Selective Hysteretic Sorption of Light ... - ACS Publications

This is an open access article published under a Creative Commons Attribution (CC-BY) License, which permits unrestricted use, distribution and reproduction in any medium, provided the author and source are cited.

Article pubs.acs.org/cm

Selective Hysteretic Sorption of Light Hydrocarbons in a Flexible Metal−Organic Framework Material Shan Gao,†,‡ Christopher G. Morris,†,‡,§ Zhenzhong Lu,‡ Yong Yan,‡ Harry G. W. Godfrey,‡ Claire Murray,§ Chiu C. Tang,§ K. Mark Thomas,∥ Sihai Yang,*,‡ and Martin Schröder*,‡ †

School of Chemistry, University of Nottingham, Nottingham NG7 2RD, U.K. School of Chemistry, University of Manchester, Manchester M13 9PL, U.K. § Diamond Light Source, Harwell Science Campus, Oxfordshire OX11 0DE, U.K. ∥ Wolfson Northern Carbon Reduction Laboratories, School of Chemical Engineering and Advanced Materials, University of Newcastle upon Tyne, Newcastle upon Tyne NE1 7RU, U.K. ‡

S Supporting Information *

ABSTRACT: Porous MFM-202a (MFM = Manchester Framework Material, replacing the NOTT designation) shows an exceptionally high uptake of acetylene, 18.3 mmol g−1 (47.6 wt %) at 195 K and 1.0 bar, representing the highest value reported to date for a framework material. However, at 293 K and 10 bar C2H6 uptake (9.13 mmol g−1) is preferred. Dual-site LangmuirFreundlich (DSLF)- and Numerical Integration (NI)-based IAST methods have been used to analyze selectivities for C1 to C3 hydrocarbons. MFM-202a exhibits broadly hysteretic desorption of acetylene; such behavior is important for practical gas storage since it allows the gas to be adsorbed at high pressure but stored at relatively low pressure. Stepwise uptake and hysteretic release were also observed for adsorption of other unsaturated light hydrocarbons (ethane and propene) in MFM-202a but not for saturated hydrocarbons (methane, ethane, and propane). MFM-202a has been studied by in situ synchrotron X-ray powder diffraction to reveal the possible phase transition of the framework host as a function of gas loading. A comprehensive analysis for the selectivities between these light hydrocarbons has been conducted using both IAST calculation and dual-component mixed-gas adsorption experiments, and excellent agreement between theory and experiment was achieved.

■

INTRODUCTION The large-scale separation of hydrocarbon mixtures for the production and purification of relevant energy resources and feedstocks is an extremely energy-consuming process. Natural gas, the largest reservoir for methane (CH4), also contains ethane (C2H6), propane (C3H8), and other higher alkanes. The purification of CH4 has attracted much attention1,2 since it is an important target among hydrocarbon separations due to its wide ranging applications. Additionally, the purification of two very important industrial starting petrochemicals, ethene and propene (the former being the largest volume organic in the world), involves the removal of other light hydrocarbons (e.g., acetylene, ethane, and propane).3 The state-of-the-art separation method for light hydrocarbon mixtures is cryogenic distillation at high pressure based upon the small differences in vapor pressure for each component.4 This process is highly energy-intensive, and reductions in cost and energy consumption are required. To overcome these problems, selective adsorption by traditional porous materials (e.g., mesoporous silica, zeolites, and activated carbon) as adsorbents have been employed to separate hydrocarbon mixtures.5−8 Compared with these traditional porous materials, metal− organic framework (MOFs) materials, as a newly emerged class © XXXX American Chemical Society

of porous crystalline solids, show high adsorption capacities. 9−11 Adsorbed gas molecules often form specific intermolecular interaction with open metal and other surface sites within MOFs, leading to the selective adsorption of certain gas species and substrates.3,12−14 Owing to their tailored porous structure and the availability of strong adsorption sites, there is great interest for the utilization of MOFs in hydrocarbon separations.3,12,15−23 The efficacy of MOFs to separate hydrocarbons is generally estimated by ideal adsorbed solution theory (IAST), which gives the selectivity value for multicomponent mixed gases derived from the corresponding single component gas adsorption isotherms. For instance, the series of M2(dobdc) (M = Mg, Mn, Fe, Co, Ni, Zn; dobdc4− = 2,5dioxido-1,4-benzenedicarboxylate) has been analyzed for C2 and C3 hydrocarbon separations by IAST and confirms the potential of M2(dobdc) for the separation of ethene/ethane and propene/propane mixtures owing to the preferred binding of unsaturated hydrocarbons to the open metal sites.15 MFM300(Al) (MFM = Manchester Framework Material, replacing Received: January 31, 2016 Revised: March 2, 2016

A

DOI: 10.1021/acs.chemmater.6b00443 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

and 1000 mbar loadings. The PXRD data and Le Bail refinements were analyzed via the TOPAS software program. Calculation of Selectivity. Myers and Prausnitz developed Ideal Adsorption Solution Theory (IAST) to estimate the amount adsorbed of each component of a mixture using the isotherms for the single components.26 The theory assumes that the adsorbed gases form an ideal solution, and therefore the method is applicable at relatively low pressures and surface coverages. It uses spreading pressure to characterize the adsorbed phase and assumes that the adsorbent structure is thermodynamically inert with its surface area being independent of the adsorbate. The pure component isotherms must be measured accurately at low surface coverage because the integration to obtain spreading pressure is sensitive to this part of the isotherm. This is particularly critical for components in low gas phase concentrations. The IAST calculations for multicomponent mixtures can be carried out by the following methods. The isotherms of the pure components are fitted to appropriate equations to provide a mathematical description of the isotherm. Previous studies have used the Langmuir and/or dual site Langmuir Freundlich equations to describe the isotherms for the pure components.3 These equations are then integrated to obtain the spreading pressure as a function of amount adsorbed for each component. Alternatively, the spreading pressures can be obtained as a function of the amount adsorbed by direct numerical trapezoidal integration of the isotherm data for the pure components. The amounts adsorbed for mixtures and gas pressures are then calculated for specific spreading pressures with the estimation of selectivity data. We have calculated the binary mixed gases isotherms by two methods using a dual-site Langmuir−Freundlich (DSLF) method and a Numerical Integration (NI) method combined with IAST.26 DSLF is a frequently used model to fit the isotherms, which adopts the leastsquares errors to estimate the overall error and finally gives six parameters that determine the sorption profile with the minimum residual sum of squares. The equation for DSLF13 is given as

the NOTT designation) also shows great potential for the separation of C2 hydrocarbons because of its hydroxyldecorated pore environment and suitable pore size.16 Direct comparison of the IAST method and experimental measurement of dual-component adsorption isotherms allows a better understanding of the selectivity for mixed gas adsorption. However, this experiment is usually very challenging in practice. Here, we report the adsorption of light hydrocarbons (both single component and binary mixtures) in an In(III)tetracarboxylate system, MFM-202a, which has been reported as a flexible and defect material. This coordination complex exhibits the highest BET surface area (2220 m2 g−1) among all In(III)-MOFs, high CO2 (19.7 mmol g−1 at 195 K and 1 bar),24 and the highest SO2 (13.6 mmol g−1 at 268 K and 1 bar)25 uptake capacities with stepwise sorption and hysteretic desorption associating with a structural phase transition. MFM-202a exhibits type-I isotherms for C1−C3 hydrocarbons at 273−303 K. However, at 195 K, the adsorption of acetylene exhibits marked stepwise and an exceptionally high uptake coupled with significant hysteretic desorption. A similar adsorption−desorption hysteresis loop was also observed for ethene but not for ethane. The uptakes of propane and propene are similar at 1 bar and 273−303 K but show differences both in terms of reversibility and capacities (12.1 mmol g−1 at 201 K for propene and 9.0 mmol g−1 at 195 K for propane) at low temperatures. The unsaturated hydrocarbon, propene, shows a hysteretic desorption isotherm, whereas propane exhibits fully reversible uptake. In situ synchrotron PXRD experiments revealed an absence of framework phase change as a function of gas loading, and thus the observed stepwise adsorption is attributed to sequential pore filling. Selectivity data obtained from the IAST calculation and the analysis of adsorption isotherms for gas mixtures are in excellent agreement and suggest MFM-202a has potential for hydrocarbon separation, particularly for natural gas purification.

■

N = P1 ×

P2x P 3 1 + P2x P 3

+ Q1 ×

Q 2xQ 3 1 + Q 2xQ 3

(1)

Here, x is the pressure of the pure gas at equilibrium with the adsorbed phase (kPa), N is the gas amount adsorbed per gram of adsorbent (mmol g−1), P1 and Q1 are the simulated saturation capacities of sites P and Q (mmol g−1), P2 and Q2 are the simulated affinity coefficients of sites P and Q (kPa−1), and 1 and 1 represent

EXPERIMENTAL SECTION

Materials and Methods. All chemical reagents were used as received from commercial suppliers without further purification. MFM-202 was prepared from a previously reported method.24 The acetone-exchanged sample of MFM-202 was activated at 323 K and 10−7 mbar for approximately 1 day to remove the free solvent in the pore, and hydrocarbon adsorption measurements were then carried out on an IGA-003 system (Hiden). The temperature was controlled by water bath (for the measurements at 273−303 K), acetone−dry ice bath (for the 195 K measurements), and acetonitrile−dry ice bath (for the 201 K measurement) in separate experiments. All the hydrocarbons (CH4, C2H4, C2H6, C3H6, and C3H8) apart from C2H2 used for were ultrapure research grade (>99.99%) purchased from BOC or Air Liquide. C2H2 was purified by dual-stage cold trap systems and an activated carbon filter to remove acetone before use. The equimolar binary gases adsorption measurements were performed on the IGA system equipped with multiple identical mass flow controllers enabling mixing two different gas components to an equimolar mixture in this study. In Situ Synchrotron Powder X-ray Diffraction (PXRD). The in situ PXRD experiments as a function of hydrocarbon loading were carried out at Beamline I11 Diamond Light Source (Oxford, UK) using high-resolution synchrotron diffraction (λ = 0.825774 Å). The sample was loaded into capillary gas cell, and the temperature was controlled by an Oxford open-flow Cryosystems. Desolvated MFM202a was generated by heating the sample in situ under vacuum overnight. The C2H2 was loaded first, the resultant PXRD was measured at 0, 64, 209, 410, and 870 mbar loadings, and then the sample was degassed to 0 mbar. C2H4 and C2H6 were then loaded in sequence, and the corresponding PXRD were measured at 0, 100, 500,

P3

Q3

deviations from an ideal homogeneous surface. The fitted parameters were then used to predict multicomponent adsorption using IAST.13 The equation for the calculation of selectivity (S) between gas 1 and gas 2 is shown as

S=

x1/y1 x 2/y2

(2)

where xi and yi are the mole fractions of component i in the adsorbed and gas phase, respectively, and i equals 1 or 2. In this work, selectivities for equimolar binary gases of C2H2/CH4, C2H4/CH4, C2H6/CH4, C3H6/CH4, C3H8/CH4, C2H6/C2H4, C3H6/C2H4, and C3H8/C2H6 have been calculated for MFM-202a at ambient conditions (i.e., 293 K and up to 1 bar). The other method used in this work, NI, is adopted as a means to simulate the uptake of binary gases. The application of an NI method to the prediction of isotherm or selectivity data precludes the uncertainty originating from the use of a certain type of adsorption model (e.g., the Langmuir model). The predicted isotherm was obtained by the numerical trapezoidal integration of the appropriate graph for each pure gas adsorption isotherm. The limits for the calculation of the spreading pressures are set by the limits for the pure component experimental isotherms and the gas phase concentrations in the mixture. In this paper, IAST calculations within the range covered by the experimental isotherms are given as solid lines, whereas extrapolations beyond this range using equations to describe the isotherm with the lowest spreading pressure outside the range of the B


Article


Figure 1. (a) View of H4L used to construct MFM-202a. (b) View of channels running along c axis for MFM-202a (indium: green polyhedron; oxygen: red; carbon: gray; hydrogen is omitted for clarity; pore space (9 × 9 Å) is highlighted by yellow balls).

Figure 2. (Top) Hydrocarbon sorption isotherms at low temperatures for MFM-202a (solid symbol: adsorption; open symbol: desorption). (Bottom) Comparison of in situ high resolution powder diffraction patterns for MFM-202a as a function of different C2H2 (left), C2H4 (middle) and C2H6 (right) loading at 195 K (λ = 0.825774 Å). Le Bail refinement and the summary of lattice parameters are shown in Supporting Information. experimental isotherm data are shown by dashed lines in all figures. However, it is worth mentioning that the IAST method becomes less accurate when a) the host material surfaces contain strong binding sites and/or b) the two gases show significantly different sorption capacities. The errors of predicted selectivity values based upon DSLF-IAST and NI-IAST methods are calculated by the equation x − x2 y= 1 × 100% x2 (3)

reported method.24 The material has a 4,4-diamondoid framework structure with each ligand coordinating to four In(III) centers through the deprotonated carboxylates and vice versa to give an interpenetrated open structure (Figure 1b). The desolvated sample MFM-202a shows a different (more porous) structure due to the phase change (pore rearrangement) on removal of free solvent molecules from the pore. Desolvated MFM-202a has a pore size of 9 × 9 Å, a BET surface area of 2220 m2 g−1, and a large pore void comprising ∼70% of the cell volume as estimated by PLATON/SOLV.27 Adsorption Properties. The uptake of CH4 in MFM-202a at 195 K and 1 bar is 6.21 mmol g−1 (Figure 2), corresponding to a storage density of 104.5 kg m−3, reaching 25% of the density of liquid methane at its boiling point (422.4 kg m−3 at 111.7 K). The uptake reaches saturation at 195 K and 16 bar (13.1 mmol g−1, Figure S1). At higher temperature and 20 bar, CH4 adsorption in MFM-202a shows moderate uptakes (6.7−

where y is the error, and x1 and x2 are the uptake amounts for the experimental measurement and IAST prediction (mmol g−1), respectively, at specific pressure and temperature.

■

RESULTS AND DISCUSSION Crystal Structure. MFM-202 was synthesized from 5′,5″bis(4-carboxyphenyl)-[1,1′:3′,1″:3″,1‴-quaterphenyl]-4,4‴-dicarboxylic acid (H4L, Figure 1a) and In(NO3)3 using the C


Article

Chemistry of Materials Table 1. Summary of Physical Parameters of Light Hydrocarbons15,33 gas

kinetic diameter (Å)

critical temp (°C)

dipole moment (× 10−30 C m)

CH4 C2H2 C2H4 C2H6 C3H6 C3H8

3.758 3.3 4.163 4.4443 4.678 4.3−5.118

−82.60 35.75 9.2 32.17 91.06 96.7

0 0 0 0 1.22 0.28

quadrupole moment (× 10−40 C m2) 0 5.00 2.17

4.7 mmol g−1 from 273 to 303 K) compared with other MOFs with similar pore volume. For example, MOF-5 exhibits a CH4 storage capacity of ca. 6.9 mmol g−1, and HKUST-1 shows ca. 9.4 mmol g−1 at 298 K 20 bar.1,2 The moderate uptake of methane in MFM-202a at ambient conditions is consistent with the pore size and lack of open metal sites. To determine the maximum uptake of C 2 and C 3 hydrocarbons and evaluate the potential of hydrocarbon storage in MFM-202a, we carried out isotherm measurements at low temperature. Adsorption of C2H2, C2H4, and C2H6 in MFM202a shows exceptionally high uptakes of 18.3, 14.8, and 11.9 mmol g−1, respectively at 195 K and 1 bar, corresponding to 500.0 kg m−3, (the density of adsorbed C2H2 is 69% of the solid acetylene density at 192 K, 729 kg m−3), 435.6 kg m−3 (the density of adsorbed C2H4 is 77% of liquid ethene density at 169 K, 568 kg m−3), and 375.5 kg m−3 (the density of adsorbed C2H6 is 69% of liquid ethane density at 185 K, 544 kg m−3). Interestingly, the unsaturated hydrocarbons, C2H2 and C2H4, both exhibit stepwise adsorption and hysteretic desorption isotherms, whereas the saturated C2H6 retains the highly reversible type-I isotherm. C2H2 adsorption shows a step at 27 mbar with an initial uptake of 5.3 mmol g−1; an additional 11.4 mmol g−1 adsorption of C2H2 was gained at 450 mbar where it reaches saturation. The desorption shows unprecedented hysteretic release of adsorbed C2H2 molecules with 75% adsorbed molecules retained at 100 mbar, 37% retained at 10 mbar, and 24% retained at 2 mbar, indicating a strong interaction between adsorbed C2H2 molecules and the host at 195 K. 100% of adsorbed C2H2 molecules were released upon applying vacuum to the sample. Similar observation has been observed for CO2 uptakes at low temperature.24 Like CO2, C2H2 has a significant quadrupole moment, which may induce specific host−guest interactions together with other soft binding interactions resulting in the stepwise adsorption and the desorption hysteresis. Such broad hysteretic sorption behavior is rarely observed for C 2 H 2 adsorption in MOFs3,13,14,28,29 and is particularly advantageous to its practical storage and transportation, with C2H2 adsorbed at high pressure but stored at relatively low pressure. This will significantly decrease the possibility of explosion by compression. A smaller hysteretic C2H2 desorption has been observed in M′MOF-3a,30 MOF-508,31 and HOF-5a,32 which also show much lower uptakes (ca. 6.6 mmol g−1 at 195 K, ca. 4.0 mmol g−1 at 290 K, and ca. 8.1 mmol g−1 at 273 K, respectively). A tentative explanation to the hysteresis can be given based upon the flexible host structure of MOF-508 with a potential narrowpore to large-pore framework phase transition proposed upon gas loading.31 C2H4 adsorption at 195 K shows a similar but narrower step at 40 mbar with an uptake of 9.4 mmol g−1, with an additional 3.7 mmol g−1 adsorbed up to 200 mbar where it has reached saturation. The desorption shows slightly hysteretic release of

polarizability (× 10−25 cm3)

molecular radius (Å)

25.93 33.3−39.3 42.52 44.3−44.7 62.6 62.9−63.7

2.276

2.655 2.946

adsorbed C2H4 molecules with 13% retained at 2 mbar, indicating the interaction between adsorbed C2H4 molecules and the host is weaker than that of C2H2. This represents a very rare example of marked stepwise and broadly hysteretic adsorption of ethene in a porous MOF material. In comparison, C2H6 shows fully reversible type-I adsorption without steps. Given the similar molecular structures and volatilities of these three C2 hydrocarbons (Table 1),15,33 this result indicates that the π electrons in C2H2 and C2H4 induce stronger interactions with the MOF host than C2H6 leading to the presence of adsorption steps and hysteresis. C2H2 has more π electron density than C2H4 resulting in a more significant step and broader hysteresis loop in the isotherm data. Similarly, the unsaturated hydrocarbon propene, C3H6, also shows a narrow adsorption step with hysteretic desorption process (12.1 mmol g−1 at 201 K and 200 mbar, Figure 2), while propane, C3H8, exhibits fully reversible uptakes (9.0 mmol g−1 at 195 K and 125 mbar). This result is consistent with the different behaviors of saturated and unsaturated C2 hydrocarbon adsorption in MFM-202a, representing the first example of π electron-dependent hysteretic gas adsorption in MOFs. At more ambient temperatures (273−303 K), both C3H6 and C3H8 isotherms show reversible type-I profile with very sharp uptakes at low pressure (Figure 4a), indicating the presence of strong affinity between the MOF host and C3 hydrocarbons. At 293 K and 1 bar, adsorption capacity of C3H6 and C3H8 in MFM-202a are measured as 7.18 and 6.76 mmol g−1, respectively, comparable to the best-behaving MOFs reported to date.3,14,17 Investigation of the Mechanism of Hysteretic Adsorption. To investigate the binding interaction between adsorbed C2 hydrocarbons and the framework host and the possible framework phase transition, we carried out in situ synchrotron X-ray powder diffraction of MFM-202a on separate loadings of C2H2, C2H4, and C2H6 at 195 K. The PXRD of MFM-202a shows shifts in peaks slightly to a higher value of 2θ as the pressure of C2H2 increases from 0 to 1000 mbar indicating a contraction of the unit cell parameters (Figures 2, S7 and S10) with a slight broadening of peak widths, possibly due to the breakdown of MOF particles and presence of guest−host disorder upon C2H2 adsorption. In contrast, the in situ PXRD patterns of C2H4- and C2H6-loaded MFM-202a show small shifts to lower values 2θ as the pressure increases. The unit cell volumes consequently increase on loading of C2H4 and C2H6 (Figures S11 and S12). Upon desorption of C2H2, C2H4, and C2H6, the PXRD patterns of MFM-202a return to that of the desolvated bare MOF confirming the overall stability of the framework structure. This study confirms the absence of framework phase change albeit with the presence of adsorption steps and hysteretic desorption for C2H2 and C2H4. This observation is similar to that of CO2-loaded MFM-202a but contrasts to that observed D


Article

Chemistry of Materials Table 2. Comparison of Hydrocarbons Uptakes and Isosteric Enthalpies of Adsorption (Qst) for a Series of MOFs uptake (mmol g−1) compounds a

MFM-202a Fe2(dobdc)b Cu-TDPATv PAF-40d PAF-40-Fed PAF-40-Mnd MFM-300a UTSA-35ae M′MOF-3af PAF-1-SO3Agg

Qst (kJ mol−1) at low coverage

CH4

C2H2

C2H4

C2H6

C3H6

C3H8

CH4

C2H2

C2H4

C2H6

C3H6

C3H8

0.45 0.77 1.26 0.54 0.62 0.49 0.29 0.43

3.44 6.89 7.93

2.90 6.02 7.34 1.80 2.31 2.18 4.28 2.16 ∼0.4 4.06

4.21 5.00 6.89 1.95 1.85 2.05 0.85 2.43

7.18 6.66

6.76 5.67

14 20 21 18 23 16 n.a. ∼18

23 47 43

18 45 50 25 30 25 16 ∼28 27 106

18 25 30 36 48 35 11 ∼30

33 44

27 33

∼33

∼42

6.34 2.90 ∼1.9

2.39 2.58 2.51 3.29

2.97

32 ∼29 27

2.23

27

ref this work 3 13 42 42 42 16 14 16, 30 16, 30

a

Data was measured and processed at 293 K and 1 bar. bData was measured and processed at 318 K and 1 bar. vData was measured and processed at 298 K and 1 bar. dData was measured and processed at 298 K and 1.1 bar. eData was measured and processed at 296 K and 1 atm. fData was measured and processed at 295 K and 1 bar. gData was measured and processed at 296 K and 1 bar.

Table 3. Comparison of Uptake Ratios and Absolute Differences for a Series of MOFs absolute uptake differences between two hydrocarbons (mmol g−1)

ratios of uptakes compounds MFM-202aa Fe2(dobdc)b Cu-TDPATc PAF-40d PAF-40-Fed PAF-40-Mnd MFM-300a UTSA-35ae

C2H2 /CH4

C2H4 /CH4

C2H6 /CH4

C3H6 /CH4

C3H8 /CH4

C2H2 CH4

C2H4 CH4

C2H6 CH4

C3H6 CH4

C3H8 CH4

7.64 8.95 6.29

6.44 7.82 5.83 3.33 3.73 4.45 14.76 5.02

9.36 6.49 5.47 3.61 2.98 4.18 2.93 5.65

15.96 8.65

15.02 7.36

2.99 6.12 6.67

2.45 5.25 6.08 1.26 1.69 1.69 3.99 1.73

3.76 4.23 5.63 1.41 1.23 1.56 0.56 2.00

6.73 5.89

6.31 4.90

21.86 6.74

4.43 4.16 5.12 7.65

6.05 2.47

6.91

1.85 1.96 2.02 2.86

2.54

ref this work 3 13 42 42 42 16 14

a

Data was measured and processed at 293 K and 1 bar. bData was measured and processed at 318 K and 1 bar. cData was measured and processed at 298 K and 1 bar. dData was measured and processed at 298 K and 1.1 bar. eData was measured and processed at 296 K and 1 atm.

Figure 3. Variation of the thermodynamic parameters Qst and ΔS with error bars for MFM-202a as a function of hydrocarbon uptake.

for SO2-loaded MFM-202a.24,25 Given the large pore size of 9 × 9 Å, no noticeable confinement effect of adsorbed gas molecules in the pore of MFM-202a was seen upon desorption.

Therefore, the observed adsorption step and hysteresis are most likely due to the pore filling effect in this defect MOF material.32−37 Owing to the complexity of the crystal structure E


Article


Figure 4. Experimental and modeling of pure and binary hydrocarbon sorption isotherms. a) Hydrocarbon isotherms for MFM-202a measured at 293 K. b, c, d) Comparison between experimental equimolar binary hydrocarbons uptakes and theoretical (IAST) prediction from the pure components: b) C2H4/C2H6, c) C2H4/C3H6, d) C2H6/C3H8. Solid symbols: experimental gas uptake; solid blue/cyan lines: equimolar binary gases uptake predicted by DSLF-IAST/NI-IAST methods based on isotherm experimental data; solid black/gray lines and solid red/orange lines: the calculated contribution of pure gas in the predicted binary uptakes based on experimental data; dashed lines: the relevant extrapolations beyond the isotherm experimental range.

of MFM-202a, satisfactory Rietveld refinements to extract the location of adsorbed gas molecules could not be obtained. Analysis of the lattice parameters via Le Bail refinements are shown in the SI. Room Temperature Adsorption. The adsorption of C2 hydrocarbons in MFM-202a shows near-linear reversible type-I isotherms at 273−303 K and 1 bar (Figures S2−S4). C2H6 exhibits the highest uptake capacity of 6.69 mmol g−1 at 273 K and 1 bar with the observed capacity following the order C2H6 > C2H2 > C2H4. However, this sequence for uptake capacity differs at 195 K in the order C2H2 > C2H4 > C2H6. This result reveals that at low temperature the interaction between π electrons and the host plays an important role in gas adsorption in addition to van der Waals force. Furthermore, MFM-202a shows an impressive C2H6 uptake capacity at 293 K and 10 bar of 9.13 mmol g−1 (saturation uptake). The ethane uptake is comparable to the best-behaving MOFs reported to date (Tables 2 and 3), such as UTSA-35a (2.43 mmol g−1 at 296 K and 1 atm near saturated uptake),14 Cu-TDPAT (6.89 mmol g−1 at 298 K and 1 bar near saturated uptake),13 and Fe2(dobdc) (5.00 mmol g−1 at 318 K and 1 bar near saturated uptake).3 Interestingly, the C3 hydrocarbon uptake exhibits an intersection at 293 K between 0.25 and 0.3 bar, at which the adsorption capacity of C3H6 overtakes that of C3H8. Compared with CH4, C2H2, C2H4, C2H6, and C3H6, the steep isotherm of C3H8 below 0.25 bar indicates that C3H8 has the highest affinity to the framework, consistent with the polarizability of this molecule (Table 1). Under the same conditions, MFM-202a shows relatively low uptake of CH4 compared with C2 and C3

hydrocarbons, indicating the potential of this system for the purification of natural gas. Analysis of Thermodynamics. The isosteric enthalpies of adsorption (Qst) and entropy (ΔS) were calculated using the Clausius−Clapeyron equation as a function of gas loadings from the singe adsorption isotherms at 273−303 K (Figure 3). The Clausius−Clapeyron plots showed good linearity even when structural dynamics and specific interactions were apparent. The values of Qst for CH4 uptake increase steadily from 14 to 19 kJ mol−1 upon increasing CH4 loading (Table 2). Similar observations were also found for uptake of C2H4 (18− 20 kJ mol−1) and C2H6 (18−21 kJ mol−1). In comparison, the values of Qst for C2H2 (22 to 23 kJ mol−1) were almost unchanged with C 2 H 2 uptake up to 3 mmol g −1 . C 3 hydrocarbons have higher enthalpies of adsorption (C3H6, 27−42 kJ mol−1; C3H8, 27−39 kJ mol−1) than CH4 and C2 hydrocarbons, and both exhibit a gradual increase with increasing gas uptakes. At low surface coverage, where the adsorbate−adsorbent interaction plays a major role, the sequence of Qst values for saturated hydrocarbon is C3H8 > C2H6 > CH4. This observation is consistent with the increased polarizability of gas molecules (Table 1) that contributes to the molecular interaction with the framework.17,18 For the unsaturated hydrocarbons, the sequence of the interaction strength is C3H6 > C2H2 > C2H4. It is noticeable that the uptake of C3H6 overtakes C3H8 at ∼5.7 mmol g−1, where the Qst value of the former shows a dramatic increase but the latter reaches a near plateau region. The entropies for the hydrocarbons all follow a trend of gradual decrease corresponding to the continuous ordering of the system except F


Article


Figure 5. Comparison between experimental equimolar binary hydrocarbons uptakes and IAST prediction from the pure components. Solid symbols: experimental gas uptake; dashed cyan lines: extrapolation of the DSLF-IAST predicted equimolar binary gases uptake beyond the isotherm experimental range; dashed gray/orange lines: extrapolations of the calculated contribution of pure gas in the predicted binary gases uptakes beyond the isotherm experimental range.

that the C3H8 shows a fluctuation below the uptake amount of 2 mmol g−1 and thereafter decreases gradually. Overall, the enthalpies of adsorption for MFM-202a are generally lower than the reported MOFs with open metal sites [e.g., ∼ 30 kJ mol−1 for C2 hydrocarbon adsorption in UTSA-35a,14 CuTDPAT,13 and Fe2(dobdc)3], indicating a possible reduced energy consumption for the regeneration of MFM-202a as solid absorbents. Selectivity Studies. Both DSLF- and NI-based IAST methods used here can predict the equimolar binary gas adsorption isotherm and the contribution of each gas in the total uptake for the mixture adsorption at a given pressure as shown in Figure 4. By comparison of predicted and experimental adsorption isotherm data for gas mixtures, we can validate the calculated IAST selectivity. The comparison of two IAST approaches gives insight into the errors and uncertainties generated by the adoption of an isotherm model to describe the experimental data. The calculated C2H6/C2H4 binary gases uptake by the NI-IAST prediction has encountered lower divergence than that by the DSLF-IAST method within the experimental spreading pressure range, which is probably due to the inherent drawback of DSLF-IAST based on nearlinear uptake isotherms as shown in Figure 4b. In contrast,

DSLF-IAST and NI-IAST are consistent with each other in the calculation of C3H6/C2H4 and C3H8/C2H6 mixed gases uptake, and both show good agreement with the experimental data. Apart from the binary gases uptake prediction, we have further calculated the contribution of each gas to the total uptakes as shown in Figures 4 and 5. The gas with high uptake shows high contribution to the total uptake, indicating the retention of the competitive interaction with the framework even in the mixed gas system. From the calculation of the single gas contributions in Figure 4b−d, we conclude that DSLF-IAST and NI-IAST show good agreement with each other for a gas with high uptake within the experimental spreading pressure range. However, the two IAST methods show divergence in the low-uptake gas component, and this difference grows as the pressure increases. On the other hand, the DSLF-IAST method can predict the mixed gases uptake at any given pressure, while the NI-IAST method only predicts the uptake within the range of the experimental isotherm, which is determined by the spreading pressure of the gas component with lower uptake. The DSLFIAST method allows extrapolation of the spreading pressure obtained from the isotherm data for the lower uptake adsorbate beyond the experimental isotherm range allowing extensions of G


Article


Figure 6. IAST selectivity for adsorption of equimolar mixture between different hydrocarbons in MFM-202a at 293 K. These selectivity data have also been confirmed by the dual-component gas adsorption experiments as mentioned above. Solid lines: selectivities predicted by DSLF-IAST methods based on experimental data; dashed lines: extrapolations of the IAST predicted selectivities beyond the experimental range.

and may have potential for the selective removal of saturated hydrocarbons (propane, ethane). Direct comparison of the uptake ratios and difference between two hydrocarbon adsorption isotherms also gives insights to the selectivity. Interestingly, the ratios of the adsorption uptakes for C2H6/CH4 (9.36), C3H6/CH4 (15.96), and C3H8/CH4 (15.02) at 1 bar and 293 K are to the best of our knowledge among the highest values in comparison with other MOFs (Table 3).40,41 For example, the PAF-40 series (PAF-40, PAF-40-Fe, and PAF-40-Mn)42 shows ca. three to five times higher uptake for C3H8, C2H6, and C2H4 over CH4 at 273−298 K and 1.1 bar. This analysis indicates the potential application of MFM-202a in the purification of natural gas by the selective removal of higher saturated hydrocarbons.

predictions for gas mixture to higher uptakes subject to larger uncertainty. In this study, the NI-IAST method cannot be applied to calculate the uptake for C2 or C3 hydrocarbons in equimolar mixtures with CH4 over a considerable pressure range where the uptakes of C2 and C3 hydrocarbons are much higher (approximately 6−16 times) than that of CH4. Thus, we adopted the DSLF-IAST method for prediction of C2 and C3 hydrocarbons over CH4 as shown in Figure 5. In the DSLFIAST calculations for C2/CH4 mixtures, the results for C2 gases over CH4 are consistent with the experimental measurements at 293 K, with extrapolation errors of −1.7% for C2H2/CH4, 3.3% for C2H4, and 2.5% for C2H6/CH4 at 1 bar. The error increases with pressure; for example, the DSLF-IAST result for C2H6/ CH4 overestimates the experimental data by 9.4% at 10 bar and 293 K. For adsorption of C3/CH4 mixtures, the DSLF-IAST predictions also fit very well with the experimental data within the experimental spreading pressure range, with extrapolation errors of 2.1% for C3H6/CH4 and −0.4% for C3H8/CH4 at 293 K and 1 bar. This again validates that DSLF-IAST is reliable within the experimental spreading pressure range, but it shows noticeable errors when significant extrapolation of isotherm data is applied. It is worth noting that although IAST has been widely used to estimate the selectivity of competitive adsorption in MOFs, the comprehensive analysis of the associated errors is reported here for the first time. The selectivity data for MFM-202a were calculated by the DSLF-IAST method based upon the single component isotherms at 293 K.26 The C3H8/CH4 selectivity is estimated as 105 at 0.01 bar and gradually drops to 87 at 1 bar (Figure 6). The C3H6/CH4 selectivity lies in the range of 63 and 75 at pressure below 1 bar. These values are comparable to the bestbehaving MOFs reported to date (Table S3). With combined high uptake capacities and selectivities of C3H6 and C3H8 at 293 K, MFM-202a has excellent potential for C3H6/CH4 and C3H8/CH4 separations. In comparison, MFM-202a only shows moderate selectivities between C2 hydrocarbons and CH4 (generally between 12 and 7 at pressure below 1 bar). Moreover, MFM-202a also shows selectivity for C3H8/C2H6 (7−8) and C3H6/C2H4 (8−9). Interestingly, MFM-202a shows an inverse selectivity for C2H4/C2H6, 0.7 at 1 bar (i.e., C2H6/ C2H4 selectivity = 1.4). This is an unusual result.15,38 For example, Fe2(dobdc),3 PAF-1-SO3Ag,39 and MFM-30016 show a C2H4/C2H6 selectivity of ca. 13−18, ∼27, and ∼49, respectively. This result indicates that MFM-202a has stronger binding to the saturated hydrocarbons at ambient conditions

■

CONCLUSIONS The adsorption of light hydrocarbons in a flexible porous MOF material, MFM-202a, has been comprehensively investigated at various temperatures. MFM-202a shows reversible isotherms for hydrocarbon adsorption at 273−303 K. However, the unsaturated hydrocarbons, acetylene, ethene, and propene, exhibit marked stepwise adsorption isotherms at low temperatures due to the pore filling effect in the flexible framework material with structural defects, as confirmed by in situ synchrotron X-ray powder diffraction experiments. MFM202a shows very high acetylene uptake of 18.3 mmol g−1 at 195 K and 1 bar. A comprehensive analysis of the selectivity using both the NI- and DSLF-based IAST methods and measurement of mixed gas adsorption isotherms indicates that MFM-202a has great potential for purification of CH4 (natural gas). The extrapolation errors for the widely used DSLF-IAST method for the estimation of selectivity data from single component uptake isotherms and the validity of the selectivity via IAST calculations have been quantified and discussed.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.chemmater.6b00443. Substrate sorption isotherms, in situ PXRD, unit cell parameters analyses, lists of DSLF parameters, dual-site Langmuir fitting model pure hydrocarbon sorption isotherms and comparison of substrate selectivities (PDF) H


Article


■

ethane and propylene over propane in the metal-organic frameworks M2(dobdc) (M = Mg, Mn, Fe, Co, Ni, Zn). Chem. Sci. 2013, 4, 2054− 2061. (16) Yang, S.; Ramirez-Cuesta, A. J.; Newby, R.; Garcia-Sakai, V.; Manuel, P.; Callear, S. K.; Campbell, S. I.; Tang, C. C.; Schröder, M. Supramolecular binding and separation of hydrocarbons within a functionalized porous metal-organic framework. Nat. Chem. 2015, 7, 121−129. (17) Belmabkhout, Y.; Mouttaki, H.; Eubank, J. F.; Guillerm, V.; Eddaoudi, M. Effect of pendant isophthalic acid moieties on the adsorption properties of light hydrocarbons in HKUST-1-like tboMOFs: application to methane purification and storage. RSC Adv. 2014, 4, 63855−63859. (18) Li, J.; Kuppler, R. J.; Zhou, H. Selective gas adsorption and separation in metal-organic frameworks. Chem. Soc. Rev. 2009, 38, 1477−1504. (19) Fu, H.; Wang, F.; Zhang, J. A stable zinc-4-carboxypyrazole framework with high uptake and selectivity of light hydrocarbons. Dalton Trans. 2015, 44, 2893−2896. (20) Tan, Y.; He, Y.; Zhang, J. High and selective sorption of C2 hydrocarbons in heterometal-organic frameworks built from tetrahedral units. RSC Adv. 2015, 5, 7794−7797. (21) He, Y.; Zhang, Z.; Xiang, S.; Wu, H.; Fronczek, F. R.; Zhou, W.; Krishna, R.; O’Keeffe, M.; Chen, B. High separation capacity and selectivity of C2 hydrocarbons over methane within a microporous metal-organic framework at room temperature. Chem. - Eur. J. 2012, 18, 1901−1904. (22) He, Y.; Song, C.; Ling, Y.; Wu, C.; Krishna, R.; Chen, B. A new MOF-5 homologue for selective separation of methane from C2 hydrocarbons at room temperature. APL Mater. 2014, 2, 124102− 1−6. (23) Alawisi, H.; Li, B.; He, Y.; Arman, H. D.; Asiri, A. M.; Wang, H.; Chen, B. A microporous metal-organic framework constructed from a new tetracarboxylic acid for selective gas separation. Cryst. Growth Des. 2014, 14, 2522−2526. (24) Yang, S.; Lin, X.; Lewis, W.; Suyetin, M.; Bichoutskaia, E.; Parker, J. E.; Tang, C. C.; Allan, D. R.; Rizkallah, P. J.; Hubberstey, P.; Champness, N. R.; Thomas, K. M.; Blake, A. J.; Schröder, M. A partially interpenetrated metal-organic framework for selective hysteretic sorption of carbon dioxide. Nat. Mater. 2012, 11, 710−716. (25) Yang, S.; Liu, L.; Sun, J.; Thomas, K. M.; Davies, A. J.; George, M. W.; Blake, A. J.; Hill, A. H.; Fitch, A. N.; Tang, C. C.; Schröder, M. Irreversible network transformation in a dynamic porous host catalyzed by sulfur dioxide. J. Am. Chem. Soc. 2013, 135, 4954−4957. (26) Myers, A. L.; Prausnitz, J. M. Thermodynamics of mixed-gas adsorption. AIChE J. 1965, 11, 121−127. (27) Spek, A. L. Single-crystal structure validation with the program PLATON. J. Appl. Crystallogr. 2003, 36, 7−13. (28) Nijem, N.; Wu, H.; Canepa, P.; Marti, A.; Balkus, K. J., Jr.; Thonhauser, T.; Li, J.; Chabal, Y. J. Tuning the gate opening pressure of metal-organic frameworks (MOFs) for the selective separation of hydrocarbons. J. Am. Chem. Soc. 2012, 134, 15201−15204. (29) He, Y.; Xiang, S.; Zhang, Z.; Xiong, S.; Fronczek, F. R.; Krishna, R.; O’Keeffe, M.; Chen, B. L. A microporous lanthanide-tricarboxylate framework with the potential for purification of natural gas. Chem. Commun. 2012, 48, 10856−12858. (30) Xiang, S.; Zhang, Z.; Zhao, C.; Hong, K.; Zhao, X.; Ding, D.; Xie, M.; Wu, C.; Das, M.; Gill, R.; Thomas, K. M.; Chen, B. Rationally tuned micropores within enantiopure metal-organic frameworks for highly selective separation of acetylene and ethylene. Nat. Commun. 2011, 2, 204. (31) Xiang, S.; Zhou, W.; Gallegos, J. M.; Liu, Y.; Chen, B. Exceptionally high acetylene uptake in a microporous metal-organic framework with open metal sites. J. Am. Chem. Soc. 2009, 131, 12415− 12419. (32) Wang, H.; Li, B.; Wu, H.; Hu, T.; Yao, Z.; Zhou, W.; Xiang, S.; Chen, B. A flexible microporous hydrogen-bonded organic framework for gas sorption and separation. J. Am. Chem. Soc. 2015, 137, 9963− 9970.

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank the University of Manchester and the University of Nottingham for funding. S.G. and C.G.M acknowledge receipts of a CSC scholarship and a Diamond Light Source studentship, respectively. M.S. acknowledges receipt of an EPSRC Programme Grant and ERC Advanced Grant. We are especially grateful to Diamond Light Source for access to Beamline I11.

■

REFERENCES

(1) Mason, J. A.; Veenstra, M.; Long, J. R. Evaluating metal-organic frameworks for natural gas storage. Chem. Sci. 2014, 5, 32−51. (2) He, Y.; Zhou, W.; Qian, G.; Chen, B. Methane storage in metalorganic frameworks. Chem. Soc. Rev. 2014, 43, 5657−5678. (3) Bloch, E. D.; Queen, W. L.; Krishna, R.; Zadrozny, J. M.; Brown, C. M.; Long, J. R. Hydrocarbon separations in a metal-organic framework with open iron(II) coordination sites. Science 2012, 335, 1606−1610. (4) He, Y.; Krishna, R.; Chen, B. Metal-organic frameworks with potential for energy-efficient adsorptive separation of light hydrocarbons. Energy Environ. Sci. 2012, 5, 9107−9120. (5) Newalkar, B. L.; Choudary, N. V.; Turaga, U. T.; Vijayalakshmi, R. P.; Kumar, P.; Komarneni, S.; Bhat, T. S. G. Potential adsorbent for light hydrocarbon separation: role of SBA-15 framework porosity. Chem. Mater. 2003, 15, 1474−1479. (6) Newalkar, B. L.; Choudary, N. V.; Turaga, U. T.; Vijayalakshmi, R. P.; Kumar, P.; Komarneni, S.; Bhat, T. S. G. Adsorption of light hydrocarbons on HMS type mesoporous silica. Microporous Mesoporous Mater. 2003, 65, 267−276. (7) De Stefanis, A.; Perez, G.; Tomlinson, A. A. G. Pillared layered structures vs. zeolites as sorbents and catalysts. J. Mater. Chem. 1994, 4, 959−964. (8) Malek, A.; Farooq, S. Kinetics of hydrocarbon adsorption on activated carbon and silica gel. AIChE J. 1997, 43, 761−776. (9) Yaghi, O. M.; O’Keeffe, M.; Ockwig, N. W.; Chae, H. K.; Eddaoudl, M.; Kim, J. Reticular synthesis and the design of new materials. Nature 2003, 423, 705−714. (10) Wang, T. C.; Bury, W.; Goḿ ez-Gualdroń, D. A.; Vermeulen, N. A.; Mondloch, J. E.; Deria, P.; Zhang, K. N.; Moghadam, P. Z.; Sarjeant, A. A.; Snurr, R. Q.; Stoddart, J. F.; Hupp, J. T.; Farha, O. K. Ultrahigh surface area zirconium MOFs and insights into the applicability of the BET theory. J. Am. Chem. Soc. 2015, 137, 3585− 3591. (11) Yan, Y.; Telepeni, I.; Yang, S.; Lin, X.; Kockelmann, W.; Dailly, A.; Blake, A. J.; Lewis, W.; Walker, G. S.; Allan, D. R.; Barnett, S. A.; Champness, N. R.; Schröder, M. Metal-organic polyhedral frameworks: high H2 adsorption capacities and neutron powder diffraction studies. J. Am. Chem. Soc. 2010, 132, 4092−4094. (12) Herm, Z. R.; Bloch, E. D.; Long, J. R. Hydrocarbon separations in metal-organic frameworks. Chem. Mater. 2014, 26, 323−338. (13) Liu, K.; Ma, D.; Li, B.; Li, Y.; Yao, K.; Zhang, Z.; Han, Y.; Shi, Z. High storage capacity and separation selectivity for C2 hydrocarbons over methane in the metal-organic framework Cu-TDPAT. J. Mater. Chem. A 2014, 2, 15823−15828. (14) He, Y.; Zhang, Z.; Xiang, S.; Fronczek, F. R.; Krishna, R.; Chen, B. A robust doubly interpenetrated metal-organic framework constructed from a novel aromatic tricarboxylate for highly selective separation of small hydrocarbons. Chem. Commun. 2012, 48, 6493− 6495. (15) Geier, S. J.; Mason, J. A.; Bloch, E. D.; Queen, W. L.; Hudson, M. R.; Brown, C. M.; Long, J. R. Selective adsorption of ethylene over I


Article

Chemistry of Materials (33) Drago, R. S.; Webster, C. E.; McGilvray, J. M. A multipleprocess equilibrium analysis of silica gel and HZSM-5. J. Am. Chem. Soc. 1998, 120, 538−547. (34) Lin, X.; Wilson, C.; Sun, X.; Champness, N. R.; Hubberstey, P.; Mokaya, R. Schröder. A porous framework polymer based on a zinc(II) 4,4′-bipyridine-2,6,2′,6′-tetracarboxylate: synthesis, structure, and “zeolite-like” behaviors. J. Am. Chem. Soc. 2006, 128, 10745− 10753. (35) Yang, S.; Lin, X.; Blake, A. J.; Walker, G. S.; Hubberstey, P.; Champness, N. R.; Schröder, M. Cation-induced kinetic trapping and enhanced hydrogen adsorption in a modulated anionic metal-organic framework. Nat. Chem. 2009, 1, 487−493. (36) Henke, S.; Schneemann, A.; Wütscher, A.; Fischer, R. A. Directing the breathing behavior of pillared-layered metal-organic frameworks via a systematic library of functionalized linkers bearing flexible substituents. J. Am. Chem. Soc. 2012, 134, 9464−9474. (37) Gücüyener, C.; van den Bergh, J.; Gascon, J.; Kateijn, F. Ethane/ ethene separation turned on its head: selective ethane adsorption on the metal-organic framework ZIF-7 through a gate-opening mechanism. J. Am. Chem. Soc. 2010, 132, 17704−17706. (38) Ma, H.; Ren, H.; Meng, S.; Yan, Z.; Zhao, H.; Sun, F.; Zhu, G. A 3D microporous covalent organic framework with exceedingly high C3H8/CH4 and C2 hydrocarbon/CH4 selectivity. Chem. Commun. 2013, 49, 9773−9775. (39) Li, B.; Zhang, Y.; Krishna, R.; Yao, K.; Han, Y.; Wu, Z.; Ma, D.; Shi, Z.; Pham, T.; Space, B.; Liu, J.; Thallapally, P. K.; Liu, J.; Chrzanowski, M.; Ma, S. Introduction of π-complexation into porous aromatic framework for highly selective adsorption of ethylene over ethane. J. Am. Chem. Soc. 2014, 136, 8654−8660. (40) Xu, H.; Cai, J.; Xiang, S.; Zhang, Z.; Wu, C.; Rao, X.; Cui, Y.; Yang, Y.; Krishna, R.; Chen, B.; Qian, G. A cationic microporous metal-organic framework for highly selective separation of small hydrocarbons at room temperature. J. Mater. Chem. A 2013, 1, 9916− 9921. (41) Fu, H.; Zhang, J. Structural transformation and hysteretic sorption of light hydrocarbons in a flexible Zn-pyrazole-adenine framework. Chem. - Eur. J. 2015, 21, 5700−5703. (42) Meng, S.; Ma, H.; Jiang, L.; Ren, H.; Zhu, G. A facile approach to prepare porphyrinic porous aromatic frameworks for small hydrocarbon separation. J. Mater. Chem. A 2014, 2, 14536−14541.

J


Selective Hysteretic Sorption of Light ... - ACS Publications

Recommend Documents