Synthesis and characterization of zeolitic imidazolate framework ZIF-7 for CO2 and CH4 separation

Microporous and Mesoporous Materials 190 (2014) 189–196

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso

Synthesis and characterization of zeolitic imidazolate framework ZIF-7 for CO2 and CH4 separation Xiaofei Wu a, Mahdi Niknam Shahrak a,b, Bin Yuan a, Shuguang Deng a,c,⇑ a

Chemical Engineering Department, New Mexico State University, Las Cruces, NM 88003, USA Department of Chemical Engineering, Quchan University of Advanced Technologies, Quchan P.O. Box 84686-94717, Iran c Key Laboratory of Biomass Chemical Engineering of the Ministry of Education, Department of Chemical and Biological Engineering, Zhejiang University, Hangzhou 310027, China b

a r t i c l e

i n f o

Article history: Received 1 October 2013 Received in revised form 29 January 2014 Accepted 5 February 2014 Available online 14 February 2014 Keywords: ZIF-7 Adsorption CO2/CH4 separation Gate-opening

a b s t r a c t There are surging demands to separate carbon dioxide from methane in upgrading low-quality natural gas including biogas with an energy efficient process. In this work, adsorption equilibria and kinetics of CO2 and CH4 on a zeolitic imidazolate framework ZIF-7 were determined at various temperatures and gas pressures up to 1 bar, to evaluate the feasibility of removing CO2 from CH4 in a vacuum swing adsorption process using ZIF-7 as adsorbent. The as-synthesized ZIF-7 sample was characterized with scanning electron microscopy for crystal structure, powder X-ray diffraction for phase structure, thermal gravimetric analysis for thermal stability, and carbon dioxide adsorption at 0 °C for pore textural properties. To study the gate-opening of ZIF-7, CO2 adsorption was measured at five different temperatures (273, 298, 323, 348 and 373 K) and pressures up to 10 bar. A thermodynamic model was employed to estimate the free energy of phase transition between narrow pore and large pore, and a modified dualsite Sips equation could fit the S shape CO2 isotherms well. Adsorption equilibrium selectivity (a) and adsorbent selection parameter for pressure swing adsorption processes (S) were calculated from the adsorption equilibrium data. The relatively high values of adsorption selectivities suggest that ZIF-7 is a promising adsorbent candidate for CO2/CH4 separation in a pressure swing adsorption process. Ó 2014 Elsevier Inc. All rights reserved.

1. Introduction Natural gas, which consists primarily of methane, with impurities such as CO2, N2 and heavier hydrocarbons, is expected in great demand in the coming years due to the relatively clean-burning quality, compared to the other fuel source, like gasoline [1,2]. In addition, methane from coal beds, coalmines, and landfill gas becomes a fast growing source of natural gas, but the problem is that it often contains unacceptable levels of contaminants [3]. For example, Landfill gas is about 40–60% methane, with the remainder being mostly carbon dioxide (CO2). Therefore, an energy efficient process separating CO2 from CH4 is required for the wide utilization of natural gas in order to improve the purity and also to prevent the corrosion of equipment and pipelines caused by the acid CO2 gas [4,5]. Several technologies, such as absorption, cryogenic distillation, membrane separation, and adsorption are applied for the separa⇑ Corresponding author at: Chemical Engineering Department, New Mexico State University, Las Cruces, NM 88003, USA. Tel.: +1 575 646 4346; fax: +1 575 646 7706. E-mail address: [email protected] (S. Deng). http://dx.doi.org/10.1016/j.micromeso.2014.02.016 1387-1811/Ó 2014 Elsevier Inc. All rights reserved.

tion of CO2 from CH4 [6–8]. Of all these approaches, adsorptionbased methods such as pressure-swing adsorption (PSA), vacuum swing adsorption (VSA), and temperature swing adsorption (TSA), are promising due to the inherent simplicity, ease of control, low cost, low energy demand and high regeneration rate [8,9]. Various adsorbents have been considered for CO2 separation and capture, especially for zeolites [3,4,10] and carbon-based adsorbents [9–11]. As a new class of crystalline porous materials, metal–organic frameworks (MOFs), has been developed into one of the most prolific areas of research in chemistry and materials in the past decade, and of course, has stimulated many research interests. MOFs are inorganic–organic hybrid materials with one-, two-, or threedimensional structures, which are comprised of single metal ions or polynuclear metal clusters linked by organic ligands principally through coordination bonds [12]. Compared to other traditional porous materials, such as zeolites or carbon-based adsorbents, one major advantage of MOFs is that by wisely choosing metalbased building blocks and modular synthesis of organic linkers [13], the structures and properties of MOFs can be well controlled and designed [14,15], to meet different demands, like gas separations and storage, sensing, and catalysis [12,16–23]. Recently, a

190

X. Wu et al. / Microporous and Mesoporous Materials 190 (2014) 189–196

subclass of MOFs named zeolitic imidazolate frameworks (ZIFs) with structures having zeolite framework topologies in which all tetrahedral atoms are transition metals (Co, Cu, Zn, etc.), and all bridging ones are imidazolate (IM) units, has emerged [24,25]. ZIFs soon become one of the hottest research areas because of the highly combining properties from both zeolites and MOFs [26], such as microporosity, high surface areas, and exceptional thermal and chemical stability [24], making them ideal candidates for gas storage [27,28] and separation [29,30], especially in membrane separation [31,32]. Zeolitic imidazolate framework ZIF-7 (zinc benzimidazolate), as one important representative of ZIFs, was first discovered by Huang et al. [33] at 2003. Six-membered ring (6MR) pore structure is formed by connecting Zn metal clusters through benzimidazole (BIM) linkers. According to the calculation results, it has a largest cavity diameter of 5.579 Å and a pore limiting diameter of 2.371 Å [34]. The size of cage is comparable with kinetic diameters of CO2 (3.3 Å) and CH4 (3.8 Å), which suggests that ZIF-7 may be promising in separating CO2 and CH4. So far, most research work of ZIF-7 focus on the application in membrane separations, including separating H2/CO2 [35–38], H2/N2 [38,39], H2/CH4 [38], CO2/N2 [40], CO2/CH4 [40]. Aguado and co-workers [41] demonstrated a thermodynamic analysis to describe a breathing phenomenon for ZIF-7 upon CO2 adsorption. Morris and co-workers [42] carried out a combined experimental and computational study on the effect of topology on CO2 adsorption in ZIFs, including ZIF-7. ZIF-7 was also found promising in separating light alkane/alkene mixtures [43,44]. But there’s a lack of comprehensive study of CO2 and CH4 adsorption on ZIF-7. For example, gas adsorption kinetics is a very important aspect to design a PSA process, while no adsorption kinetic data of CO2 and CH4 on ZIF-7 could be found so far, and heats of adsorption of both these two gases calculated based on experimental data have not been reported neither. On the other hand, the investigations of gate-opening of ZIF-7 upon CO2 were limited either in narrow temperature range [42] or low pressure [41]. Here, we reported the pure gas adsorption equilibria and kinetics of CO2 and CH4 on ZIF-7 using a volumetric analyzer pressure up to 1 bar and temperatures at 273, 298 and 323 K. Intracrystalline diffusivities of CO2 and CH4 were obtained from the adsorption kinetics data, and the heats of adsorption were determined from the adsorption equilibrium data at different temperatures. Adsorptive separation selectivities of CO2/CH4 were calculated based on single pure gas adsorption data. These data will provide necessary information for PSA process design for CO2/CH4 separation using ZIF-7 as adsorbent and help us understand adsorption fundamentals. In order to better understand the gate-opening mechanism of ZIF-7, we also extended CO2 adsorption up to 10 bar and temperature up to 373 K.

2. Experimental 2.1. Synthesis of ZIF-7 The ZIF-7 was successfully synthesized by conventional hydrothermal method, following the procedures given elsewhere [24]. All chemicals, zinc nitrate hydrate (Zn(NO3)26H2O, 98%, Fluka), benzimidazole (C7H6N2, 98%, Aldrich), N,N-dimethylformamide (DMF) (99%, Aldrich), and methanol (CH3OH, 99%, Aldrich), were used as received without further purification. A solid mixture of Zn(NO3)26H2O (0.8025 g, 2.7 mmol) and benzimidazole (0.2347 g, 2 mmol) were dissolved in 75 ml DMF under sonication for 10 min. The homogeneous solution was then transferred to a 125-mL Teflon lined stainless-steel autoclave. The autoclave was capped tightly and heated to 130 °C with a heating rate of 5 °C/min in an oven. After the reaction under the autog-

enous pressure for 48 h, the samples were then removed from the oven and allowed to cool to the room temperature. The mother liquor was then carefully decanted from the product and replaced with methanol. Fresh methanol was used to exchange the DMF for 48 h at room temperature. After decanting the extra methanol and drying in air for 24 h, white crystals was obtained. The guest molecules incorporated in the crystals were removed under a dynamic vacuum at 150 °C for 12 h. 2.2. Material characterization To characterize and analyze produced ZIF-7 sample, powder Xray diffraction (PXRD), scanning electron microscopy (SEM) images and thermal gravimetric analysis (TGA) methods were employed. The XRD pattern was recorded using a Rigaku Miniflex-II X-ray diffractometer with Cu Ka (k = 1.5406 Å) radiation, 30 kV/15 mA current, and kb-filter. A step scan with an increment of 0.02° in 2h and a scan rate of 1°/min was employed to obtain the high-resolution patterns. ZIF-7 sample for SEM analysis was coated with a thin layer of gold using as putter coater. A thermogravimetric analyzer (Perkin Elmer, Pyris 1) was used to get the TGA curve with sample held in a platinum pan in a continuous flow nitrogen atmosphere. Heating rate was 10 °C/min during the measurements. Also CO2 adsorption and desorption isotherm at 0 °C was employed to determine pore textural properties including the specific Brunauer–Emmet–Teller (BET) surface area, pore volume and pore size distribution by using a Micromeritics ASAP 2020 adsorption porosimeter. 2.3. Adsorption measurements The adsorption isotherms of CO2 and CH4 on ZIF-7 sample at three temperatures (273, 298, and 323 K) and gas pressure up to 800 mmHg were measured volumetrically in the Micromeritics ASAP 2020 adsorption apparatus. CO2 adsorption at high pressure was measured volumetrically in the Micromeritics ASAP 2050 adsorption apparatus. All temperatures were achieved by using a Dewar with a circulating jacket connected to a thermostatic bath with a precision of ±0.01 °C. About 0.1 g of adsorbent sample was used for the gas adsorption studies. The initial degassing process was carried out at 150 °C for 12 h under a 0.0001 mmHg vacuum pressure. The helium gas was used to determine the free space of the system. The degas procedure was repeated on the same sample between measurements at 150 °C for 6 h. Ultrahigh purity grade CO2, CH4, and He from Matheson Co. were used as received. When the adsorption equilibrium data were collected, the kinetic adsorption characteristics were measured at the same time by a previously described method [45–51]. At a given dose, the changes in gas pressure and adsorption volume as a function of time were recorded and converted into the uptake profiles, generating the adsorption kinetics, and the final adsorption amount at the terminal pressure determined the adsorption equilibrium amount at a given pressure. 3. Results and discussion 3.1. Physical properties of ZIF-7 samples Fig. 1 shows the scanning electron microscopy (SEM) image of the ZIF-7 sample prepared in this work. It can be seen that the crystals of ZIF-7 are in cubic or rectangular shape, with length 5–10 lm, which are larger than Li’s result but with very similar morphologies [38]. The phase structure of the ZIF-7 sample was confirmed by the powder X-ray diffraction patterns shown in Fig. 2. The close match of the locations of the main peaks (2h) from

191

X. Wu et al. / Microporous and Mesoporous Materials 190 (2014) 189–196

Fig. 1. Scanning electronic micrograph of the as-synthesized ZIF-7 crystals.

simulated as-synthesized ZIF-7

200

one weight-loss step corresponds to the decomposition of framework at about 550 °C, indicating the high thermal stability. The as-synthesized ZIF-7 sample was degassed in situ at 150 °C under a vacuum (0.0001 mmHg) for 12 h before adsorption measurements in order to ensure the micro-channels in the structure were guest-free. In order to determine pore textural properties including the specific Brunauer–Emmet–Teller (BET) surface area, Langmuir surface area, pore volume and pore size distribution, CO2 adsorption and desorption isotherms on ZIF-7 sample at 0 °C were measured in an ASAP-2020 adsorption apparatus (Micromeritics). The pore size distribution of ZIF-7 as shown in Fig. 4 was obtained by correlating the desorption isotherm data with the Horvath–Kawazoe model built in the ASAP-2020 software. The median pore diameter is found to be 3.9 Å, which is close to the theoretical value (the hexagonal window size in the SOD cage estimated from crystallographic data is about 0.3 nm) [52]. The pore textural properties of ZIF-7 sample are summarized in Table 1. The BET surface area we got here (312 m2/g) is a little bit smaller than Li’s result (BET surface area is 362 m2/g, calculated by nitrogen adsorption at 77 K) [40] and simulated result (BET surface area is 405 ± 20 m2/g) [42].

150

Intensity

3.2. Adsorption isotherms 100

50

0 10

20

30

Fig. 2. Comparison of the experimental XRD patterns of as-synthesized ZIF-7 sample (red, bottom) along with the simulated pattern (black, top) using the single X-ray crystal structure data. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

simulated and the experimental XRD pattern suggests that ZIF-7 sample synthesized in this work has the correct phase structure with a good crystallinity. The small crystalline impurity phase could be attributed to the amorph-ish behavior of the XRD pattern or very small amount of ZnO. Furthermore, Fig. 3 gives the TGA curve of the ZIF-7 sample. Because the guest molecules incorporated in the samples were removed in advance, so there is only

The adsorption isotherms of CO2 and CH4 at three temperatures (273, 298 and 323 K) on ZIF-7 sample are plotted in Fig. 5a and b, respectively. All adsorption isotherms of CH4 are of typical Langmuir type and can be well correlated with the Langmuir adsorption isotherm model. While, adsorption isotherms of CO2 show a step change during CO2 uptake with an important hysteresis, which are in well agreement with the work done by Aguado et al. [41]. In order to fit these S shape isotherms, a modified dual-site Sips equation (Langmuir–Freundlich equations) which combined Langmuir equation was proposed as: 1=n

q¼a

ð1Þ

where q is the amount adsorbed of the pure component in mole per unit mass (mmol/g), P is the pressure of the bulk gas at equilibrium (kPa), qm,A, qm,B, and a (mmol/g) are the maximum loading capacities at adsorption sites A, B and C of the adsorbent, bA, bB, and b (kPa1) are the affinity parameters for sites A, B and C, nA and nB are solid heterogeneity parameters for sites A and B. The dash lines and solid lines in Fig. 5 represent the Langmuir and modified dual-site Sips isotherm model using the equation

100

0.18

90

Smoothed dV/dw (cm³/g·Å)

0.15

80

Weight (%)

1=n

bP ðbA PÞ A ðbB PÞ B þ qm;A þ qm;B 1=nA 1=n 1 þ bP 1 þ ðbA PÞ 1 þ ðbB PÞ B

70 60 50 40

0.12

0.09

0.06

0.03

0.00

30 0

200

400

600

800

Temperature (°C) Fig. 3. TGA trace of as-synthesized ZIF-7 sample.

1000

3.0

3.5

4.0

4.5

5.0

Pore Width (Å) Fig. 4. H–K pore size distribution of the ZIF-7 sample.

192

X. Wu et al. / Microporous and Mesoporous Materials 190 (2014) 189–196

Table 1 Textural properties of ZIF-7 sample.

a

BET surface area (m2/g)

Langmuir surface area (m2/g)

H–K median pore diameter (Å)

Maximum pore volumea (cm3/g)

312

355

3.9

0.0846

At P/Po = 0.030.

3.0

(a)

CO2 uptake (mmol/g)

2.5

3.3. Gate-opening

2.0

273 K 298 K 323 K

1.5

1.0

0.5

0.0 0

20

40

60

80

100

120

Pressure (kPa)

(b)

0.4

CH4 uptake (mmol/g)

on ZIF-7 at pressure around 1 bar at 298 K is found to be 2.34 mmol/g (10.3 wt.%), showing modest CO2 capacity, compared to other adsorbents for CO2 capture (At 1 bar and 298 K, the adsorption amount of CO2 on Cu3(BTC)2, MOF-177, and MOF-505 is found to be 4.1, 0.8, and 3.3 mmol/g, respectively [53]. The CO2 uptake on zeolite 5A and MOF-5 is about 20.8 and 4 wt.%, respectively [51]). In contrast, the adsorption capacity of CH4 is only 0.13 mmol/g, under the same condition. Compared to the obvious hysteresis of CO2, the adsorption and desorption branches for CH4 almost coincide, implying that it is more difficult for CO2 to diffuse out of the pores and faster kinetics for CH4 is expected.

273 K 298 K 323 K

0.3

The reversible gate-opening upon changes in temperature and CO2 partial pressure on ZIF-7 is worth noticing. Similar gate-opening phenomena have also been observed on several light hydrocarbons (ethane, ethylene, propane, propylene, butane and butene) isotherms on ZIF-7 [43,44], which are the common features of flexible frameworks [54]. In order to better understand this get-opening mechanism, we extended CO2 adsorption up to 10 bar and 373 K as shown in Fig. 6. Close inspection of these CO2 adsorption isotherms at different temperatures reveals that all of them experience a single-step gate-opening, at which reach saturation loadings at about 2.5 mmol/g, and keep increasing slowly after that. The opening and closing pressures could then be defined as the pressure at which the loading equals 1.2 mmol/g-half the adsorption/desorption step [43]. As displayed in Table 3, the opening pressure dramatically increases as increasing temperature, while the closing pressure remain almost the same except the case at

0.2

7

273 K 298 K 323 K 348 K 373 K

6

CO2 uptake (mmol/g)

0.1

0.0 0

20

40

60

80

100

120

Pressure (kPa) Fig. 5. Adsorption isotherms of CO2 (a) and CH4 (b) on ZIF-7; filled symbols, adsorption; open symbols, desorption; solid line, modified dual-site Sips equation; dash line, Langmuir equation.

5 4 3 2 1 0

parameters listed in Table 2(a and b). Because of the small uptake amount, the adsorption data points of CH4 at 323 K are relatively scattered. Other than that, from Fig. 5 and the high R2 in the table, we can see both the models fit the isotherms quite well. As shown in Fig. 5, neither gas reaches its saturation loading over the pressure range examined. The adsorption amount of CO2

1

10

100

1000

Pressure (kPa) Fig. 6. CO2 adsorption on ZIF-7 at 273 K (black square), 298 K (blue triangle), 323 K (red circle), 348 K (olive diamond) and 373 K (magenta star) up to 10 bar. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 2 Equation parameters for the CH4 isotherms (the Langmuir equation) (a), and CO2 isotherms (modified dual-site Sips equation) (b). T (K)

am (mmol/g)

b (kPa1)

R2

273 298 323

1.486 0.287 0.020

0.00271 0.00753 0.02847

>0.999 0.997 0.95

T (K)

a (mmol/g)

b (kPa1)

qm,A (mmol/g)

bA (kPa1)

nA

qm,B (mmol/g)

bB (kPa1)

nB

R2

273 298 323

1.7207 0.50316 –

0.00279 0.03173 –

0.09248 0.74482 0.37752

0.03759 0.01377 0.01953

0.033 0.286 0.762

1.44604 1.36212 1.0727

0.04919 0.01732 0.0067

0.067 0.066 0.205

>0.999 >0.999 >0.999

193

X. Wu et al. / Microporous and Mesoporous Materials 190 (2014) 189–196 Table 3 The opening and closing pressures of CO2 adsorption/desorption isotherms. T (K)

pads (kPa) at 1.2 mmol/g

pdes(kPa) at 1.2 mmol/g

273 298 323 348 373

20 58 140 240 406

11 36 35 40 45

DF HOST ¼ RT½N max lnð1 þ K II pÞ  Nmax lnð1 þ K I pÞ II I

273 K. Also, different guest gases would result in different gateopening pressures on ZIF-7 [43,44]. Very useful work has been done on gate-opening of ZIF-7 [41,43]. Aguado and co-workers [41] demonstrated a thermodynamic analysis of a reversible phase-to-phase transformation from a narrow pore (np) to a large pore (lp) phase of ZIF-7. Below the gate-opening pressure, the well fit of the Langmuir adsorption isotherm model shows that ZIF-7 adsorbs CO2 in a similar way to CH4 adsorption, indicating that the gates are entirely closed until the gate-opening pressure reaches, which is similar with that of MIL53 [55]. The interaction between the specific adsorbate and the BIM linker is responsible for the gate-opening effect [56,57]. The free-energy difference of a unit cell (DFHost) of the ZIF-7 (host) between two possible phases, before and after opening is calculated by using the CO2 adsorption isotherms and the phase transition pressure based on the equation originally derived by Coudert et al. [58] and then improved by Bergh et al. [43]:

DXOS ¼ DF HOST þ pDV 

RT½Nmax II

lnð1 þ K II pÞ 

Nmax I

omitted because of its very small contribution. Obviously, after applying above assumptions the following Eq. (3) would be easily applicable for DF calculations:

As displayed in Table 4, the estimated free energy increases as increasing temperature. The values are reasonable compared to the reported results for different adsorbates [43]. Furthermore, both energy and entropy differences also could be estimated from the following Eq. (4) for the ZIF-7 structure between two phases [58]:

DF HOST ðTÞ ¼ DU HOST ðTÞ  T DSHOST

þ K I pÞ

ð2Þ

In this equation, DXOS and DV are the differences of osmotic potentials and unit cell volumes of the structure between two possible phases, respectively. It is worthwhile to note that the osmotic potential (Xos) is defined as the potential of a host framework to move from one phase to another phase (phase transitions) induced by the adsorption of the fluid and it could be employed for considering the relative stability between the structure phases [43,58]. Furthermore, based on two forms of pore structures open (II) and closed (I), N and K are defined as Langmuir adsorption equation constants which could be distinctly obtained from fitted Langmuir equation for each phases (see Fig. S1, Supporting Information). All the Langmuir equation constants of both adsorption (closed) and desorption (open) branches of the CO2 isotherms in different temperatures are listed in Table 4. The pressures at which structural transitions occur may be determined from the experimental adsorption and desorption isotherms. Herein, we employ an average of the pressures that the desorption and adsorption branches deviate from the fitted Langmuir isotherms [58]. This pressure is calculated and shown for all isotherms in Table 4 as Ptrans. At the transition pressure, the thermodynamic equilibrium will result in the same osmotic potential for both phases. In other words, the osmotic potential of the structure in phase 1 (before opening (closed) Xos(1) (Ptrans)) is equal to phase 2 (after opening (open) Xos(2) (Ptrans)) [58]. Moreover, the second term in the right hand of the Eq. (2) (pDV or pressure multiplied by the difference in unit cell volume) can be

ð4Þ

The calculated energy and entropy differences of ZIF-7 are found to be around 22.74 kJ/mol and 0.086 kJ/(mol.K), respectively. 3.4. Isosteric heat of adsorption and Henry’s constants Isosteric heat or, isosteric enthalpy of adsorption represents the strength of the interactions between adsorbate molecules and the adsorbent lattice atoms, which is helpful to measure the energetic heterogeneity of a solid surface and evaluate the potential adsorption processes for adsorbents. The isosteric heat of adsorption (Qst) can be calculated as following procedure [48,49]. At a given amount, the isosteric heat of adsorption can be expressed by the Clausius–Clapeyron equation.

Q st ¼ RT 2

lnð1

ð3Þ

  @ ln P na @T

ð5Þ

where Qst (kJ mol1) is the isosteric heat of adsorption, P is the pressure (kPa), T is the temperature (K), R is the gas constant, and na is the adsorption amount (mmol g1). Integrating Eq. (5) as constant na, we can get

ln P ¼

Q st þC RT

ð6Þ

So under a certain na, different corresponding pressures at 273, 298, and 323 K were read from adsorption equilibrium data, then plot ln P vs 1/T, we can get the Qst. Fig. 7 shows the isosteric heat of adsorption of CO2 and CH4 with increasing loading. As we can see, the heat of adsorption of CO2 decreases with loading, while the trend for CH4 is the opposite, indicating the energetic heterogeneity of the adsorbent surface. The decreasing Qst for CO2 may due to the interaction between the quadrupole moment of CO2 with adsorbent, and the intersorbate interactions that are promoted in the restricted pore space may be used to explain the increasing Qst for CH4 [59]. At the same time, the values of Henry’s constants give an alternative way to understand the interaction between adsorbate and adsorbent. For isotherms well correlated with Langmuir adsorption isotherm model, the product of the two Langmuir constants (am and b) generates the Henry’s constant (K) directly. But for CO2 adsorption, this way does not work, so another method was chosen to calculate the Henry’s constant [48]. An isotherm model in Virial form is given by

P¼

q expðA1 q þ A2 q2 þ   Þ K

ð7Þ

Table 4 The estimated free-energy change of ZIF-7 associated with the phase transition induced by CO2 at different temperatures. T (K)

Ptrans (kPa)

KI (kPa1)

N max (mol/kg) I

KII (kPa1)

N max (mol/kg) II

DF (kJ/mol)

273 298 323 348 373

12.00 36.66 106.68 183.20 318.15

0.0773 0.0251 0.0035 0.0034 0.0025

0.670 0.614 1.350 1.290 1.000

0.127 0.039 0.023 0.022 0.043

2.964 2.988 4.990 4.350 3.950

1.75 2.09 5.18 6.21 10.43

194

X. Wu et al. / Microporous and Mesoporous Materials 190 (2014) 189–196

CO2 na (mmol/g) 0.0

0.1

0.2

0.3

0.4

and Cu-MOF (21.4 kJ/mol) [49], smaller than that of carbon molecular sieve (CMS 3 K, 33.7 kJ/mol) [61]. CO2/CH4 selectivity obtained in this work is comparable with the calculated value (8.01, charge of Zn is set as +1.0 e, Lennard– Jones parameters are set to be 100–90) [60]. The a value of 12.22 at 298 K, is more than double than that of Cu-MOF (4.9) [49] and MOF-5 (5.6) [51], but is significantly lower than that of zeolite 5A (106.6) [51] and Mg-MOF-74 (283) [50]. The pressure-dependent selectivity profile (Fig. 8) estimated from the ratio of adsorption uptakes at a given pressure indicates that CO2/CH4 selectivity also depends on the pressure greatly. The selectivity rapidly increases as the gate starts to open, and reaches the maximum value about 20.

0.5

40

Qst (kJ/mol)

35

30

25

20

3.5. Adsorption kinetics 15 0.00

0.02

0.04

0.06

0.08

0.10

CH4 na (mmol/g) Fig. 7. Isosteric heat of adsorption of CO2 (filled symbols) and CH4 (open symbols) on ZIF-7.

where P is pressure (kPa), q is adsorption amount (mmol/g), A1 and A2 are Virial coefficient, and K is the Henry’s constant (mmol/gkPa). When q or P approaches zero, Eq. (7) could be developed as

  P ¼ A1 q  ln K ln q

ð8Þ

So by plotting ln (P/q) vs q, we can get the Henry’s constant. Van’t Hoff equation then can be used to calculate the heat of adsorption at zero-coverage (Eq. S1, Fig. S2, Supporting Information). The ratio of the Henry’s constants of CO2 and CH4 yields the intrinsic thermodynamic selectivity a of CO2/CH4 on the ZIF-7 adsorbents. For pressure swing adsorption process, the adsorbent selection parameter S defined in the following equation is more useful in adsorbent evaluation and selection because it includes the ratio of adsorption capacity difference of components i and j [48,50,51]:

To explore the diffusion rate difference between CO2 and CH4, the kinetic adsorption characteristics were measured on ZIF-7 at the same time when the adsorption equilibrium data were collected by a previously described method [45–51]. As plotted in Fig. 9, both CO2 and CH4 diffuse faster at higher temperature. To better understand the kinetic behavior, the following micropore diffusion model [48–50] was used to fit the fractional adsorption uptake curves and to extract the diffusion time constant (Dc/rc2, s1):

mt 6  pffiffiffiffi ma p

sffiffiffiffiffiffiffi Dc t Dc t 3 2 r2c rc

ðmt =m1 < 0:85Þ

ð10Þ

where mt =m1 is the fractional adsorption uptake, Dc (m2/s) is the intracrystalline diffusivity of gas molecules in porous media, rc (m) is the crystal radius, t (s) is the time. This model assumes the mass transfer resistance for gas adsorption is dominated by the intracrystalline diffusion and the adsorbent crystals can be regarded as an approximately spherical object, which are reasonable on occasions where the kinetic diameter of gas is comparable with apertures. Examples are given in Fig. S3 in Supporting Information to

22

S¼

Dq i aij Dqj

ð9Þ

20 18 16

qCO2/qCH4

where Dqi and Dqj are the working capacity that is calculated as the adsorption equilibrium capacity difference at adsorption pressure and desorption pressure for components i and j, respectively. The adsorption and desorption pressures are assumed to be 100 and 1 kPa, which is a typical operating condition for vacuum swing adsorption processes. Table 5 summarizes Henry’s constants, heat of adsorption at zero-coverage and adsorption selectivities (a and S) on ZIF-7 sample. Generally, the heat of adsorption for CO2 is in the range of Sevillano’s calculation’s work [60], and it is obviously smaller than that of Mg-MOF-74 (73 kJ/mol) [50], comparable with that obtained on carbon molecular sieve (CMS 3 K, 38.9 kJ/mol) [61] and larger than that of Cu-MOF (29.7 kJ/mol) [49]. The heat of adsorption for CH4 is larger than that of Mg-MOF-74 (18.5 kJ/mol) [50]

14 12 10 8 6 4 0

20

40

60

80

100

Pressure (kPa) Fig. 8. The pressure-dependent selectivity profiles on ZIF-7 at 298 K.

Table 5 Summary of Henry’s constants, heat of adsorption at zero-coverage and adsorption selectivities. T (K)

273 298 323

Henry’s constant (mmol/gkPa) CO2

CH4

0.0634 0.0264 0.0039

0.00403 0.00216 0.00057

a

S

DHCO2 (kJ/mol)

DHCH4 (kJ/mol)

15.74 12.22 6.8

144 179 148

40.44

28.27

X. Wu et al. / Microporous and Mesoporous Materials 190 (2014) 189–196

1.0

(a)

CO2 Fractional uptake

0.8

0.6

0.4

273 K 298 K 323 K

0.2

0.0 0

100

200

300

400

500

Time (s)

1.0

(b)

CH4 Fractional uptake

0.8

195

for CO2/CH4 separation. ZIF-7 sample prepared in this work has a BET specific surface area of 312 m2/g and a median pore width about 3.9 Å, based on the CO2 adsorption/desorption at 0 °C. Adsorption equilibria, isosteric heats of adsorption, Henry’s constants, and diffusion time constants for CO2 and CH4 on ZIF-7 sample were determined experimentally. Gate-opening was studied in detail on CO2 adsorption isotherms temperature ranging from 273 to 373 K and pressure up to 10 bar. A thermodynamic model was employed to estimate the free energy of phase transition between narrow pore and large pore, and a modified dual-site Sips equation could fit the S shape isotherms well. Adsorption equilibrium selectivity (a) and adsorbent selection parameter for pressure swing adsorption processes (S) were calculated to predict the adsorption performance of ZIF-7 sample for separating CO2 and CH4. At 298 K, a is 12.22, S is 179, implying the potential to separate CO2/CH4 in a vacuum swing adsorption process using ZIF-7 as adsorbent. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.micromeso. 2014.02.016.

0.6

References 0.4

273 K 298 K 323 K

0.2

0.0 0

50

100

150

200

250

300

Time (s) Fig. 9. Fractional adsorption uptake of CO2 (a) and CH4 (b) on ZIF-7 at different temperatures.

Table 6 Summary of diffusion time constants and diffusion activation energy of CO2 and CH4 on ZIF-7. T (K)

CH4 Dc/rc2 (s1)

CO2 Dc/rc2 (s1)

CH4 Ea (kJ/mol)

CO2 Ea (kJ/mol)

273 298 323

0.01751 0.01878 0.0196

0.00911 0.01087 0.01233

1.7

4.4

demonstrate how to correlate the adsorption kinetic data shown in Fig. 9 with the diffusion models. Diffusion activation energy then can be calculated by the temperature-dependent Arrhenius equation (Eq. S3, Fig. S4, Supporting Information). Table 6 summarizes the diffusion time constants and diffusion activation energy of CO2 and CH4 on ZIF-7 at pressure (4 mmHg) and three different temperatures. The diffusion time constants of CH4 and CO2 are in the same order, indicating that a kinetic-based separation is difficult to achieve on ZIF-7 As temperature increases, these two values get closer to each other, which is in accordance with the permeability of CO2 and CH4 measured on the ZIF-7 membrane [37,52]. The values of diffusion energy are comparable with those of Mg-MOF-74 (CO2: 2.9 kJ/mol, CH4: 1.7 kJ/mol) [50], and much smaller than those obtained on CuMOF (CO2: 11.0 kJ/mol, CH4: 20.4 kJ/mol) [49]. 4. Conclusions A zeolitic midazolate framework, ZIF-7, was successfully synthesized, characterized, and evaluated as a potential adsorbent

[1] P.K. Sahoo, M. John, B.L. Newalkar, N.V. Choudhary, K.G. Ayappa, Ind. Eng. Chem. Res. 50 (2011) 13000–13011. [2] S.J. Bhadra, S. Farooq, Ind. Eng. Chem. Res. 50 (2011) 14030–14045. [3] S. Cavenati, C.A. Grande, A.E. Rodrigues, J. Chem. Eng. Data 49 (2004) 1095– 1101. [4] S. Cavenati, C.A. Grande, A.E. Rodrigues, Energy Fuels 20 (2006) 2648–2659. [5] Y.S. Bae, R.Q. Snurr, Angew. Chem. Int. Ed. 50 (2011) 11586–11596. [6] R.W. Baker, K. Lokhandwala, Ind. Eng. Chem. Res. 47 (2008) 2109–2121. [7] Y.S. Bae, K.L. Mulfort, H. Frost, P. Ryan, S. Punnathanam, L.J. Broadbelt, et al., Langmuir 24 (2008) 8592–8598. [8] D. Aaron, C. Tsouris, Sep. Sci. Technol. 40 (2005) 321–348. [9] X. Peng, W.C. Wang, R.S. Xue, Z.M. Shen, AIChE J. 52 (2006) 994–1003. [10] R. Babarao, Z.Q. Hu, J.W. Jiang, S. Chempath, S.I. Sandler, Langmuir 23 (2007) 659–666. [11] R. Van der Vaart, C. Huiskes, H. Bosch, T. Reith, Adsorption 6 (2000) 311–323. [12] J.R. Li, R.J. Kuppler, H.C. Zhou, Chem. Soc. Rev. 38 (2009) 1477–1504. [13] S. Kitagawa, R. Kitaura, S. Noro, Angew. Chem. Int. Ed. 43 (2004) 2334–2375. [14] O.M. Yaghi, M. O’Keeffe, N.W. Ockwig, H.K. Chae, M. Eddaoudi, J. Kim, Nature 423 (2003) 705–714. [15] M.J. Zaworotko, Chem. Commun. (1) (2001) 1–9. [16] L. Hamon, N. Heymans, P.L. Llewellyn, V. Guillerm, A. Ghoufi, S. Vaesen, et al., Dalton T. 41 (2012) 4052–4059. [17] L.J. Murray, M. Dinca, J.R. Long, Chem. Soc. Rev. 38 (2009) 1294–1314. [18] J.L.C. Rowsell, O.M. Yaghi, Angew. Chem. Int. Ed. 44 (2005) 4670–4679. [19] A.W.C. van den Berg, C.O. Arean, Chem. Commun. (6) (2008) 668–681. [20] S.Q. Ma, H.C. Zhou, Chem. Commun. 46 (2010) 44–53. [21] H.L. Jiang, Y. Tatsu, Z.H. Lu, Q. Xu, J. Am. Chem. Soc. 132 (2010) 5586. [22] F.X.L.I. Xamena, A. Abad, A. Corma, H. Garcia, J. Catal. 250 (2007) 294–298. [23] J. Lee, O.K. Farha, J. Roberts, K.A. Scheidt, S.T. Nguyen, J.T. Hupp, Chem. Soc. Rev. 38 (2009) 1450–1459. [24] K.S. Park, Z. Ni, A.P. Cote, J.Y. Choi, R.D. Huang, F.J. Uribe-Romo, et al., Proc. Natl. Acad. Sci. USA 103 (2006) 10186–10191. [25] X.C. Huang, Y.Y. Lin, J.P. Zhang, X.M. Chen, Angew. Chem. Int. Ed. 45 (2006) 1557–1559. [26] S.R. Venna, M.A. Carreon, J. Am. Chem. Soc. 132 (2010) 76. [27] H. Wu, W. Zhou, T. Yildirim, J. Am. Chem. Soc. 129 (2007) 5314. [28] D. Fairen-Jimenez, R. Galvelis, A. Torrisi, A.D. Gellan, M.T. Wharmby, P.A. Wright, et al., Dalton T. 41 (2012) 10752–10762. [29] J. Perez-Pellitero, H. Amrouche, F.R. Siperstein, G. Pirngruber, C. Nieto-Draghi, G. Chaplais, et al., Chem. Eur. J. 16 (2010) 1560–1571. [30] H. Amrouche, S. Aguado, J. Perez-Pellitero, C. Chizallet, F. Siperstein, D. Farrusseng, et al., J. Phys. Chem. C 115 (2011) 16425–16432. [31] M.C. McCarthy, V. Varela-Guerrero, G.V. Barnett, H.K. Jeong, Langmuir 26 (2010) 14636–14641. [32] J.A. Thompson, K.W. Chapman, W.J. Koros, C.W. Jones, S. Nair, Microporous Mesoporous Mater. 158 (2012) 292–299. [33] X.C. Huang, J.P. Zhang, X.M. Chen, Chin. Sci. Bull. 48 (2003) 1531–1534. [34] E. Haldoupis, S. Nair, D.S. Sholl, J. Am. Chem. Soc. 132 (2010) 7528–7539. [35] Y.S. Li, H. Bux, A. Feldhoff, G.L. Li, W.S. Yang, J. Caro, Adv. Mater. 22 (2010) 3322. [36] L.L. Zhang, Z.Q. Hu, J.W. Jiang, J. Phys. Chem. C 116 (2012) 19268–19277. [37] T.X. Yang, Y.C. Xiao, T.S. Chung, Energy Environ. Sci. 4 (2011) 4171–4180. [38] Y.S. Li, F.Y. Liang, H.G. Bux, W.S. Yang, J. Caro, J. Membr. Sci. 354 (2010) 48–54.

196

X. Wu et al. / Microporous and Mesoporous Materials 190 (2014) 189–196

[39] J. Yao, D. Li, K. Wang, L. He, G. Xu, H. Wang, J. Nanosci. Nanotechnol. 13 (2013) 1431–1434. [40] T. Li, Y.C. Pan, K.V. Peinemann, Z.P. Lai, J. Membr. Sci. 425 (2013) 235–242. [41] S. Aguado, G. Bergeret, M.P. Titus, V. Moizan, C. Nieto-Draghi, N. Bats, et al., New J. Chem. 35 (2011) 546–550. [42] W. Morris, N. He, K.G. Ray, P. Klonowski, H. Furukawa, I.N. Daniels, et al., J. Phys. Chem. C 116 (2012) 24084–24090. [43] J. van den Bergh, C. Gucuyener, E.A. Pidko, E.J.M. Hensen, J. Gascon, F. Kapteijn, Chem. Eur. J. 17 (2011) 8832–8840. [44] C. Gucuyener, J. van den Bergh, J. Gascon, F. Kapteijn, J. Am. Chem. Soc. 132 (2010) 17704–17706. [45] D. Saha, S.G. Deng, Langmuir 25 (2009) 12550–12560. [46] D. Saha, S.G. Deng, J. Chem. Eng. Data 54 (2009) 2245–2250. [47] D. Saha, S.G. Deng, J. Colloid Interface Sci. 345 (2010) 402–409. [48] Z.B. Bao, S. Alnemrat, L. Yu, I. Vasiliev, Q.L. Ren, X.Y. Lu, et al., Langmuir 27 (2011) 13554–13562. [49] Z.B. Bao, S. Alnemrat, L.A. Yu, I. Vasiliev, Q.L. Ren, X.Y. Lu, et al., J. Colloid Interface Sci. 357 (2011) 504–509.

[50] Z.B. Bao, L.A. Yu, Q.L. Ren, X.Y. Lu, S.G. Deng, J. Colloid Interface Sci. 353 (2011) 549–556. [51] D. Saha, Z.B. Bao, F. Jia, S.G. Deng, Environ. Sci. Technol. 44 (2010) 1820–1826. [52] Y.S. Li, F.Y. Liang, H. Bux, A. Feldhoff, W.S. Yang, J. Caro, Angew. Chem. Int. Ed. 49 (2010) 548–551. [53] A.R. Millward, O.M. Yaghi, J. Am. Chem. Soc. 127 (2005) 17998–17999. [54] S. Kitagawa, K. Uemura, Chem. Soc. Rev. 34 (2005) 109–119. [55] D.N. Dybtsev, H. Chun, K. Kim, Angew. Chem. Int. Ed. 43 (2004) 5033–5036. [56] J. Seo, R. Matsuda, H. Sakamoto, C. Bonneau, S. Kitagawa, J. Am. Chem. Soc. 131 (2009) 12792–12800. [57] G. Ferey, C. Serre, Chem. Soc. Rev. 38 (2009) 1380–1399. [58] F.X. Coudert, M. Jeffroy, A.H. Fuchs, A. Boutin, C. Mellot-Draznieks, J. Am. Chem. Soc. 130 (2008) 14294–14302. [59] J. An, S.J. Geib, N.L. Rosi, J. Am. Chem. Soc. 132 (2010) 38. [60] J.J.G. Sevillano, S. Calero, C.O. Ania, J.B. Parra, F. Kapteijn, J. Gascon, et al., J. Phys. Chem. C 117 (2013) 466–471. [61] S. Cavenati, C.A. Grande, A.E. Rodrigues, Energy Fuels 19 (2005) 2545–2555.