.W. Baker, Carbon dioxide capture with membranes atan IGCC power plant, J. Membr. Sci. 389 (2012) 441–450.

[6] Cost and Performance Baseline for Fossil Energy Plants (V 1): Bituminous Coaland Natural Gas to Electricity, Report Number DOE/NETL-2010-1397, Septem-ber 2013.

[7] R. Bredesen, K. Jordal, O. Bolland, High-temperature membranes in powergeneration with CO2 capture, J. Membr. Sci. 43 (2004) 1129–1158.

[8] Y. Ma, I.P. Mardilovich, E.E. Engwall, Thin composite palladium and palladium/alloy membranes for hydrogen separation, Ann. N. Y. Acad. Sci. 984 (2003)346–360.

[9] Y.H. Ma, Hydrogen separation membranes, in: N.N. Li, A.G. Fane, W.S. Ho,T. Matsuura (Eds.), Advanced Membrane Technology and Applications, JohnWiley & Sons, Hoboken, New Jersey, 2008, pp. 671–684.

[10] A.Y. Ku, P. Kulkarni, R. Shisler, W. Wei, Membrane performance requirementsfor carbon dioxide capture using hydrogen-selective membranes in integratedgasification combined cycle (IGCC) power plants, J. Membr. Sci. 367 (2011)233–239.

[11] F. Guazzone, J. Catalano, I.P. Mardilovich, J. Kniep, S. Pande, T. Wu, R.C. Lambrecht, S. Datta, N.K. Kazantzis, Y.H. Ma, Gas permeation field tests ofcomposite Pd and Pd–Au membranes in actual coal derived syngas atmo-sphere, Int. J. Hydrogen Energy 37 (2012) 14557–14568.

[12] P.K. Liu, M. Sahimi, T. Tsotsis, Process intensification in hydrogen productionfrom coal and biomass via the use of membrane-based reactive separations,Curr. Opin. Chem. Eng. 1 (2012) 342–351.

[13] T.S. Chung, L. Shao, P.S. Tin, Surface modification of polyimide membranes bydiamines for H2 and CO2 separation, Macromol. Rapid Commun. 27 (2006)998–1003.

[14] V.I. Bondar, B.D. Freeman, I. Pinnau, Gas transport properties of poly(ether-b-amide) segmented block copolymers, J. Polym. Sci.: Part B: Polym. Phys. 38(2000) 2051–2062.

[15] H. Lin, E. Van Wagner, B.D. Freeman, L.G. Toy, R.P. Gupta, Plasticization-enhanced H2 purification using polymeric membranes, Science 311 (2006)639–642.

[16] H. Lin, B.D. Freeman, Materials selection guidelines for membranes thatremove CO2 from gas mixtures, J. Mol. Struct. 739 (2005) 57–74.

[17] W. Yave, A. Car, K. Peinemann, Nanostructured membrane materials designedfor carbon dioxide separation, J. Membr. Sci. 350 (2010) 124–129.

[18] W. Yave, A. Car, J. Wind, K. Peinemann, Nanometric thin film membranesmanufactured on square meter scale: ultra-thin films for CO2 capture,Nanotechnology 21 (2010) 1–7.

[19] S.R. Reijerkerk, K. Nijmeijer, C.P. Ribeiro, B.D. Freeman, M. Wessling, On theeffects of plasticization in CO2/light gas separation using polymeric solubilityselective membranes, J. Membr. Sci. 367 (2011) 33–44.

[20] K. Ramasubramanian, Y. Zhao, W. Ho, CO2 capture and H2 purification:prospects for CO2-selective membrane processes, AIChE J. 59 (2013)1033–1045.

[21] M.K. Barillas, R.M. Enick, M. O0Brien, R. Perry, D. Luebke, B.D. Morreale, TheCO2 permeability and mixed gas CO2/H2 selectivity of membranes composedof CO2-philic polymers, J. Membr. Sci. 372 (2011) 29–39.

[22] D. Grainger, M. Hagg, Techno-economic evaluation of a PVAm CO2-selectivemembrane in an IGCC power plant with CO2 capture, Fuel 87 (2008) 14–24.

[23] J. Kim, S. Ha, Y. Lee, Gas permeation of poly(amide-6-b-ethylene oxide)copolymer, J. Membr. Sci. 190 (2001) 179–193.

[24] N.P. Patel, A.C. Miller, R.J. Spontak, Highly CO2-permeable and selectivepolymer nanocomposite membranes, Adv. Mater. 15 (2003) 729–733.

[25] N.P. Patel, A.C. Miller, R.J. Spontak, Highly CO2-permeable and -selectivemembranes derived from crosslinked poly(ethylene glycol) and its nanocom-posites, Adv. Funct. Mater. 14 (2004) 699–707.

[26] C.A. Scholes, J. Bacus, G.Q. Chen, W.X. Tao, G. Li, A. Qader, G.W. Stevens, S.E. Kentish, Pilot plant performance of rubbery polymeric membranes forcarbon dioxide separation from syngas, J. Membr. Sci. 389 (2012) 470–477.

[27] H. Lin, Z. He, Z. Sun, A. Ng, R.W. Baker, T.C. Merkel, CO2-selective membranes forhydrogen production and CO2 capture. Part II: Process designs, in preparation.

[28] J.G. Wijmans, R.W. Baker, The solution-diffusion model: a review, J. Membr.Sci. 107 (1995) 1–21.

[29] B.D. Freeman, Basis of permeability/selectivity tradeoff relations in polymericgas separation membranes, Macromolecules 32 (1999) 375–380.

[30] K. Lokhandwala, I. Pinnau, Z. He, K. Amo, J.C.D. Da Costa, H. Wijmans, R. Baker,Membrane separation of nitrogen from natural gas: a case study from membranesynthesis to commercial deployment, J. Membr. Sci. 346 (2010) 270–279.

[31] I. Pinnau, L.G. Toy, Solid polymer electrolyte composite membranes for olefin/paraffin separation, J. Membr. Sci. 184 (2001) 39–48.

[32] R.W. Baker, Membrane Technology and Applications, 3rd ed., John Wiley andSons, Ltd., Chichester, UK, 2012.

[33] K.A. Lundy, I. Cabasso, Analysis and construction of multilayer compositemembranes for the separation of gas mixtures, Ind. Eng. Chem. Res. 28 (1989)742–756.

H. Lin et al. / Journal of Membrane Science 457 (2014) 149–161160

[34] J. Schwinge, P.R. Neal, D.E. Wiley, D.F. Fletcher, A.G. Fane, Spiral woundmodules and spacers: review and analysis, J. Membr. Sci. 242 (2004) 129–153.

[35] H. Lin, B.D. Freeman, Permeation and Diffusion, in: H. Czichos, L.E. Smith,T. Saito (Eds.), Springer Handbook of Materials Measurement Methods,Springer, Germany, 2006, pp. 371–387.

[36] H. Lin, S.M. Thompson, A. Serbanescu-Martin, H.G. Wijmans, K.D. Amo,K. Lokhandwala, T.C. Merkel, Dehydration of natural gas using membranes.Part I: composite membranes, J. Membr. Sci. 413–414 (2012) 70–81.

[37] D.T. Coker, T. Allen, B.D. Freeman, G.K. Fleming, Nonisothermal model for gasseparation hollow-fiber membranes, AIChE J. 45 (1999) 1451–1468.

[38] S. Kelman, H. Lin, E.S. Sanders, B.D. Freeman, CO2/C2H6 separation usingsolubility selective membranes, J. Membr. Sci. 305 (2007) 57–68.

[39] J. Northington, F. Morton, R.A. Yongue, Advanced Technology Testing at theNational Carbon Capture Center, The 29th Annual International PittsburghCoal Conference, Pittsburgh, PA, 2012.

[40] F. Morton, National Carbon Capture Center at Power System DevelopmentFacility, The U.S. DOE NETL CO2 Capture Technology Meeting, Pittsburgh, PA,2011.

[41] M. Nelson, CO2 Capture at the Kemper County IGCC Project, The U.S. DOE NETLCO2 Capture Technology Meeting, Pittsburgh, PA, 2011.

[42] T. Wu, Pre-combustion CO2 Capture at National Carbon Capture Center(NCCC), The U.S. DOE NETL CO2 Capture Technology Meeting, Pittsburgh, PA,2013.

[43] H. Lin, B.D. Freeman, Gas solubility, diffusivity and permeability in poly(ethylene oxide), J. Membr. Sci. 239 (2004) 105–117.

[44] B.D. Freeman, I. Pinnau, Separation of gases using solubility-selective poly-mers, Trends Polym. Sci. 5 (1997) 167–173.

[45] L.M. Robeson, The upper bound revisited, J. Membr. Sci. 320 (2008) 390–400.[46] L.M. Robeson, Correlation of separation factor versus permeability for poly-

meric membranes, J. Membr. Sci. 62 (1991) 165–185.[47] A. Car, C. Stropnik, W. Yave, K. Peinemann, PEG modified poly(amide-b-

ethylene oxide) membranes for CO2 separation, J. Membr. Sci. 307 (2008)88–95.

[48] H. Lin, E. Van Wagner, S.J. Swinnea, B.D. Freeman, S.J. Pas, A.J. Hill,S. Kalakkunnath, D.S. Kalika, Transport and structural characteristics of cross-linked poly(ethylene oxide) rubbers, J. Membr. Sci. 276 (2006) 145–161.

[49] S.R. Reijerkerk, M.H. Knoef, K. Nijmeijer, M. Wessling, Poly(ethylene glycol)and poly(dimethyl siloxane): Combining their advantages into efficient CO2

gas separation membranes, J. Membr. Sci. 352 (2010) 126–135.[50] C.H. Lau, S. Liu, D.R. Paul, J. Xia, Y. Jean, H. Chen, L. Shao, T.S. Chung, Silica

nanohybrid membranes with high CO2 affinity for green hydrogen purifica-tion, Adv. Energy Mater. 1 (2011) 634–642.

[51] T.C. Merkel, V.I. Bondar, K. Nagai, B.D. Freeman, I. Pinnau, Gas sorption,diffusion, and permeation in poly(dimethylsiloxane), J. Polym. Sci.: Part B:Polym. Phys. 38 (2000) 415–434.

[52] T.C. Merkel, H. Lin, X. Wei, R.W. Baker, Power plant post-combustion carbondioxide capture: an opportunity for membranes, J. Membr. Sci. 359 (2010)126–139.

H. Lin et al. / Journal of Membrane Science 457 (2014) 149–161 161