CH4 separation

CH4 separation

Accepted Manuscript Mixed-matrix membranes containing inorganically surface-modified 5A zeolite for enhanced CO2/CH4 separation Heqing Gong, Siew Sian...

10MB Sizes 3 Downloads 25 Views

Accepted Manuscript Mixed-matrix membranes containing inorganically surface-modified 5A zeolite for enhanced CO2/CH4 separation Heqing Gong, Siew Siang Lee, Tae-Hyun Bae PII:

S1387-1811(16)30398-5

DOI:

10.1016/j.micromeso.2016.09.017

Reference:

MICMAT 7902

To appear in:

Microporous and Mesoporous Materials

Received Date: 20 February 2016 Revised Date:

13 September 2016

Accepted Date: 13 September 2016

Please cite this article as: H. Gong, S.S. Lee, T.-H. Bae, Mixed-matrix membranes containing inorganically surface-modified 5A zeolite for enhanced CO2/CH4 separation, Microporous and Mesoporous Materials (2016), doi: 10.1016/j.micromeso.2016.09.017. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Graphical Abstract

ACCEPTED MANUSCRIPT

Mixed-matrix membranes containing inorganically surface-

RI PT

modified 5A zeolite for enhanced CO2/CH4 separation

SC

Heqing Gong,[a], [1] Siew Siang Lee,[b], [1] and Tae-Hyun Bae*[a],[b]

[a] School of Chemical and Biomedical Engineering, Nanyang Technological University,

M AN U

Singapore 637459, Singapore

[b] Singapore Membrane Technology Center, Nanyang Technological University, Singapore 637141, Singapore

TE D

* Corresponding author

These authors contributed equally to this work.

AC C

[1]

EP

Email: [email protected]

1

ACCEPTED MANUSCRIPT

Abstract LTA zeolite with highly roughened surfaces were prepared in an aqueous phase by the ion-exchange-induced growth of Mg(OH)2 nanostructures on the zeolite surfaces. The

RI PT

morphology of LTA crystals was tuned by the systematic modification of reaction parameters such as the pH of the reaction medium and the amount of magnesium loaded in the substrates. After converting surface-modified LTA to 5A, which has been known as

SC

a good candidate for selective CO2 uptake and transport, by replacing extra-framework cations remaining in LTA with Ca2+, a series of mixed-matrix membrane incorporating

M AN U

Matrimid® and 5A was fabricated. Owing to improved zeolite/polymer adhesion property along with the selective transport of CO2 by the fillers, mixed-matrix membranes containing surface modified 5A showed enhanced CO2/CH4 separation properties, which were measured by a binary mixture permeation testing. In light of that, a dramatic

TE D

increase in CO2 permeability (ca. 120 %) was observed for Matrimid®/20 wt% 5A membrane which also showed an improved CO2/CH4 selectivity. In contrast, untreated 5A decreased CO2/CH4 selectivity of Matrimid® membrane due to defects formed at

EP

zeolite/polymer interfaces.

AC C

Keywords: Mixed-matrix membrane, 5A zeolite, Mg(OH)2, Gas separation, CO2 capture

2

ACCEPTED MANUSCRIPT

Highlights

TE D

M AN U

SC

RI PT

Graphical Abstract

5A zeolite was inorganically modifed to increase surface roughness.



Selective uptake of CO2 over CH4 was observed in the surface modifed 5A.



Defect-free mixed-matrix membranes were succesfully synthesized.



CO2/CH4 separation properties were improved in mixed-matrix membranes.

AC C

EP



3

ACCEPTED MANUSCRIPT

1. Introduction The rising level of atmospheric CO2 from anthropogenic sources is one of the major environmental concerns facing the society today due to its global warming effect.

RI PT

Furthermore, separation of CO2 from gas mixtures is an essential process in many energy production processes such as natural gas treatment, biogas upgrading and hydrogen purification. Such concern and need have led to the burgeoning research and development

SC

of energy-efficient CO2 separation technology among scientists and engineers; one of

M AN U

which being the membrane-based gas separation technology [1, 2].

Mixed-matrix membranes consisting of polymers and nanoporous materials have attracted vast interest as they combine the benefits of processability and mechanical stability of polymers with the molecular sieving property of fillers [3, 4]. Among various

TE D

nanoporous materials that can be used for mixed-matrix membrane fabrications, metalorganic frameworks have recently been one of the most popular fillers owing to their large pore volume and tunable functionality [5-7]. Especially, a large number of CO2-

EP

selective mixed-matrix membranes have been designed using metal-organic framework that can selectively adsorb/transport CO2. However, metal-organic framework has a

AC C

relatively short history compared to other conventional nanoporous materials such as zeolites, such that the long-term stability of material has not yet been validated in industrial applications. Furthermore, the production cost is generally higher than that for conventional nanoporous materials. In contrast, zeolites possessing a sharp molecular sieving property owing to their rigid pore structures are currently produced in large scales with reasonable costs and used in many industrial processes [8]. In particular, LTA (Linde

4

ACCEPTED MANUSCRIPT

Type A) zeolite is a promising filler for mixed-matrix membrane used in CO2 capture due to their high CO2 uptake as well as chemical and thermal stability [9]. Nevertheless, the preparation of a high-quality mixed-matrix membranes containing zeolites has been

RI PT

greatly challenged by the formation of non-ideal polymer-zeolite interfaces. One of the main concerns is the sieve-in-a-cage (voids) morphology [3, 10-12] which is caused by the poor adhesion between organic polymer and inorganic zeolite crystals; thus yielding

SC

poor gas separation performance due to non-selective bypass of gas molecules through voids [12]. Hence, the control of polymer-zeolite interface is essential to attaining a

M AN U

successful fabrication of mixed-matrix membrane with good gas separation performance.

Various methods have been demonstrated to improve the polymer-zeolite interaction; one of which was chemical coupling of zeolite and polymer chain using a linker such as

TE D

silane coupling agent [12-15]. However, this method has been applicable to limited polymer-sieve pairs and often results in another non-ideal interfacial morphology such as the rigidified polymer chain which would adversely impact the permeability of gaseous

EP

molecules [3]. Furthermore, significant decrease in the pore volumes of the chemically treated zeolite which would impair its molecular sieving property have been reported.

AC C

Meanwhile, enhancing surface roughness of zeolite crystals via the formation of nanostructures such as Mg(OH)2 nanowhiskers could be a potential alternative to improve the polymer-zeolite interface through the adsorption and entanglement of polymer chains within the nanostructures [16, 17]. Nonetheless, the conventional Grignard treatment method which have been commonly employed to synthesize the nanostructures entailed complex procedures and sensitive reactants [18, 19]. To replace such complicated surface

5

ACCEPTED MANUSCRIPT

treatment, a facile solvothermal treatment to deposit Mg(OH)2 nanostructures on surfaces of Silicalite-1 and Na-A (or 4A) zeolites has been developed [20, 21]. Mixed-matrix membranes containing the surface modified zeolites showed an improved gas separation

RI PT

performance as a result of an enhanced polymer-zeolite interfacial property. However, the solvothermal treatment is not environmentally benign since a large amount of ethylenediamine which is corrosive and difficult to be regenerated should be used in the

SC

reaction. As an alternative, a method to grow inorganic nanostructure on the surface of Na-A in an aqueous condition has been developed [19]. This reaction was induced by

M AN U

ion-exchange of extra-framework cation of aluminosilicate zeolite, such that Mg(OH)2 nanostructure were formed only on the surface of zeolite. Defect-free membranes could be synthesized with the surface modified zeolite resulting in enhanced CO2/CH4 selectivity of Matrimid® membrane. However, owing to the low gas permeability of Na-

TE D

A, the CO2 permeability of mixed-matrix membrane was lower than that of pure Matrimid® membrane of which CO2 permeability is merely 10 barrers. Many other studies also reported the low CO2 permeability of membranes comprising zeolite Na-A

EP

(less than 10 barrers) [22, 23]. Meanwhile, a recent study showed that 5A (or Ca-A) has an even better CO2 separation property than Na-A in terms of the uptake capacity, the

AC C

binding energy, the adsorption kinetics and the selectivity over N2 since 5A has larger pores and stronger binding sites for CO2 than Na-A [9]. The results highlighted the potential of 5A for improving CO2 separation property of polymeric membranes producible in a large industrial scale.

Herein, we report defect-free Matrimid®/5A mixed-matrix membranes for enhanced

6

ACCEPTED MANUSCRIPT

CO2/CH4 separation. To improve polymer-zeolite adhesion, Mg(OH)2 nanostructures were deposited on the surfaces of 5A. To this end, the previous ion-exchange based surface treatment has been systematically modified to give the best candidate for our

RI PT

mixed-matrix membrane fabrication. A series of mixed-matrix membranes was fabricated by varying the loading of fillers and their CO2/CH4 binary mixture permeation properties

SC

were investigated.

2. Experimental

M AN U

2.1. Materials

Sodium hydroxide (NaOH, ≥98%), sodium metasilicate pentahydrate (≥95%), sodium aluminate (technical grade), calcium nitrate (Ca(NO3)2, ≥99%), magnesium chloride (MgCl2, ≥98%) and sodium nitrate (NaNO3, ≥99%) were purchased from Sigma-Aldrich

TE D

and used as received without further purification. Methanol (CH3OH) and dichloromethane (DCM, ≥99.9%) were purchased from VWR. Commercial Matrimid®

EP

was used without further treatment.

2.2. Synthesis of Na-A zeolite

AC C

Na-A or 4A zeolite was first synthesized based on the process as published by W.A. Khandayet et al [24]. 0.723 g of NaOH was dissolved in 80 ml water. After the NaOH was completely dissolved, the solution was divided into two equal volumes. One half of the solution was further added with 8.258 g sodium aluminate while the second half of the solution was mixed with 15.48 g sodium metasilicate. Both solutions were mixed gently until they become clear. Subsequently, the sodium aluminate solution was

7

ACCEPTED MANUSCRIPT

transferred into a round bottom flask. A gel will form as the silicate solution was being poured into the aluminate solution. The gel was stirred with egg shape stirrer for more than 4 h at 80 °C. After the reaction, Na-A crystals were collected by filtration and

RI PT

washed with deionized water. Finally, the zeolite particles were dried in a convection oven.

SC

2.3. Surface treatment of zeolites and transformation into 5A zeolites

In this treatment, the magnesium source for the formation of Mg(OH)2 nanostructure was

M AN U

supplied from inside the zeolite particles. To introduce Mg2+ ions to the substrates, 2 g of unactivated Na-A zeolite particles were dispersed into 100 mL of aqueous MgCl2 solutions and the mixture was stirred for 1 h at room temperature. The amounts of Mg2+ ions in zeolites were varied by using three different MgCl2 solutions of 0.1, 0.2, and 0.4

TE D

M, for the ion-exchange. Elemental analysis via EDX as summarized in Table S1 of the Supporting Information showed that the Mg/Na ratio increased with increasing concentration of Mg used. Subsequently, 0.1 g of the hydrated Mg-containing zeolites,

EP

which were not yet been activated, were added into 10 ml of 0.1 M NaNO3 solutions of which pH were adjusted to 9.5 or 11 with a dilute aqueous NaOH solution [25-27]. Then,

AC C

the mixture were transferred into a Teflon-lined autoclave and heated at 160 °C for 12 h. During this hydrothermal treatment, ion-exchange happened between Mg2+ ions introduced into the zeolites and Na+ ions in basic solution outside. As the free Mg2+ ions diffused out through micropores of the zeolites, Mg(OH)2 nanostructures were formed at the zeolite surfaces by the reaction between the free Mg2+ ions and hydroxyl ions in bulk solution. After the reaction, the zeolite crystals were collected by filtration and washed

8

ACCEPTED MANUSCRIPT

with deionized water. Subsequently, an ion-exchange process was carried out at 60 °C for 20 h with 20 mL of 0.5M calcium nitrate to replace residual extra-framework cations in zeolites with Ca2+. This ion-exchange process was repeated twice for the same substrate

RI PT

collected via filtration. Bare 5A samples were also prepared by the ion-exchange of untreated Na-A with Ca(NO3)2. In summary, four different surface-treated (ST) 5A zeolites were prepared by a systematic modification of reaction conditions which are

SC

described in Table 1.

Mg loading Samples

Mg(Cl)2

Hydrothermal treatment(a)

Stirring Duration

concentration

NaNO3

(h)

concentration

pH

0.1 M

1

0.1 M

9.5

ST2-5A

0.2 M

1

0.1 M

9.5

ST3-5A

0.4 M

1

0.1 M

9.5

1

0.1 M

11

EP

TE D

ST1-5A

ST4-5A

0.1 M

conducted at 160 oC for 12 h

AC C

(a)

M AN U

Table. 1 Reaction conditions for surface treatment of zeolites

2.4. Fabrication of membrane A solution casting technique was employed to prepare dense films of mixed-matrix membranes. 0.2 g zeolite particles (bare or surface treated 5A, hydrated) were dispersed in 2.5 g DCM in a 20 ml glass vial followed by ultrasonication for 5 minutes.

9

ACCEPTED MANUSCRIPT

Subsequently, 0.8 g of polymer (Matrimid®) was added to the suspension following which the mixture was stirred overnight with an egg-shaped stirrer to obtain a homogenous solution. Subsequently, the solution was poured on a glass plate for casting

RI PT

of a nascent film using a "doctor's knife", in a glove bag saturated with DCM vapor. The membrane film on the glass plate was left to complete evaporation in the glove bag for 1

SC

h before annealed at 230 °C for 1 d in a vacuum oven.

2.5. Characterization of zeolite particles and membranes

M AN U

The powder X-ray diffraction (XRD) of zeolite was characterized using a Bruker D2 phase diffractometer with Cu Kα radiation. N2 adsorption-desorption analysis at 77K were carried out using AutoSorb-1 Quantachrome Instrument to determine the textural properties of zeolites such as specific surface area, pore size and pore volume.

TE D

Differential scanning calorimetry (DSC) was employed to observe physical transformation of zeolite materials as a function of temperature using the Pyris Diamond Thermogravimetric Analysis. Temperature range was 50-500 °C, with an increase rate of

EP

10 °C/min and air as the purging gas. Elemental compositions of the zeolite materials were analyzed using an energy dispersive X-ray spectrometer (EDX) attached to the SEM

AC C

(FESEM, JSM-7600F). The gas uptake properties of zeolites and membranes were measured by a high-pressure gas adsorption analyzer (Quantachrome, i-Sorp HP-1). The activation of the zeolites and membranes were carried out at 200 °C for 24 h under high vacuum whereby the samples were placed into a stainless steel chamber and heated by a heating bag. Following this, pure component CO2 and CH4 isotherms were obtained in a pressure range of 0 to 1 bar at 40 °C. The morphology of the zeolites and cross-section of

10

ACCEPTED MANUSCRIPT

the membranes were observed with a field emission scanning electron microscopy (FESEM, JSM-7600F). The membranes were first cryogenically fractured in liquid nitrogen and coated with platinum prior to FESEM imaging. In the case of zeolites, the

RI PT

image was taken without any platinum coating to preserve the inorganic morphological structure. Mechanical properties of membranes were also compared with and without

SC

fillers using Instron Tester 5543 Model System.

2.6. Gas adsorption analysis

the following Langmuir model.

M AN U

To quantify the gas uptake properties, all gas adsorption isotherm data were fitted with

q sat bp (1) 1 + bp

TE D

q=

Here q is the quantity adsorbed, qsat represents the saturation loading of the adsorbent, b

EP

is the Langmuir parameter, and p is the partial pressure of gas. As the measurement of the saturation capacity of CH4 at 40 °C is difficult, it was assumed that all adsorption sites for

AC C

CO2 in zeolites are accessible to CH4. Thus, the saturation capacities for CO2 and CH4 were assumed to be equal in this work.

Then, using the fitting parameters in equation 1, Henry’s coefficients (kH) were estimated as follow,

11

ACCEPTED MANUSCRIPT

lim q = k H p = qsat bp P →0

k H = qsat b

(2)

2.7. Gas permeation test

(3)

M AN U

selectivit y = k H ,CO 2 / k H ,CH 4

SC

coefficients for CO2 over CH4.

RI PT

In this work, the sorption selectivity of CO2 over CH4 was defined as the ratio of Henry’s

Gas permeation tests were carried out using the method descried in our previous work [28]. The testing gas was CO2/CH4 (50/50) binary mixture supplied from Air Liquide. The membrane sample was first mounted onto the sample cell and then vacuum pumped

TE D

to remove all the residual gases. Then, the upper chamber of the cell was pressurized with the test gas at 1 bar. The permeate gas in the downstream was swept by He and the composition was subsequently analyzed with a gas chromatograph (GC). The

EP

permeability of individual component was determined using the flow rate of sweeping gas (He), the composition of permeated gas measured by GC, and the properties of

AC C

membrane such as effective exposed area and thickness. The operating temperature throughout the whole measurement process was fixed at 40 °C. Gas permeation test was repeated for another 2 times using 2 other samples to determine reproducibility of the experiment.

3. Results and Discussion

12

ACCEPTED MANUSCRIPT

3.1. Zeolite synthesis and characterization Figure 1 shows an SEM image of bare 5A crystals synthesized by a hydrothermal method. At given reaction conditions, 1-2 µm cubic-shaped crystals with smooth surface

TE D

M AN U

SC

RI PT

were successfully synthesized.

EP

Figure 1. FESEM image of bare LTA zeolite particles prior to surface treatment (before

AC C

ion exchange with MgCl2).

Following the surface treatment, the surfaces of LTA zeolite was transformed into highly roughened surfaces owing to the formation of Mg(OH)2 nanostructures; as clearly depicted in the FESEM images in Figure 2. A vivid morphological evolution of particles at different reaction conditions, in particular Mg loading in the substrate (LTA) and reaction pH, was observed. When the amount Mg loaded in the substrate is lowest, the surfaces of zeolite were coated with fine structure of Mg(OH)2 as shown in Figure 2a. In 13

ACCEPTED MANUSCRIPT

this condition, the growth of Mg(OH)2 was limited by the supply of Mg from the inside of substrate. However, as the amount of Mg loading in the substrate was increased by increasing concentration of aqueous MgCl2 solution from 0.1 M to 0.2 M and 0.4 M, the

RI PT

size of Mg(OH)2 structures formed on the surface of zeolite increased which could enhance its overall surface roughness and surface area (Figure 2b and c). Meanwhile, when pH increased from 9.5 to 11.0 at the same Mg loading, overall morphology was

SC

evolved into a sphere enwrapped with Mg(OH)2 nanostructures (Figure 2d). Nevertheless, it is also possible that a core-shell-like system where the ST4-5A (which

M AN U

was the core) was entirely enwrapped with a thick Mg(OH)2 shell could have resulted in the spherical shape formation of ST4-5A. In general, this observation revealed that the final products could be tuned via adjustment of pH and Mg loading to attain different

AC C

EP

TE D

desirable properties for different application need and efficiency.

14

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

TE D

Figure 2. FESEM images of surface treated 5A zeolite (ST-5A) with different loading of Mg and pH: (a) 0.1 mol/L Mg, pH = 9.5, (b) 0.2 mol/L Mg, pH = 9.5, (c) 0.4 mol/L Mg,

EP

pH = 9.5, (d) 0.1 mol/L Mg, pH = 11.0.

AC C

Figure 3 shows the XRD patterns of bare 5A and surface treated 5A zeolites which were synthesized under various Mg loadings in the substrate and pHs of reaction media. Both the untreated and treated 5A exhibited major peaks of LTA zeolite crystals [29] thus indicating that the crystallinity of zeolite was retained even after the surface treatments at 160 °C under basic conditions and not affected by different Mg loading and pH. Interestingly, Mg(OH)2 peaks were not prominent for all ST-5A samples which were synthesized at all the four different reaction conditions. The low loading of Mg(OH)2 15

ACCEPTED MANUSCRIPT

and/or the possibly amorphous or lower degree of Mg(OH)2 crystallinity relative to the 5A zeolite component could have limited the detection of Mg(OH)2 via XRD. Nevertheless, the presence of Mg was evidenced through EDX spectrum of the surface

RI PT

modified 5A (Figure S1 of the Supporting Information). In addition, analysis using DSC further verified the successful formation on Mg(OH)2 on LTA surfaces where a characteristic peak for transition of Mg(OH)2 to MgO was observed at ~380 oC (as shown

SC

in Figure S2 of the Supporting Information). In essence, observation from XRD, DSC and EDX consistently evidenced the successful surface modification of LTA with

M AN U

Mg(OH)2 without changing the crystal structure of LTA.

TE D

EP

Intensity (a.u.)

ST4-5A

AC C

ST2-5A ST1-5A

(2,2,2) (2,0,0) (2,2,0) (4,0,0)

10

ST3-5A

Bare 5A

20

30 40 2 Theta (degree)

50

60

Figure 3. Powder X-ray diffraction patterns of bare 5A and ST-5A which were synthesized under different reaction conditions.

Figure 4 depicted the N2 physisorption isotherms for bare 5A and ST-5A which evidenced 16

ACCEPTED MANUSCRIPT

the microporosity nature of 5A before and after its surface treatment. As summarized in Table 2, the reduction of micropore volume and BET specific surface area were marginal for ST1-5A which was treated at pH 9.5 using the substrate with the lowest Mg loading.

RI PT

Such small losses may be due to the increase in mass of sample by the formation of Mg(OH)2 structures which are not porous. The pore channel and window of LTA seemed to be intact in this reaction condition. Nevertheless, increased loading of Mg in the

SC

substrate had considerably reduced the porosity of resulting surface-treated zeolite due to the growing coverage of the zeolite surface by Mg(OH)2 and this was consistent with the

M AN U

FESEM images from Figure 2b-c. The microporosity of surface-treated zeolite was significantly changed by the pH of reaction medium. Even though both ST1-5A and ST45A were conducted with the same substrate containing the lowest Mg content, the micropore volume of ST4-5A treated at pH 11 were significantly lower than that of ST1-

TE D

5A. The microporosity could be reduced at the expense of higher pH when the external layers of zeolites were partially dissolved under highly alkaline condition, leading to a morphological change as shown in Figure 2d. External surface area seemed to have

EP

increased at higher Mg loading and pH given the surface roughness from the Mg(OH)2 nanostructures as corroborated by the FESEM images in Figure 1. In essence, sufficient

AC C

micropore volume is necessary to ensure molecular diffusivity of gaseous products during gas separation, thus must not be overly compromised for filler enhancement. Thus, ST1-5A was chosen for the gas adsorption study and mixed-matrix membrane fabrications.

17

ACCEPTED MANUSCRIPT

Bare 5A ST1-5A ST2-5A ST3-5A ST4-5A

250 200

RI PT

N2 volume adsorbed (cm3 S.T.P/g)

300

150

SC

100 50

0.2 0.4 0.6 0.8 Relative pressure (p/p0 )

M AN U

0 0.0

1.0

Figure 4. N2 physisorption isotherms of bare 5A and surface treated 5A zeolites which

TE D

were synthesized under different reaction conditions.

Table 2. Properties of bare and surface treated 5A zeolites calculated from N2

AC C

Zeolites

EP

physisoprtion at 77K

External surface

Micropore

area(a)

volume(a)

(m2/g) )

(cm3/g)

BET surface 2

area (m /g) )

Bare 5A

830

20

0.29

ST1-5A

742

33

0.27

ST2-5A

273

35

0.11

ST3-5A

55

55

0

18

ACCEPTED MANUSCRIPT

ST4-5A (a)

510

80

0.17

calculated by the t-plot method.

RI PT

3.2. Gas adsorption properties of zeolites

Figure 5 shows CO2 and CH4 uptake properties of bare 5A and ST1-5A zeolites measured at 40 °C. Both zeolites selectively adsorb large quantity of CO2 over CH4 at the pressure

SC

range tested. For further analysis, all isotherms were fitted with Langmuir model which successfully described all adsorption isotherms with the R2 values greater than 0.99. All

M AN U

parameters used for curve fittings, Henry’s coefficients, and sorption selectivity are summarized in Table 3. The surface treatment of 5A zeolite caused marginal decrease in both CO2 adsorption capacity and CO2/CH4 selectivity compared to those of bare 5A zeolite, as Mg(OH)2 nanostructures that couldn’t contribute to CO2 uptake were added to

TE D

the zeolite. Nevertheless, the ST1-5A still retains a good potential to improve CO2 separation property of certain polymeric membranes.

4 3

AC C

Gas uptake (mmol/g)

EP

5

CO2 Bare 5A CO2 ST1-5A

2

CH4 Bare 5A CH4 ST1-5A

1

0 0.0

0.2

0.4 0.6 Pressure (bar)

0.8

1.0

19

ACCEPTED MANUSCRIPT

Figure 5. Pure component CO2 and CH4 adsorption isotherms of bare 5A and ST1-5A measured at 40 °C, along with corresponding Langmuir fits (solid lines).

selectivities CO2

CH4

RI PT

Table 3. Fitting parameters for Langmuir model, Henry’s coefficients, and sorption

CO2/CH4

qsat

b

qsat -1

(bar )

Bare 5A

4.61

18.4

ST1-5A

4.20

12.3

(mmol/g)

selectivity

-1

(bar )

M AN U

(mmol/g)

b

SC

Zeolites

0.138

130

4.20

0.138

89

TE D

4.61

3.3. Fabrications and characterization of mixed-matrix membranes Figure 6 shows the cross-section morphology of mixed-matrix membranes made up from

EP

Matrimid®and zeolites (bare 5A and surface treated 5A). The appearance of undesirable common non-ideal polymer-zeolite interfacial morphology in particular the 'sieve-in-a-

AC C

cage' as depictedin Figure 6a and b was seen alleviated in the mixed-matrix membrane made up of Matrimid® with ST1-5A, (Figure 6c and d). This was highly likely endowed by the surface roughness and enhanced surface area of the ST1-5A which favoured physical interaction thus facilitating ST1-5A's adhesion with the polymer chain via the Mg(OH)2 nanostructures. Furthermore, owing to such improved interaction between zeolite and polymer, the surface treated zeolite particles were uniformly distributed in the polymer matrix, which is highly desirable for an improved gas separation performance. 20

ACCEPTED MANUSCRIPT

Our results are remarkable because unlike the conventional research where pore volume was usually sacrificed to improve polymer-filler interfacial morphology, the synthesis herein was successfully optimized to minimize pore volume reduction of zeolite which

AC C

EP

TE D

M AN U

SC

RI PT

facilitate gaseous diffusion while achieving better polymer-filler adhesion.

Figure 6. Cross section FESEM images of mixed-matrix membranes to show the enhanced polymer-zeolite interfacial morphology: (a) low magnification Matrimid®/bare 5A, (b) high magnification Matrimid®/bare 5A, (c) low magnification Matrimid®/ST15A, and (d) high magnification Matrimid®/ ST1-5A.

To further investigate polymer/filler adhesion property, the mechanical properties of

21

ACCEPTED MANUSCRIPT

mixed-matrix membranes were measured. Mixed-matrix membrane consisting of ST-5A exhibited greater mechanical strength than that made up from bare 5A, as shown in Table 4. Stress-strain curves for mixed-matrix membranes were included in Figure S3 of the

RI PT

Supporting Information. This was possibly rendered by the Mg(OH)2 flakes on zeolite 5A which have facilitated a stronger and more effective adhesion between Matrimid® and ST1-5A. On the other hand, with bare 5A, the voids at the polymer-zeolite interface

SC

could be the weak links which has led to the significantly lower tensile strength than even pure Matrimid® membrane. It is noteworthy that an enhanced mechanical strength

M AN U

observed for mixed-matrix membrane with ST1-5A is a highly desirable property for production and application of membranes in industrial scales.

Table 4. Mechanical strength of mixed-matrix membranes

Membranes

TE D

Tensile strength

Modulus

(MPa)

(MPa)

42.7 ± 6.7

1711 ± 134

Matrimid®/20 wt% Bare 5A

24.8 ± 4.8

1686 ± 416

Matrimid®/20 wt% ST1-5A

45.2± 6.2

1995 ± 328

AC C

EP

Matrimid®

3.4. CO2 adsorption of mixed-matrix membranes Figure 7 shows the CO2 adsorptions in pure polymer and mixed-matrix membrane containing ST1-5A. As expected, the amount of CO2 uptake in mixed-matrix membrane was significantly higher than that of pure polymer membrane owing to the addition of zeolite 5A that can capture a large quantity of CO2 as illustrated in Figure 5. However, the 22

ACCEPTED MANUSCRIPT

CO2 uptake in mixed-matrix membrane was observed to be slightly lower than its respective theoretical prediction calculated assuming the pores in all of the zeolite crystals in the membrane are available for CO2 adsorption. The partial blockage of zeolite

RI PT

pores by the polymer chain following a strong polymer-zeolite binding might possibly limit the access of CO2 to some of zeolite pores. Nevertheless, majority of zeolite pores are still available for CO2 uptake and transport as the adsorption isotherm for the mixed-

SC

matrix membranes is closer to the theoretical calculation than the CO2 isotherm of pure polymer membrane. Such enhanced CO2 uptake property along with a good

AC C

EP

TE D

membranes comprising ST1-5A.

M AN U

zeolite/polymer adhesion can render a high CO2 separation efficiency to mixed-matrix

23

ACCEPTED MANUSCRIPT

Figure 7. Pure component CO2 adsorption by Matrimid®/20 wt% bare 5A and

3.5. Gas permeation properties of mixed-matrix membranes

RI PT

Matrimid®/20 wt% ST1-5A mixed-matrix membrane measured at 40 °C

Table 5 shows CO2/CH4 binary mixture permeation properties of mixed-matrix membranes. The CO2 permeabilities were significantly improved in mixed-matrix

SC

membranes containing ST1-5A capable of rapid transport of CO2 molecules. For example, the CO2 permeability of Matrimid®/20 wt% ST1-5A membrane was observed to

M AN U

be 22.4, which is 120 % greater than that of pure Matrimid® membrane. Such dramatic improvement of CO2 permeability is highly attractive for industrial applications of Matrimid® membrane which is suffering from a low permeability in spite of its many advantages as a membrane material such as inexpensive cost, good mechanical property,

TE D

and excellent gas selectivity.

The selectivity of mixed-matrix membrane also gradually increased with increasing

EP

loading of ST1-5A in the polymer matrix owing to the selective uptake or transport of CO2 through the pores of 5A. The CO2/CH4 selectivity of Matrimid®/20 wt% ST1-5A

AC C

membrane reached 36.4, which is approximately 10 % higher than the selectivity of pure polymer. On the other hand, selectivity of mixed-matrix membrane seems to decrease with increasing bare 5A likely due to the increased presence of voids at the polymerzeolite interface which promoted more non-selective bypass of CO2 molecules. Such decreased selectivity observed for bare 5A-containing membranes indicates that our surface treatment is highly effective in improving polymer/zeolite adhesion thus

24

ACCEPTED MANUSCRIPT

minimizing the 'sieve-in-a-cage' morphology formation.

pressure with a 1:1 CO2/CH4 mixture. CO2permeability

CO2/CH4 selectivity

(barrers) 10.2 ± 0.5

Matrimid®/10 wt% bare 5A

26.7 ± 3.0

Matrimid®/20 wt% bare 5A Matrimid®/10 wt% ST1- 5A

33.6 ± 2.0

31.3 ± 2.0

31.0 ± 5.0

30.8 ± 3.0

19.6 ± 2.0

35.4 ± 2.0

22.4 ± 3.0

36.4 ± 2.0

TE D

Matrimid®/20wt% ST1- 5A

M AN U

Matrimid®

SC

Membranes

RI PT

Table 5. Mixed-gas permeation properties of membranes at 40 °C and 1 bar total feed

Since permeability depends on both solubility and diffusivity, the benefits of surfacemodified zeolite were further analyzed by observing its effect on solubility and

EP

diffusivity of the resulting mixed-matrix membrane. The solubility of CO2 at 0.5 bar, which was the upstream CO2 partial pressure for permeation testing, was obtained from

AC C

CO2 adsorption isotherms in Fig 7. Then, the diffusivity of CO2 was estimated from the solubility and the permeability of membrane. The detailed procedure for calculations of solubility and diffusivity is described in our previous work [30]. As shown in Table 6, incorporating 20 % surface modified zeolite could increase the CO2 solubility of polymer by more than 150 %. While marginal reduction in diffusivity of CO2 was inevitable due to the strong binding of CO2 onto zeolite pore surfaces, the overall improvement of CO2 permeability for mixed-matrix membrane was significant owing to CO2 solubility 25

ACCEPTED MANUSCRIPT

improvement.

Matrimid®/20 wt% ST1-5A Membranes

RI PT

Table 6. CO2 solubility, diffusivity and permeability data for Matrimid® and

CO2 permeability

CO2 solubility

CO2 diffusivity

(mol-m/m2-s-bar)

(mol/m3-bar)

(m2/s)

3.4 × 10-10

340

Matrimid®/20 wt% ST1- 5A

7.5 × 10-10

870

1.0 ×10-12

8.6 × 10-13

M AN U

SC

Matrimid®

4. Conclusions

5A zeolite with enhanced surface roughness were successfully prepared via a facile treatment in an aqueous phase through which nanostructures of Mg(OH)2 were grown on

TE D

the zeolite surfaces. The systematic modification of reaction conditions resulted in an optimized 5A zeolite (ST1-5A) for the fabrication of CO2-selective mixed-matrix membranes. The mixed-matrix membrane consisted of Matrimid® andST1-5A possessed

EP

improved polymer-zeolite interfacial morphology given by the enhanced physical

AC C

interaction between polymer and zeolite through the Mg(OH)2 nanostructures. What is more, mixed-matrix membranes exhibited a greater mechanical strength thus stability, due to the improved polymer-zeolite interaction that has also led to a significant enhancement in CO2 permeability (ca. 120 %) along with a modest increase in CO2/CH4 selectivity compared to those of pure polymer membrane. This study is remarkable as it brought about an advancement through a simplified synthesis of surface-enhanced new filler candidate for use in fabricating mixed-matrix membranes witha more efficient CO2

26

ACCEPTED MANUSCRIPT

separation.

Acknowledgements

RI PT

This work was financially supported by Academic Research Fund Tier-1 (Project reference number: RG10/13) from the Ministry of Education, Singapore and Nanyang Technological University (start-up grant for T. H. B.). We would like to thank Ms. Liu

AC C

EP

TE D

M AN U

SC

Chengyin for rendering her assistance in carrying out the membrane mechanical tests.

27

ACCEPTED MANUSCRIPT

References [1] R. Khalilpour, K. Mumford, H. Zhai, A. Abbas, G. Stevens, E.S. Rubin, Membrane-

[2] A.A. Olajire, Energy, 35 (2010) 2610-2628.

RI PT

based carbon capture from flue gas: a review, J. Clean. Prod. 103 (2014) 286-300

[3] T.-S. Chung, L.Y. Jiang, Y. Li, S. Kulprathipanja, Prog. Polym. Sci., 32 (2007) 483507.

Matsuura, Prog. Polym. Sci., 39 (2014) 817-861.

SC

[4] M. Rezakazemi, A. Ebadi Amooghin, M.M. Montazer-Rahmati, A.F. Ismail, T.

M AN U

[5] W. Li, Y. Zhang, Q. Li, G. Zhang, Chem. Eng. Sci., 135 (2015) 232-257. [6] B. Seoane, J. Coronas, I. Gascon, M.E. Benavides, O. Karvan, J. Caro, F. Kapteijn, J. Gascon, Chem. Soc. Rev., 44 (2015) 2421-2454.

[7] K. Sumida, D.L. Rogow, J.A. Mason, T.M. McDonald, E.D. Bloch, Z.R. Herm, T.-H.

TE D

Bae, J.R. Long, Chem. Rev., 112 (2012) 724-781.

[8] S. Kulprathipanja, Zeolites in industrial separation and catalysis: Chapter 1 Introduction, Weinheim : Wiley-VCH, (2010).

EP

[9] T.-H. Bae, M.R. Hudson, J.A. Mason, W.L. Queen, J.J. Dutton, K. Sumida, K.J. Micklash, S.S. Kaye, C.M. Brown, J.R. Long, Energy Environ. Sci., 6 (2013) 128-138.

AC C

[10] S. Husain, W.J. Koros, J. Membr. Sci., 288 (2007) 195-207. [11] S. Shu, S. Husain, W.J. Koros, J. Phys. Chem. C, 111 (2007) 652-657. [12] A.E. Amooghin, M. Omidkhah, A. Kargari, J. Membr. Sci. 490 (2015) 364-379. [13] O.G. Nik, X.Y. Chen, S. Kaliaguine, J. Membr. Sci., 379 (2011) 468-478. [14] A.L. Khan, A. Cano-Odena, B. Gutiérrez, C. Minguillón, I.F. Vankelecom, J. Membr. Sci., 350 (2010) 340-346.

28

ACCEPTED MANUSCRIPT

[15] X.Y. Chen, O.G. Nik, D. Rodrigue, S. Kaliaguine, Polymer, 53 (2012) 3269-3280. [16] J. Liu, T.-H. Bae, W. Qiu, S. Husain, S. Nair, C.W. Jones, R.R. Chance, W.J. Koros, J. Membr. Sci., 343 (2009) 157-163.

RI PT

[17] Shu, S. Husain, W.J. Koros, Chem. Mater., 19 (2007) 4000-4006.

[18] J.-H. Lee, P. Zapata, S. Choi, J.C. Meredith, Polymer, 51 (2010) 5744-5755.

[19] M.E. Lydon, K.A. Unocic, T.-H. Bae, C.W. Jones, S. Nair, J. Phys. Chem. C, 116

SC

(2012) 9636-9645.

[20] T.-H. Bae, J. Liu, J.S. Lee, W.J. Koros, C.W. Jones, S. Nair, J. Am. Chem. Soc., 131

M AN U

(2009) 14662-14663.

[21] T.-H. Bae, J. Liu, J.A. Thompson, W.J. Koros, C.W. Jones, S. Nair, Micropor. Mesopor. Mater., 139 (2011) 120-129.

[22] H.H. Yong, H.C. Park, Y.S. Kang, J. Won, W.N. Kim, J. Membr. Sci., 188 (2001)

TE D

151-163.

[23] O. Bakhtiari, S. Mosleh, T. Khosravi, T. Mohammadi, Sep. Sci. Technol., 46 (2011) 2138-2147.

EP

[24] W.A. Khanday, R. Tomar, Catal. Commun., 43 (2014) 141-145. [25] C. Henrist, J.P. Mathieu, C. Vogels, A. Rulmont, R. Cloots, J. Cryst. Growth, 249

AC C

(2003) 321-330.

[26] X. Lv, B. Hari, M. Li, X. Ma, S. Ma, Y. Gao, L. Tang, J. Zhao, Y. Guo, X. Zhao, Z. Wang, Colloid Surf. A 296 (2007) 97-103. [27] C. Yan, D. Xue, L. Zou, X. Yan, W. Wang, J. Cryst. Growth, 282 (2005) 448-454. [28] H. Gong, T.H. Nguyen, R. Wang, T.-H. Bae, J. Membr. Sci. 495 (2015) 169-175. [29] H. Sun, D. Wu, X. Guo, B. Shen, J. Liu, A. Navrotsky, J. Phys. Chem. C, 118 (2014)

29

ACCEPTED MANUSCRIPT

25590-25596.

AC C

EP

TE D

M AN U

SC

RI PT

[30] T.-H. Bae, J.R. Long, Energy Environ. Sci., 6 (2013) 3565-3569.

30

ACCEPTED MANUSCRIPT

Highlights 5A zeolite was inorganically modifed to increase surface roughness.



Selective uptake of CO2 over CH4 was observed in the surface modifed 5A.



Defect-free mixed-matrix membranes were succesfully synthesized.



CO2/CH4 separation properties were improved in mixed-matrix membranes.

AC C

EP

TE D

M AN U

SC

RI PT