U

YS

a

ARRAA

KMZMHFC

1

vPtftTvofHp

iogcisu

no

w

0d

Journal of Membrane Science 401– 402 (2012) 76– 82

Contents lists available at SciVerse ScienceDirect

Journal of Membrane Science

j ourna l ho me pag e: www.elsev ier .com/ locate /memsci

ltem®/ZIF-8 mixed matrix hollow fiber membranes for CO2/N2 separations

ing Dai, J.R. Johnson, Oguz Karvan, David S. Sholl ∗, W.J. Koros ∗

chool of Chemical & Biomolecular Engineering, Georgia Institute of Technology, 311 Ferst Drive NW, Atlanta, GA 30332-0100, USA

r t i c l e i n f o

rticle history:eceived 18 November 2011eceived in revised form 24 January 2012ccepted 26 January 2012vailable online 7 February 2012

a b s t r a c t

Organic–inorganic hybrid (mixed matrix) membranes can potentially extend the separation performanceof traditional polymeric materials while maintaining processing convenience. Although many dense filmsstudies have been reported, there have been few reported cases of these materials being successfullyextended to asymmetric hollow fibers. In this work we report the first successful production of mixedmatrix asymmetric hollow fiber membranes containing metal-organic-framework (MOF) ZIF-8 fillers.Specifically, we have incorporated ZIF-8 into a polyetherimide (Ultem® 1000) matrix and produced dual-

eywords:OF

IFixed matrix membranesollow fiberslue gas

layer asymmetric hollow fiber membranes via the dry jet-wet quench method. The outer separating layerof these composite fibers contains 13 wt% (17 vol%) of ZIF-8 filler. These membranes have been tested overa range of temperatures and pressures for a variety of gas pairs. An increase in separation performancefor the CO2/N2 gas pairs was observed for both pure gas and mixed gas feeds.

© 2012 Elsevier B.V. All rights reserved.

O2/N2 separation

. Introduction

The impact on global warming caused by emissions of CO2 fromarious sources is a major concern. According to the Environmentalrotection Agency, the U.S. emitted 6.1 billion metric tons of CO2 tohe atmosphere in 2007 [1]. One large source of carbon dioxide isossil fuel combustion. Alkaline-based sorbents and scrubbing solu-ions can be employed to remove CO2 from various gas mixtures.hese methods have critical drawbacks, however, due to the largeolume and low pressure of flue gas. Carbon dioxide capture meth-ds have been reviewed by a number of authors [2,3]. Membranesor the selective removal of CO2 in the presence of CO, H2, H2O, and

2S (fuel gas) or N2, O2, H2O, SO2, NO, and HCl (flue gas) would beotentially advantageous relative to other approaches [4,5].

The well-known challenges of low selectivity and permeabil-ty in polymer membranes are being addressed with some degreef success with advanced engineering polymers [6,7]. There is nouarantee, however, that these materials can be processed intoost-effective asymmetric hollow fiber membranes. The “spinnabil-ty” of a polymer is often undermined by the lack of adequatetability of the nascent fiber, which prevents extrusion and takep at economical rates [8].

In addition to challenges associated with the hollow fiber spin-ing, selectivity enhancement via the introduction of new polymersffers slow progress as is evidenced by rather marginal increases in

∗ Corresponding authors.E-mail addresses: david.sholl@chbe.gatech.edu (D.S. Sholl),

jk@chbe.gatech.edu (W.J. Koros).

376-7388/$ – see front matter © 2012 Elsevier B.V. All rights reserved.oi:10.1016/j.memsci.2012.01.044

selectivity and permeability over the previous decade [9]. In con-trast to polymers, inorganic and carbon based molecular sievesoffer very high selectivities and permeabilities. These materialsare, unfortunately, prohibitively expensive to fashion into work-ing membranes with current technology. A promising route toenhanced transport properties involves the formation of hybridor mixed matrix membranes (MMM) combining the processibilityof spinnable polymers with the excellent transport properties ofmolecular sieves. In principle, adhesion of a continuous polymermatrix to a dispersed zeolite can create a hybrid material withappropriate properties [10,11]. One class of inorganic materials thathas been considered extensively for this purpose is zeolites. Twomajor drawbacks of zeolite-based MMM must be overcome. First,the number of zeolite types identified to produce successful MMMis limited. Second, the inorganic surface chemistry of zeolites leadsto additional membrane formation challenges when attempting tocreate a defect-free morphology [12].

Metal organic frameworks (MOFs) are a relatively new class ofmicroporous materials comprised of transition metals and tran-sition metal oxides connected by organic linkages to create one-,two- and three-dimensional microporous structures. MOF parti-cles offer an attractive alternative to zeolite particles in MMMapplications. Like zeolites, MOFs have characteristically high sur-face areas, sorption capacities and selective affinities for gases.The “tailorability” or flexibility of the MOF structure introducescontrolled pore sizes, surface functionalities and chemical prop-

erties which are very desirable properties in a hybrid membranesystem [13–15]. In short, MOFs appear to be attractive fillers inMMMs [16,17]. Thousands of crystalline MOF structures have beenidentified [18,19]. Zeolitic imidazolate frameworks (ZIFs) are a class

ane Science 401– 402 (2012) 76– 82 77

ocZZetcti

icefiZmttofAontpoim

fiacilrl

mc1hspc

2

2

fifnA1w

2

cMi(w

Table 1Pure polymer dope and dual layer core and sheath dope compositions (wt%).

Component Purepolymerdope

ZIF-8/Ultem®

core dopeZIF-8/Ultem®

sheath dope

Ultem® 1000 33.00 32.06 29.82NMP 51.00 55.71 51.34THF 10.00 10.02 13.23Ethanol 4.70 – –

Y. Dai et al. / Journal of Membr

f MOFs with tetrahedral networks that resemble structure types ofonventional zeolites. As one of the many different such structures,IF-8 crystallizes in the sodalite topology (SOD) and consists ofn(II) tetrahedrally coordinated by four 2-methylimidazolate link-rs [20]. The rate-limiting step for the diffusion of guest molecules ishe passage through the six-membered windows connecting eachavity. The window size is estimated to be 3.4 A by X-ray diffrac-ion structure analysis [20,21]. ZIF-8 is currently the only ZIF thats commercially available.

ZIF-8 was predicted by Haldoupis et al. to have extraordinar-ly high membrane selectivity for CO2/CH4 mixtures in theoreticalalculations based on a rigid crystal structure [22]. Bux et al. how-ver have observed much lower actual selectivity in a pure ZIF-8lm [23]. This is consistent with the idea that internal vibrations inIF-8 are sufficient to enable the window between cages to allowolecules to pass through even though the molecules are “larger”

han the pore size that would be assigned from a rigid crystal struc-ure. Although flexibility in the windows of ZIF-8 has not beenbserved directly, Zhou et al. have shown the existence of a quasi-ree rotational potential for CH3 groups in the ZIF-8 structure [24].lthough a full description of the effect of framework vibrationsn different molecules in ZIF-8 is not available, this material isonetheless a reasonable candidate for membrane-based separa-ion because of its relatively small pores. MOFs that have only largeores (i.e. when diffusion is controlled by constrictions 1 nm acrossr larger) are unlikely to show appreciable diffusion-based selectiv-ty for CO2/N2 and are therefore not as interesting as smaller pore

aterials in membranes for this separation [22,25].There has been a growing interest in MMMs containing MOF

llers as seen in the literature reports [13,16,17,26–34]. Addition-lly, there have been a few reports in recent literature of MMMsontaining a ZIF-8 [26,27,29,31,35]; moreover, a few reports existn the open literature on asymmetric mixed matrix zeolite hol-ow fiber spinning for gas separations [28,36–42]. Only a singleeport was found on MOF/polymer asymmetric mixed matrix hol-ow fibers for H2 separations [28].

The objective of this study was to create mixed matrix asym-etric hollow fiber membranes comprising ZIF type fillers in a

ommercially available glassy polymer matrix. In this work Ultem®

000 (a polyetherimide) was chosen for the matrix because it isas acceptable intrinsic separation properties and has been useduccessfully for creating mixed matrix hollow fibers [37,38]. Thisublication, to the best of our knowledge, is the first report of suc-essful ZIF/polymer mixed matrix hollow fibers for gas separation.

. Experimental

.1. Materials

Ultem® 1000 polymer was purchased from GE Plastics (Pitts-eld, MA). Anhydrous N-methyl-2-pyrrolidione (NMP), tetrahydro-

uran (THF), iso-propanol (IPA), methanol (MeOH), hexane, lithiumitrate (LiNO3), heptane and ZIF-8 were all purchased from Sigma-ldrich (Milwaukee, WI). Polydimethylsiloxane (PDMS, Sylgard84) is produced by Dow Chemicals. All chemicals and polymersere used as received without any further purification.

.2. Mixed matrix dense membrane preparation

ZIF-8/Ultem® dense membranes were made to study the con-ept prior to spinning hollow fibers. A 10 wt% ZIF-8/Ultem® dense

MM was fabricated by solution blending. Two mixtures were

nitially prepared: Ultem® 1000 (0.45 g) powder in chloroform5 ml) and ZIF-8 (0.05 g) in chloroform (5 ml). The ZIF-8 suspensionas stirred and sonicated in a bath alternately for 1 h to obtain

LiNO3 1.30 2.21 1.17ZIF-8 – – 4.44

a homogenous dispersion. The sonication bath used for indirectsonication was a VWR Ultrasonication Water Bath, operating at120 W and 40 kHz. ZIF-8 crystals were first “primed” by adding asmall amount, approximately 20%, of the Ultem® 1000 solution tothe ZIF-8 mixture, after which the mixture was further stirred andsonicated for another 1 h. After thorough mixing, the remainingpolymer solution was added gradually by pipette. Excess chloro-form was evaporated to obtain a solution with the desired viscosity(∼10 wt% polymer) for membrane casting. Membranes were castonto a glass plate in a N2 filled glove bag containing a small dish ofchloroform using a fixed casting knife (Bird Film Applicators; Nor-folk, VA). The resulting membrane was immediately covered witha watch glass, and kept for 2 days as the chloroform evaporatedthrough a small vent into a hood. The membrane was then peeledoff the glass plate and annealed in a vacuum oven at 100 ◦C forovernight.

2.3. MMM hollow fiber membrane preparation

For high rates of gas production, the dense separating layer ofa membrane must be as thin as possible, yet strong enough towithstand the applied transmembrane pressure differential drivingforces. Such an arrangement is ideally achieved with asymmetrichollow fibers, which consist of a very thin dense layer (skin) on aporous support layer. These membranes are typically formed in asingle step via a dry jet-wet quench spinning process where phaseseparation within the wet quench bath occurs as rapidly as pos-sible. Rapid phase separation traps in significant porosity for thesupport layer of the membrane, while a thin skin can be formed inthe prior step via solvent evaporation in the air gap. Pure polymerdope preparation and the spinning procedure for the hollow fiberswere performed following previous work [38].

2.3.1. MMM hollow fiber membrane dope preparation2.3.1.1. Core dope preparation. Dried (convection oven, 110 ◦C,overnight) LiNO3 was dissolved in NMP using a sonication bath.The salt is used as a pore-forming agent in this work wherein it con-trols the phase separation kinetics and morphology of the poroussupport layer. After the salt was dissolved, THF and dried poly-mer powder (vacuum oven, 110 ◦C, overnight) were added and themixture was placed on a roll mill until the polymer was completelydissolved and the solution became transparent. An infrared heatlamp was used to aid dissolution of the polymer solution on theroll mill at 50 ◦C. Once fully dissolved, the solution was loaded intoa 500 ml syringe pump (Model 500D; Teledyne Isco, Inc.; Lincoln,NE) 1 day prior to spinning to allow for degassing. The core dopecomposition used in this work is provided in Table 1.

2.3.1.2. Sheath dope preparation. The polymer and ZIF-8 were driedat 110 ◦C in a vacuum oven overnight prior to use. ZIF-8 was

added to a portion of the required NMP and sonicated using theabove mentioned ultrasonication water bath. The mixture wasstirred and sonicated alternately for 1 h to obtain a well dispersedsuspension containing 10 wt% ZIF-8. All of the THF (as per the dope

78 Y. Dai et al. / Journal of Membrane S

Table 2Dual layer mixed matrix hollow fiber spinning parameters.

Spinning parameter Range

Core flow rate (ml/h) 150–180Sheath flow rate (ml/h) 15–18Bore flow rate (ml/h) 50–60

cdUtotad1N

2

dutepfT

22empt

2etantsTv(tt(cCaqps[

uduw(i

Air gap (cm) 1–20Draw ratio 1.7–3.5

omposition; Table 1) was added to the suspension followed by ailute polymer solution comprising the remaining NMP and enoughltem® to form rough a 13 wt% solution with the LiNO3. This mix-

ure was stirred for 15–30 min until a uniform consistency wasbserved. The remaining polymer powder was added to the mix-ure to increase dope viscosity. Once all of the components weredded, the solution was stirred for 5.5 h at 45 ◦C to completelyissolve the polymer. The prepared dope was then poured into a00 ml syringe pump (Model 100DM; Teledyne Isco, Inc.; Lincoln,E) and allowed to degas overnight prior to spinning.

.3.2. Dope preparation for single-layer fibersSingle-layer, pure polymer asymmetric hollow fibers were pro-

uced for comparison with the hybrid dual-layer fibers. The dopesed to produce these fibers was prepared in the same manner ashe core dope used for the hybrid dual-layer fibers. A minor differ-nce, though, is the addition of ethanol. This solvent was added toush the dope closer to the phase boundary and to aid in skin layerormation. Ethanol was added along with THF to the salt solution.he dope composition is provided in Table 1.

.3.3. Fiber spinning

.3.3.1. Single-layer fibers. Single-layer fibers produced via co-xtrusion of the spin dope and a bore fluid solution through aonolithic spinneret. The spinning process and conditions used to

roduce these fibers was previously reported by Husain [37]. Forhe sake of brevity, this information has not reiterated in this paper.

.3.3.2. Dual-layer fibers. Dual-layer fibers were produced via co-xtrusion of the core and sheath dopes with a bore fluid solutionhrough a composite spinneret. The solutions were extruded into

water quench bath maintained at 25 ◦C using a range of spin-ing parameters (see Table 2 for details). Syringe pumps are usedo control solution extrusion because they provide excellent flowtability and are ideal for use in laboratory spinning operations.he bore fluid consisted of 10 wt% deionized water in NMP. To pre-ent blockage of the narrow spinneret orifices, a 60 �m in-line filterSwagelok®; Solon, OH) was attached upstream of the spinneret torap large particles. Spinning was carried out at 50 ◦C by heatinghe spinneret and filter blocks using multiple silicon heating tapesOmega Engineering Inc.; Stamford, CT) regulated by temperatureontrollers (Model CN9111A; Omega Engineering Inc.; Stamford,T). The nascent membrane was extruded through an adjustableir gap into the quench bath, passed under a Teflon® guide in theuench bath and collected on a controlled speed rotating drumartially immersed in tap water. A schematic of the fiber spinningystem is available in a previous publication by Husain and Koros38].

The fibers were removed from the drum by cutting cleanlysing a sharp blade and placed in a bath of deionized water for 3ays, wherein the water was changed every day, to remove resid-

al solvent. After the third day, the fibers (approximately 75 g)ere solvent exchanged by immersing in three successive aliquots

400 ml) of methanol for 20 min, followed by 20 min immersionn three successive aliquots (400 ml) of hexane. The fibers were

cience 401– 402 (2012) 76– 82

removed from the last hexane bath and allowed to dry in the hoodfor 1 h then placed in a vacuum oven and dried for 2 h at 75 ◦C.

2.3.4. Asymmetric flat sheet preparationA small piece of flat sheet membrane was made from the sheath

dope for ZIF-8 material characterization. The dope was cast onto aglass plate and kept in the air for 5 s, and then placed in a deionizedwater bath to remove the residual solvent. The resultant membranewas air dried first and then dried in a vacuum oven 100 ◦C beforefurther characterization. Permeation studies showed these mem-branes to by highly defective, and therefore these results are notpresented here. Rather, these membranes were used for elementalanalysis (EA) and X-ray photoemission spectroscopy studies (XPS).

2.4. Material characterization

Scanning electron microscopy (SEM) images were acquiredusing a model FEG LEO-1530 microscope. The fractured fibers weresputter coated with a 10–20 nm thick coating of gold (Model P-S1Sputter Coater; ISI; Mountain View, CA). Powder X-ray diffrac-tion (XRD) patterns were recorded on a PANalytical X’Pert Pro(Almelo, The Netherlands) using CuK� radiation in the range of0.6–8◦ (40 kV, 200 mA). EA was performed by Columbia AnalyticalServices, Inc (Kelso, WA) and XPS was performed on a Nordson-Dage X-Ray XD7600NT (Westlake, OH) were use to estimate ZIF-8content in hybrid membranes.

2.5. Permeation

The permeation of a gas through the dense polymer can bedescribed using the sorption-diffusion theory. In terms of thismodel, the productivity of a membrane is defined by the perme-ability of the gas through the membrane. The permeability of a gas,i, is given by:

Pi = Di · Si (1)

where Di and Si represent the diffusion and solubility coefficientsof component i, respectively. Permeability can also be expressed asthe flux normalized by film thickness (l) and the transmembranepressure (�pi), as shown by Eq. (2):

Pi = Fluxi · �

�pi(2)

A commonly accepted unit for gas permeability is the Bar-rer, where 1 Barrer = (cm3(STP)·cm)/(cm2·s·cmHg)·10−10. Whenthe thickness is difficult to define (as is often the case with asym-metric membranes), the pressure normalized flux, or permeance(Pi/l) is used instead. In this case, we use the gas permeation unit(GPU), which is defined as 1 GPU = (cm3(STP))/(cm2·s·cmHg)·10−6.The ratio of the permeabilities can be used to signify the permselec-tivity of separation of the desired component within the mixture.The ideal selectivity of the membrane, thus, is the ratio of the per-meabilities or permeances of the individual gases. For a mixture ofgas A and B the ideal selectivity is described by:

˛A/B = PA

PB(3)

Modules containing 6 fibers of 20 cm lengths were made per themethod described by Omole [43]. All the fibers tested and reportedin this paper were post-treated. Fibers were post-treated with a2 wt% high molecular weight polydimethylsiloxane solution in hep-tanes to seal any pinhole defects in the selective skin layer [37].

Transport properties of the fibers were tested with pure gas (N2,CO2) and a mixed gas (20% CO2 in N2) obtained from Airgas andPraxair, respectively. Pure-gas permeance tests were carried outusing the constant-volume method, and at least three modules

Y. Dai et al. / Journal of Membrane Science 401– 402 (2012) 76– 82 79

17 vo

wp(tofu

3

3

offiwSamfZpst

fitZclrfiou

Fig. 1. SEM images of sonicated pure ZIF-8 (A) and

ere tested for each spinning condition. Gas compositions of theermeate in these tests were analyzed using a gas chromatographModel 450; Bruker Daltonics; Billerica, MA). A stage cut of lesshan 1% was used to avoid any nonidealities in flow distributionr concentration ambiguities. Feed pressures of 100 psia were usedor the pure gas testing and a range of pressures (20–50 psia) weresed for the mixed gas measurements.

. Results and discussion

.1. SEM

SEM was used to study the membrane cross section morphol-gy as well as probe the particle-polymer interface. Sample imagesrom both sonicated pure ZIF-8 and 13 wt% (17 vol%) ZIF-8 loadedlms are shown in Fig. 1. The crystal size of the BASF ZIF-8 sampleas around 200 nm and crystals were generally not aggregated.

EM images of the dense film membrane cross sections show homogeneous dispersion of the ZIF-8 material in the polymeratrix and good adhesion between the ZIF-8 and polymer. This uni-

orm dispersion and good adhesion indicate that the polymer andIF-8 are compatible. Furthermore, the dispersion of the sieve in theolymer matrix is quite sufficient (limited agglomerates present),uggesting that the ZIF-8/Ultem® composite membranes producedhe desired morphology.

SEMs of pure Ultem® and ZIF-8/Ultem® mixed matrix hollowber membrane are displayed in Fig. 2. Circular fibers with concen-ric bores were observed. The dual-layer fibers incorporating theIF-8 did not display a noticeable interface between the sheath andore regions, thereby indicating good adhesion between the twoayers. Limited macrovoids were observed, initiating in the core

egion of the fiber as can be seen in the SEM micrographs of selectedbers in Fig. 2, but these defects do not affect the performancef the fibers at the low feed pressures of interest. In Fig. 2C, thenusual appearance at the outer surface is an artifact of the sample

Fig. 2. SEM of pure Ultem® (A) and ZIF-8/Ultem® m

l% ZIF-8 in Ultem® composite dense films (B & C).

preparation process; regardless, well-adhered ZIF-8 crystals areapparent, similar to those the dense film samples.

3.2. XRD of MMM

In addition to SEM imaging, several characterization techniqueswere used to confirm the existence of ZIF-8 in the composite hollowfibers. XRD is a commonly used technique to verify the presence ofcrystals in amorphous matrices [44]. Spectra for a hybrid dense filmshow the presence of major peaks (compared to the simulated crys-tal spectra), signifying the preservation of the crystal in the polymermatrix. This analysis, however, was inconclusive for the compositehollow fibers (see Fig. 3). The difficulty in using XRD to prove thepresence of ZIF-8 arises from the low concentration of ZIF in themembrane. Since the sieve is only located in the thin outer sheathlayer of a dual-layer hollow fibers, the total sieve concentration isquite low; albeit, the concentration in the selective skin layer is17 vol%.

3.3. Elemental analysis and XPS of MMM

In order to verify the sieve content in the hybrid membranes, XPSand EA were used to measure the zinc (Zn) content of the hybridmaterials. Zinc is unique to the sieve and therefore can be distin-guished from the polymer matrix and used as a measure of the sievecontent in the hybrid material. A summary of the EA and XPS mea-surements is provided in Table 3. The results of these analyses arecompared with calculated theoretical calculated values which arebased on dual layer fiber dimensions and anticipated weight frac-tions. Results from EA and XPS are in reasonable agreement withtheoretical expectations. The EA results reflect the fact that EA was

performed on the entire fiber, even though ZIF-8 is only present inthe thin skin layer. The flat sheet, on the other hand, has sieve dis-persed throughout. Since EA is a destructive analytical technique,the total zinc content should be measured. Alternatively, XPS only

ixed matrix hollow fiber membranes (B & C).

80 Y. Dai et al. / Journal of Membrane Science 401– 402 (2012) 76– 82

Table 3EA and XPS results for Zn content in ZIF-8/Ultem® hybrid dual-layer hollow fiber and flat sheet membranes.

Sample EA XPS

Measured (wt%) Theoretical (wt%) Measured (wt%) Theoretical (wt%)

4 0.64 0.765 0.55 0.76

ptusZ

3

3m

c

F

Table 4Permeation properties of pure polymer dense films and hollow fibers and hybrid hol-low fiber membranes containing ZIF-8 filler. Dense film feed pressures were 30 psia.Hollow fiber feed pressures were 100 psia. All measurements were conducted at35 ◦C.

Membranea PCO2 or PCO2 /l ˛CO2/N2

Pure Ultem® dense film [37] 1.4 Barrers 25Pure Ultem®hollow fiber 14 GPU 3017 vol% ZIF-8/Ultem® hybrid hollow fiber 26 GPU 36

Ft

Dual-layer hollow fiber 0.03 0.0Flat sheet membrane 3.53 3.6

enetrates a few nanometers into the sample surface; therefore,he XPS predictions reflect the zinc content in an estimated vol-me (with an expected concentration of 17 vol% at the membraneurface). The XPS results indicate that the presence and loading ofIF-8 in the membranes is at the expected levels.

.4. Gas permeation

.4.1. Pure gas permeation properties of hybrid hollow fiberembranes

The permeation properties of the hybrid fibers are presented inomparison with pure Ultem® dense film and pure Ultem® hol-

ig. 3. XRD of ZIF-8/Ultem® mixed matrix hollow fiber and dense film membranes.

a All membranes were heat treated at 75 ◦C for 6 h prior to testing. This was done

ig. 4. Permeation properties of Ultem®/ZIF-8 composite dual layers asymmetric hollow

emperature. Each sample was tested at 100 psia and a 1–10 torr vacuum downstream.

so that each membrane had a similar thermal history because of the PDMS post-treatment process.

low fiber membrane properties in Table 4. Both pure polymer andhybrid hollow fiber cases presented here were spun using simi-lar parameters (a spin temperature of 50 ◦C, an air gap of 10 cmand fiber take-up rate of 75 m/min). The reader may note thatthe pure polymer hollow fiber has a higher selectivity than thedense film. This phenomenon is likely due to polymer chain ori-entation when the high viscosity solution is extruded throughthe narrow spinneret orifice [45]. More importantly, both per-meance and selectivity of the ZIF-8/Ultem® hybrid hollow fibersare higher than the pure polymer fiber membrane, ∼85% and 20%respectively. These results are promising, especially compared toprevious hybrid hollow fiber membrane work. Husain and Korosreported hybrid hollow fiber membranes containing surface modi-fied HSSZ-13 zeolite in an Ultem®, but actually saw a slight decreasein CO2/N2 permselectivity (albeit an improvement in CO2/CH4 andO2/N2 permselectivities) [38]. The selectivity improvement pre-sumably reflects the effects of enhancement contributed by theZIF-8 additive.

3.4.2. Gas permeation properties as a function of temperatureAdditional measurements were made at various temperatures

to determine the effect of temperature on transport properties.

fibers (�) and pure Ultem® hollow fibers (©) plotted as a function of measurement

Y. Dai et al. / Journal of Membrane Science 401– 402 (2012) 76– 82 81

Table 5Pure Ultem® and Ultem®/ZIF-8 hybrid asymmetric hollow fiber permeation properties at different measurement temperatures. All tests were conducted with a 100 psia feedpressure and a 1–10 torr vacuum downstream.

Temperature (◦C) Hybrid hollow fiber Pure polymer hollow fiber

Permeance (GPUs) Permselectivity (CO2/N2) Permeance (GPUs) Permselectivity (CO2/N2)

25 18 44 11 3630 21 39 13 3435 26 36 14 3045 34 28 20 26

F ance.

r es), 35

Timtiwae8toal

3m

cfifstmmfi

4

hc

ig. 5. Mixed gas permeation results using a feed containing 20% CO2 with a N2 balesults for mixed gas measurements. Measurements were conducted at 25 ◦C (circl

hese data are presented in Table 5 and the trends are shownn Fig. 4 for pure gases measurements. The results from these

easurements follow expected trends for solution-diffusion basedransport in glassy polymers. Permeances decrease with decreas-ng temperature; however, the ideal permselectivity increases

ith decreasing temperature. An ideal permselectivity of 44 waschieved for the CO2/N2 gas pair at 25 ◦C. This is a >20% selectivitynhancement compared to the pure polymer hollow fibers. ZIF-/Ultem® membranes are more affected by the temperature changehan the pure Ultem® polymer fibers; however, the detailed causef this sensitivity is not known. Nevertheless, the fiber is not dam-ged by the higher temperature testing and upon returning to theower temperature, the performance was regained.

.4.3. Mixed gas permeation properties of hybrid hollow fiberembranes

Mixed gas permeation was measured using a dry gas mixtureontaining 20% CO2 with a N2 balance. Multiple temperatures andeed pressures have been used to develop a greater understand-ng of material performance. Permeance and permselectivity trendsor CO2 and N2 as a function of feed pressure and temperature arehown in Fig. 5. Although the mixed gas permselectivities are lesshan ideal permselectivities reported in Table 4, an increase in per-

eance is observed in the MMMs relative to the pure polymer. Aixed gas selectivity of 32 has been achieved in these hybrid hollow

ber membranes at 25 ◦C.

. Conclusions

Organic–inorganic hybrid (mixed matrix) asymmetric dual layerollow fiber membranes were spun via a dry jet-wet quench pro-ess containing the metal organic framework sieve ZIF-8 in an

(A) Permeance of CO2 (closed symbols) and N2 (open symbols). (B) Permselectivity◦C (triangles) and 45 ◦C (squares).

Ultem® polymer matrix. The fibers reported in this work con-tained 17 vol% ZIF-8 in the selective skin layer of the asymmetricmembrane. Good adhesion was observed between the bulk poly-mer and ZIF-8 particles in the membrane. This was confirmed bythe enhancement in the permeation properties of the post-treatedhybrid membranes over the pure polymer hollow fiber membranes.Hybrid fibers showed increased permeance with permselectivityenhancements as high as 20% over pure polymer. Furthermore,mixed gas measurements revealed promising permselectivity (ashigh as 32) in the hybrid membranes. Although the performanceof the MMMs we have studied is limited by the relatively low per-meability of Ultem®, our results are evidence that the MOFs holdpromise as components in MMMs for large scale gas separations,such as CO2 capture from flue gas. Moreover, this work providesa good starting point for the growing area of MOF/polymer hybridmembranes and the extension into the hollow fiber morphology.

Acknowledgments

The authors would like to thank Dr. Yougui Huang for his assis-tance with XRD measurements, Liren Xu for his assistance with SEMimaging and Wei Long for her assistance with XPS characteriza-tion. Additionally, the authors would like to acknowledge the DOEARPA-E IMPACCT Program for financial support under contract DE-AR0000074. This publication is based in part on work supportedby Award No. KUS-I1-011-21 made by King Abdullah University ofScience and Technology (KAUST).

References

[1] US Environmental Protection Agency. http://www.epa.gov/climatechange/policy/intensitygoal.html.

8 rane S

[

[

[[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

2 Y. Dai et al. / Journal of Memb

[2] E.J. Granite, T. O’Brien, Review of novel methods for carbon dioxide separationfrom flue and fuel gases, Fuel Processing Technology 86 (2005) 1423–1434.

[3] C.M. White, B.R. Strazisar, E.J. Granite, J.S. Hoffman, H.W. Pennline, Separationand capture of CO2 from large stationary sources and sequestration in geologi-cal formations: coalbeds and deep saline aquifers, Journal of the Air and WasteManagement Association 53 (2003) 645–715.

[4] A. Kohl, F. Riesenfeld, Gas Purification, Gulf Publishing, Houston, TX, 1985.[5] C.A. Scholes, S.E. Kentish, G.W. Stevens, Effects of minor components in carbon

dioxide capture using polymeric gas separation membranes, Separation andPurification Reviews 38 (2009) 1–44.

[6] P. Bernardo, E. Drioli, G. Golemme, Membrane gas separation: a review/state ofthe art, Industrial and Engineering Chemistry Research 48 (2009) 4638–4663.

[7] W.J. Koros, R. Mahajan, Pushing the limits on possibilities for large scale gas sep-aration: which strategies? Journal of Membrane Science 175 (2000) 181–196.

[8] A. Ziabicki, Fundamentals of Fibre Formation, John Wiley & Sons, London, 1976.[9] R.W. Baker, Future directions of membrane gas separation technology, Indus-

trial and Engineering Chemistry Research 41 (2002) 1393–1411.10] R. Mahajan, W.J. Koros, Factors controlling successful formation of mixed-

matrix gas separation materials, Industrial and Engineering ChemistryResearch 39 (2000) 2692–2696.

11] T.T. Moore, R. Mahajan, D.Q. Vu, W.J. Koros, Hybrid membrane materials com-prising organic polymers with rigid dispersed phases, AIChE Journal 50 (2004)311–321.

12] D.W. Breck, Zeolite Molecular Sieves, Wiley, New York, 1973.13] T.-H. Bae, J.S. Lee, W. Qiu, W.J. Koros, C.W. Jones, S. Nair, A high-performance

gas-separation membrane containing submicrometer-sized metal–organicframework crystals, Angewandte Chemie International Edition 49 (2010)9863–9866.

14] T.-S. Chung, L.Y. Jiang, Y. Li, S. Kulprathipanja, Mixed matrix membranes(MMMs) comprising organic polymers with dispersed inorganic fillers for gasseparation, Progress in Polymer Science 32 (2007) 483–507.

15] D. Tanaka, S. Kitagawa, Captured molecules in coordination frameworks, MRSBulletin 32 (2007) 540–543.

16] R. Adams, C. Carson, J. Ward, R. Tannenbaum, W.J. Koros, Metal organicframework mixed matrix membranes for gas separations, Microporous andMesoporous Materials 131 (2010) 13–20.

17] S. Keskin, D.S. Sholl, Selecting metal organic frameworks as enabling materialsin mixed matrix membranes for high efficiency natural gas purification, Energyand Environmental Science 3 (2010) 343–351.

18] C.J. Doonan, W. Morris, H. Furukawa, O.M. Yaghi, Isoreticular metalation ofmetal–organic frameworks, Journal of the American Chemical Society 131(2009) 9492–9493.

19] A.O. Yazaydın, R.Q. Snurr, T.-H. Park, K. Koh, J. Liu, M.D. LeVan, A.I. Benin,P. Jakubczak, M. Lanuza, D.B. Galloway, J.J. Low, R.R. Willis, Screening ofmetal–organic frameworks for carbon dioxide capture from flue gas usinga combined experimental and modeling approach, Journal of the AmericanChemical Society 131 (2009) 18198–18199.

20] K.S. Park, Z. Ni, A.P. Côté, J.Y. Choi, R. Huang, F.J. Uribe-Romo, H.K. Chae, M.O’Keeffe, O.M. Yaghi, Exceptional chemical and thermal stability of zeolitic imi-dazolate frameworks, Proceedings of the National Academy of Sciences 103(2006) 10186–10191.

21] J.L.C. Rowsell, O.M. Yaghi, Metal-organic frameworks: a new class of porousmaterials, Microporous and Mesoporous Materials 73 (2004) 3–14.

22] E. Haldoupis, S. Nair, D.S. Sholl, Efficient calculation of diffusion limitations inmetal organic framework materials: a tool for identifying materials for kineticseparations, Journal of the American Chemical Society 132 (2010) 7528–7539.

23] H. Bux, C. Chmelik, J.M. van Baten, R. Krishna, J. Caro, Novel MOF-membrane formolecular sieving predicted by IR-diffusion studies and molecular modeling,Advanced Materials 22 (2010) 4741–4743.

24] W. Zhou, H. Wu, T.J. Udovic, J.J. Rush, T. Yildirim, Quasi-free methyl rotation

in zeolitic imidazolate framework-8, The Journal of Physical Chemistry A 112(2008) 12602–12606.

25] A.I. Skoulidas, D.S. Sholl, Self-diffusion and transport diffusion of light gases inmetal-organic framework materials assessed using molecular dynamics simu-lations, The Journal of Physical Chemistry B 109 (2005) 15760–15768.

[

cience 401– 402 (2012) 76– 82

26] S. Basu, A. Cano-Odena, I.F.J. Vankelecom, MOF-containing mixed-matrix mem-branes for CO2/CH4 and CO2/N2 binary gas mixture separations, Separation andPurification Technology 81 (2011) 31–40.

27] K. Diıaz, L. Garrido, M. Loıpez-Gonzalez, L.F. del Castillo, E. Riande, Co2 trans-port in polysulfone membranes containing zeolitic imidazolate frameworksas determined by permeation and PFG NMR techniques, Macromolecules 43(2010) 316–325.

28] J. Hu, H. Cai, H. Ren, Y. Wei, Z. Xu, H. Liu, Y. Hu, Mixed-matrix membrane hol-low fibers of Cu3(BTC)2 Mof and polyimide for gas separation and adsorption,Industrial and Engineering Chemistry Research 49 (2010) 12605–12612.

29] M.J.C. Ordonez, K.J. Balkus Jr., J.P. Ferraris, I.H. Musselman, Molecular sievingrealized with ZIF-8/Matrimid® mixed-matrix membranes, Journal of Mem-brane Science 361 (2010) 28–37.

30] E.V. Perez, K.J. Balkus Jr., J.P. Ferraris, I.H. Musselman, Mixed-matrix membranescontaining MOF-5 for gas separations, Journal of Membrane Science 328 (2009)165–173.

31] C. Zhang, Y. Dai, J.R. Johnson, O. Karvan, W.J. Koros, High performance ZIF-8/6FDA-DAM mixed matrix membrane for propylene/propane separations,Journal of Membrane Science (2011).

32] P.S. Goh, A.F. Ismail, S.M. Sanip, B.C. Ng, M. Aziz, Recent advances of inorganicfillers in mixed matrix membrane for gas separation, Separation and Purifica-tion Technology 81 (2011) 243–264.

33] Y. Zhang, I.H. Musselman, J.P. Ferraris, K.J. Balkus Jr., Gas permeability prop-erties of Matrimid® membranes containing the metal-organic frameworkCu–BPY–HFS, Journal of Membrane Science 313 (2008) 170–181.

34] Y. Li, F. Liang, H. Bux, W. Yang, J. Caro, Zeolitic imidazolate framework ZIF-7based molecular sieve membrane for hydrogen separation, Journal of Mem-brane Science 354 (2010) 48–54.

35] S. Basu, M. Maes, A. Cano-Odena, L. Alaerts, D. De Vos, I.F.J. Vankelecom, Solventresistant nanofiltration (SRNF) membranes based on metal-organic frame-works, Journal of Membrane Science 344 (2009) 190–198.

36] O.M. Ekiner, S.S. Kulkarni, Process for Making Mixed Matrix Hollow Fiber Mem-branes for Gas Separation, US 6,663,805 B1, L’Air Liquide Societe Anonyme aDirectoire et Conseil de Surveillance pour l’Etude et l’Exploitation des ProcedesGeorges Claude, U.S.A. (2003).

37] S. Husain, Mixed Matrix Dual Layer Hollow Fiber Membranes for Natural GasSeparation, PhD Dissertation, Georgia Institute of Technology, Atlanta, GA,2006.

38] S. Husain, W.J. Koros, Mixed matrix hollow fiber membranes made with mod-ified HSSZ-13 zeolite in polyetherimide polymer matrix for gas separation,Journal of Membrane Science 288 (2007) 195–207.

39] L.Y. Jiang, T.S. Chung, C. Cao, Z. Huang, S. Kulprathipanja, Fundamental under-standing of nano-sized zeolite distribution in the formation of the mixed matrixsingle- and dual-layer asymmetric hollow fiber membranes, Journal of Mem-brane Science 252 (2005) 89–100.

40] R. Kiyono, G.H. Koops, M. Wessling, H. Strathmann, Mixed matrix microporoushollow fibers with ion-exchange functionality, Journal of Membrane Science231 (2004) 109–115.

41] W.J. Koros, D. Wallace, J.D. Wind, S.J. Miller, C. Staudt-Bickel, D.Q. Vu,Crosslinked and Crosslinkable Hollow Fiber Mixed Matrix Membrane andMethod of Making Same, US Patent 6,755,900 B2, Chevron U.S.A. Inc. & Uni-versity of Texas System, U.S.A. (2004).

42] S.J. Miller, C.L. Munson, S.S. Kulkarni, D.J. Hasse, Purification of P-Xylene UsingComposite Mixed Matrix Membranes, US Patent 6,500,233 B1, Chevron U.S.A.Inc. & Medal, L.P., U.S.A. (2002).

43] I. Omole, Crosslinked Polyimide Hollow Fiber Membranes for Aggressive Natu-ral Gas Feed Streams, PhD Dissertation, Georgia Institute of Technology, Atlanta,GA, 2008.

44] S.S. Ray, M. Okamoto, Polymer/layered silicate nanocomposites: a reviewfrom preparation to processing, Progress in Polymer Science 28 (2003)

1539–1641.

45] A.F. Ismail, S.J. Shilton, I.R. Dunkin, S.L. Gallivan, Direct measurement ofrheologically induced molecular orientation in gas separation hollow fibremembranes and effects on selectivity, Journal of Membrane Science 126 (1997)133–137.