- Email: [email protected]
paraffin separation
Journal of Membrane Science 557 (2018) 76–86
Contents lists available at ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
Ultra-stable and cost-efficient protic ionic liquid based facilitated transport membranes for highly selective olefin/paraffin separation
T
Haozhen Doua,1, Bin Jianga, Xiaoming Xiaoa, Mi Xua,1, Baoyu Wangb, Li Haoa, Yongli Suna, ⁎ Luhong Zhanga, a b
School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China School of Chemical Engineering and Food Science, Zhengzhou Institute of Technology, Zhengzhou 450000, China
A R T I C LE I N FO
A B S T R A C T
Keywords: Facilitated transport membranes Protic ionic liquids, olefin/paraffin separation Long-term stability
Facilitated transport membranes (FTMs) for olefin/paraffin separations have failed to achieve commercial success due to the instability of carriers although great efforts have been made. In this work, ultra-stable and cost-efficient protic ionic liquid based FTMs (PIL-FTMs) were firstly prepared by utilizing the Brönsted acidic property of PILs to stabilize the carrier. The gas solubility in the carrier/PILs was measured and the separation performances of PIL-FTMs were evaluated systemically. The results indicated that the structure of PILs affected the C2H4 permeability and the presence of ether group and hydroxyl group in PILs significantly enhanced the C2H4/C2H6 selectivity. The carrier concentration led to structural variation of PIL-FTMs, thus manipulating the gas separation performances of PIL-FTMs. The increase of transmembrane pressure decreased C2H4 permeability and C2H4/C2H6 selectivity, indicating a typical feature of FTMs. The increase of temperature increased the C2H4 permeability but decreased C2H4/C2H6 selectivity. The separation performances of PIL-FTMs were much higher than other results in the literature. Furthermore, the PIL-FTMs exhibited excellent stability during the long-term experiments carried out for six months. Finally, the investigation of separation mechanism revealed that the hydrogen-bonding and coordinative interactions between PILs and carrier accounted for the high separation efficiency of PIL-FTMs. In all, the excellent long-term stability, outstanding separation performances and economic feasibility of PIL-FTMs could play an important role in moving these membranes toward industrial application.
1. Introduction Short-chain olefins are among the most important feedstocks, and widely used to produce plastics, rubber, films and other chemicals [1,2]. Industrial olefin/paraffin separations heavily rely upon the extremely cost and energy intensive cryogenic distillation technology [3]. The development of economically viable olefin/paraffin separation processes is becoming increasingly important [4]. Many novel materials and alternative techniques have been investigated for olefin/paraffin separations, such as deep eutectic solvent based membranes [5,6], ionic liquid based membranes [7], polymeric ionic liquid membranes [8], mixed matrix membranes [9], metal-organic frameworks [10] and so on [11]. Facilitated transport membranes (FTMs) have gained prominence as research targets for olefin/paraffin separations because of their potential to perform separation more efficiently than the passive polymeric
⁎
1
Corresponding author. E-mail address: [email protected] (L. Zhang). These authors contributed equally to this work.
https://doi.org/10.1016/j.memsci.2018.04.015 Received 16 December 2017; Received in revised form 6 April 2018; Accepted 10 April 2018 Available online 12 April 2018 0376-7388/ © 2018 Elsevier B.V. All rights reserved.
membranes [12]. In the FTMs, dispersed transition metal cations, such as silver and copper (I) ions, can coordinate with olefins based on a πbond complex formation mechanism, thus significantly increasing the gas separation performances of FTMs [13]. Usually, according to the mobility of carriers, FTMs can be classified into two types: fixed-site carrier membranes and mobile carrier membranes [14]. Polymeric electrolyte membranes, as a typical example of fixed-site carrier FTMs, have been investigated in great detail over the past decades [15,16]. It was found that silver salts added in the polymer electrolytes effectively facilitated olefin transport. Moreover, mobile carrier FTMs can be obtained by adding silver salts into aprotic ionic liquids (AILs). Ortiz et al. firstly reported the separation of propane/propylene mixture by mobile carrier FTMs containing AgBF4 as carrier [17]. Matsuyama et al. designed mobile carrier facilitated transport ion-gel membranes for propylene/propane separation by introducing gelator into AILs [18]. Despite the intensive research of FTMs outlined above, the
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
instability of carrier is still the most challenge for the industrial application of FTMs [12]. The silver ions can be easily reduced to silver particles even under dark conditions, resulting in deterioration of separation performances of FTMs [19]. However, there have been few studies of carrier stability and little is known about how to stabilize the silver ions. Kim et al. firstly reported the reduction of silver ions could be restrained by phthalates but after one week silver particles still appeared [20]. Subsequently, Kang et al. reported the stability of FTMs could be improved by additives such as HBF4 and Al(NO3)3, which could be kept for 100–200 h [21,22], but these additives resulted in relatively low separation performances of FTMs [23]. Very recently, the instability of FTMs was solved by a regeneration method that used a peroxide/acid liquid or vapor phase treatment to oxidize reduced silver carriers within the FTMs [12]. However, this method suffers from many operating drawbacks and challenges which consequently increase the high capital and operating costs. Besides the challenge of carrier instability, economic feasibility of carrier is another decisive factor for the large-scale application of FTMs. Among the many silver salts used as carriers, AgBF4 and AgCF3SO3 have been commonly used because of their highly favorable ability to facilitate olefin transport [17,18,24]. However, their high cost has created another barrier for the large-scale application of FTMs. This barrier can be easily overcome by AgNO3, which has advantages of low cost and facile availability [25]. Unfortunately, AgNO3 is rather inactive in facilitated transport of olefin due to its high lattice energy [25]. Therefore, exploitation of greener and more sustainable solvents to stabilize and activate the carrier of AgNO3 is in high urgency for designing novel and promising mobile carrier FTMs. Protic ionic liquids (PILs) have aroused our great interest due to their useful Brönsted acidic property, which is expected to stabilize silver ions [26]. PILs are a subset of ILs formed by proton transfer from the Brønsted acid to Brønsted base [27]. Their straightforward preparation reduces or eliminates additional work-up stages, such as solvent separation and ion-exchange reactions, which makes them significantly less expensive than common AILs [28]. Moreover, PILs are often characterized by a reduced environmental impacts, such as lower toxicity and higher biodegradability with respect to the corresponding AILs [29]. And most importantly, the structural designability of PILs endows their ability to tune the activity of AgNO3. Compared with AILs, which are widely applied in the olefin/paraffin separations [17,18], PILs have never been reported for olefin/paraffin separations. In this study, a series of PIL-FTMs were designed and fabricated for the C2H4/C2H6 separation. To the best of our knowledge, this is the first report on the fabrication of PIL-FTMs exhibiting excellent long-term stability, C2H4 permeability and C2H4/C2H6 selectivity. The gas solubility in AgNO3/PILs was measured and the gas permeability as well as the selectivity of PIL-FTMs were investigated. The structure-property relationship of PIL-FTMs has been highlighted, the main affected factors of the separation process were optimized systemically and the longterm stability of PIL-FTMs was evaluated intensively. Finally, the facilitated transport mechanism was analyzed by various characterization techniques, which shed insight into the smart molecule design of novel FTMs for the C2H4/C2H6 separation.
Fig. 1. A synthetic route to prepare PILs through the proton-transfer reaction (a); the chemical structure of PILs studied in this work (b).
obtained from Haining Zhongli Filtering equipment Corporation (China). 2.2. Synthesis of PILs All the PILs were prepared according to a similar procedure in the literature [30]. As a typical example, the aqueous solution of nitric acid (0.2 mol) was added dropwise into 2-methoxyethylamine (0.2 mol) using methanol as solvent at 0 °C or below. Then, the reaction was carried out under stirring vigorously at RT for 12 h (Fig. 1). After the reaction was completed, the product was purified by rotary evaporating at 60 °C and then dried under vacuum at 80 °C for 48 h to remove the trace of water and obtain colorless and transparent PIL. The water contents of the PILs were determined to be 0.25–0.58 wt% by Karl-Fischertitration method (DL37 KF Coulometer, Mettler Toledo). 2.3. PIL-FTMs preparation PIL-FTMs were prepared using the similar procedure reported by our group previously [5,6]. Initially, AgNO3 was dissolved in the PIL under magnetic stirring at RT to obtain reactive AgNO3/PIL mixture. Then, after the residual water and other trace volatile compounds inside the pores of nylon membranes were removed under vacuum (< 2 mbar) at 60 °C for 1 h, 3 mL of AgNO3/PIL mixture was spread onto the top surface of the treated membrane support and impregnated into membrane pores by using nitrogen gas with a transmembrane pressure difference of 1 bar in the permeation cell for 10 min. In order to ensure the completed impregnation, the above procedure was carried out several times until getting a thin and visible layer liquid on the bottom surface of the membrane support. Finally, the excessive AgNO3/PIL mixture was removed from the membrane surface using tissue paper. As shown in Fig. 2, the nylon membrane was a highly porous supported material with a typical sponge-like structure. After the impregnation, the pores of the nylon membrane were homogeneously and completely saturated by AgNO3/PIL mixture, which confirmed the successful fabrication of PIL-FTMs.
2. Experimental 2.1. Materials
2.4. Determination of physical properties Ethylene and ethane gases (99.9 mol%) were purchased from Tianjin Tang Dynasty Gas Co., Ltd (China). Propylamine (99 wt%), 2methoxyethylamine (99.5 wt%), 2-(2-Aminoethoxy) ethanol (99 wt%) and nitric acid (65–68 wt%) were all supplied from Shanghai Aladdin Biochem Technology Co., Ltd. (China) and used as received. The microporous membrane support was hydrophilic nylon flat sheet membrane with porosity of 70%, the average thickness of 100 µm, a nominal pore size of 0.1 µm and effective area of 19.625 cm2, which was
The density measurements were conducted at 25 °C with an Anton Paar DMA 5000 type automatic densimeter with a precision of 0.0001 g/cm3, which was calibrated with distilled water prior to the measurements. The viscosities were measured on a Brookfield LVDV-II +Pro viscometer with an uncertainty of ± 1% in relation to full scale. The melting points of PILs were determined by a DSC 204 differential scanning calorimeter. 77
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 2. SEM images of the surface for Nylon membrane (a) and PIL-FTMs (b).
gas at different equilibrium states (P0, P1 and P2), which can be calculated by the Eq. (3):
2.5. Characterization The 1H NMR spectra were recorded on the VARIA INOVA 500 MHz spectrometer using d6-DMSO as the solvent with TMS as the internal standard. The attenuated total reflection infrared spectra (ATR-IR) were recorded on a Bio-Rad FTS 6000 FT-IR spectrometer. Raman measurements were carried out by a Bruker RFS 100 Fourier transform Raman spectrometer. The microscopic morphology of the PIL-FTMs was characterized by the Hitachi S-4800 field emission scanning electron microscope (SEM). Thermogravimetric measurements were conducted on a Netzsch TG 209 thermal gravimetric analyser.
Zi = 1+
Ci =
ni VPILs
BC 2H 4 = 8.7776·10−6.0061·10 4 / T +1.2857·106/ T 2−1.0861·109/ T 3
(4)
BC 2H 6 = 1.0773·102−8.2548·10 4 / T +5.2387·106/ T 2−1.9764·109/ T 3
(5)
Where BC2H4 and BC2H4 are the second-order virial coefficient of C2H4 and C2H6, respectively, cm3/mol. T is the temperature (298.15 K).
The gas solubility was determined in light of pressure drop method described in our previous publication [5]. Briefly, the apparatus is illustrated in Fig. 3. The apparatus consisted of an equilibrium vessel (7) with volume of 133.4 cm3 equipped with a magnetic stirrer, and a gas storage vessel (6) for the introduction of known amounts of 467.5 cm3 gas. The pressures in these two vessels were monitored using pressure transducers (Model CYYZ11, China) with an accuracy of 0.001 bar. The temperature was maintained constant by a DC-5060 thermostatic air bath (10) with an accuracy of ± 0.1 K. The solubility of pure gas in PIL or AgNO3/PIL is determined by Eqs. (1) and (2):
P (V − VPILs ) P0 Vr PV − 1 s − 2 r Z0 RT Z1 RT Z2 RT
(3)
Where the subscripts of “0”, “1”, and “2” represent different equilibrium states (P0, P1 and P2). B is the second-order virial coefficient of ethylene or ethane, and its correlation formula for different gases is illustrated by the Eq. (4) and Eq. (5) [31]:
2.6. Gas absorption
ni = ninitial − nfinal =
Pi⋅B (i = 0,1,2) R·T
2.7. Gas permeation Mixed-gas permeation experiments were performed as reported by our group previously [6], and the setup is shown in Fig. 4. In brief, the operation temperature and transmembrane pressure were regulated by a micrometer value and an air blowing thermostatic oven according to the experimental requirements, respectively. The flow rates of individual gases (ethylene, ethane, sweep gas and outlet gas) were controlled and measured by using four mass flow controllers and the composition of permeated gases was analyzed by the online gas chromatograph. Three parallel experiments were carried out to obtain the reliability values. For mixed gas experiments, the permeability coefficient (Pi ) of a particular gas is defined as the permeation flux (Ji ) normalized to the pressure difference across the membrane ( ∆Pi ), as well as the membrane thickness (δ ), which is described by the Eq. (6):
(1) (2)
Where P0 is the pressure of the total gas introduced into the gas storage vessel while P1 and P2 are the pressure of equilibration vessel and gas storage vessel at equilibrium. Vr, Vs, and VPILs are the volume of the equilibration vessel, gas storage vessel, and the absorbents in equilibration vessel. Z0, Z1, and Z2 are the compressibility factors for the pure
Pi = Ji
δ ∆Pi
(6)
The gas transport through FTMs can be interpreted by the solutiondiffusion theory [8,17]. Therefore, the permeability can further be represented as the product of thermodynamic solubility coefficient (S), and kinetic effective diffusion coefficient (D).
Pi = Si⋅Di
(7)
The gas solubility coefficient can be determined by the gas solubility (ni ) . Once determined the gas permeability through the membrane and gas solubility in the PILs, the effective diffusion coefficient can be derived from Eq. (7). The permeability selectivity (αi, j ) is obtained by dividing the permeability of the more permeable specie i to the permeability of the less permeable specie j for the case where the downstream pressure is negligible relative to the upstream feed pressure. As shown in Eq. (8), the αi, j can be also expressed as the product of the diffusivity selectivity and the solubility selectivity.
Fig. 3. Diagram of gas absorption experimental setup. 1-gas bottles, 2-micrometric values, 3-three way valves, 4-fourway valves, 5-vacuum pump, 6-gas storage vessel, 7-gas equilibrium vessel, 8-pressure transducers, 9-magnetic stirring, 10-constant temperature circulating water bath.
αij = 78
Pi S D = i⋅ i Pj Sj Dj
(8)
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 4. Diagram of gas permeation experimental setup.1-gas bottles, 2-buffer tanks, 3-mass flow controllers, 4-two-way valves, 5-static mixer, 6-three-way valves, 7permeation cell, 8-air blowing thermostatic oven, 9-mass flow meter, 10-pressure transducers, 11-micrometric values, 12-gas chromatograph.
Fig. 5. The PILs characterizations: the 1H NMR spectra (a), the FTIR spectra (b).
respectively. For 2MEAN, the chemical shifts of C1H, C2H and C3H groups occurred at 3.005, 3.512 and 3.293 ppm, respectively. In the case of 22HEEAN, the chemical shift at 3.013 ppm was assigned to C3H group, while the peaks around 3.595 ppm were ascribed to C2H groups. As also seen from Fig. 5a, the chemical shifts at 7.827, 7.909 and 7.897 ppm were assigned to the N‒H proton of PAN, 2MEAN and 22HEEAN, respectively. The chemical shift at 4.185 ppm was attributed to the O-H of 22HEEAN. As shown in Fig. 5b, the broad peaks between 3600 and 3000 cm−1 confirmed the formation of hydrogen bonds between the cation and anion of PIL, which maybe existed as N‒H···O and O‒H···O. However, the peaks for three PILs varied slightly, which was attributed to different structural arrangement of hydrogen bonds and electrostatic interactions. For example, the symmetric N‒H stretching vibrations appeared at 3041, 2996, 3058 cm−1 for PAN, 2MEAN and 22HEEAN, respectively. The peaks at 1350, 1382, 1380 cm−1 were ascribed to the N‒O asymmetric stretching vibrations of PAN, 2MEAN and 22HEEAN. It also should be noted that the broad peak at 3395 cm−1 was nominated as the O‒H stretching vibration of 22HEEAN. The physical properties of PILs are particularly important in view of the use of PILs in FTMs and the main properties of PILs are collected in Table 1. As shown in Table 1, the PILs exhibited low densities and viscosities, which favored the gas permeation through FTMs. As visible,
Table 1 The water content, glass transition temperature (Tg), melting points (Tm), densities and viscosities of PILs investigated in this study. PILs
Water content (wt%)
Tg (°C)
Tm (°C)
ρ (g/cm3)
η (cp, 25 °C)
PAN 2MEAN 22HEEAN
0.58 0.73 0.65
– − 83.8 − 73.3
4.0 – –
1.193 1.251 1.275
64.5 104.2 194.5
Note that the “-” represents that the data cannot be detected.
3. Results and discussion 3.1. Chemical structure and physical properties of PILs The chemical structure of PILs was characterized by the 1H NMR and FTIR spectra, which confirmed the successful proton transfer from the Brønsted acid to Brønsted base. As shown in Fig. 5a, the ratio of integrations of the peaks in the same PIL was in agreement with the ratio of the number of hydrogen atoms. The chemical shifts of CH2 and CH3 groups in three PILs located in the range from 0.901 to 3.512 ppm. As for the 1H NMR spectrum of PAN, the chemical shifts at 0.901, 1.550 and 2.767 ppm were correspond to the C1H, C2H and C3H groups, 79
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 6. The viscosities of PILs at different temperatures (a) and the TGA curves of PILs (b).
Fig. 7. The absorption isotherms of C2H4 (a) and C2H6 (b) at 298.15 K (Note: the dashed line is fitted curve).
chain leads to an increase in the viscosity. It was worth noting that the presence of one extra ether group (in 2MEAN) or hydroxyl group (in 22HEEAN) led to an increase in viscosity probably due to the establishment of extra hydrogen bonds. The effect of temperature on the viscosity was also highlighted. As expected, the viscosities of three PILs decreased with the temperature increasing (Fig. 6a). Furthermore, the viscosity decreased with addition of AgNO3, which was probably because of the destruction of cation-anion interactions. Fig. 6b confirmed the satisfactory thermal stability of the investigated PILs. The decomposition temperatures of PILs and AgNO3/PIL mixture were higher than 150 °C, which assured the thermal stability of PIL-FTMs. 3.2. Gas absorption in the PIL and AgNO3/PIL The gas permeability through PIL-FTMs depends on both thermodynamic and kinetic mechanisms, the gas absorption and the gas diffusion, respectively. Therefore, it is reasonable and necessary to measure the gas solubility in the AgNO3/PIL. Fig. 7 clearly showed the solubility of C2H4 in such AgNO3/PIL was much higher than that of C2H6. The gas pressure presented a positive effect on the solubilities of both gases. The C2H4 equilibrium data clearly demonstrated the combination of chemical and physical absorption (Fig. 7a). In contrast, the C2H6 solubility exhibited a linear increase with the pressure, which could be simply described by Henry's law and implied a physical absorption (Fig. 7b). As also shown in Fig. 7a, the C2H4 solubility was enhanced with the AgNO3 concentration increasing, which was because
Fig. 8. The 1H NMR investigation of dissolving behavior of C2H4 in the AgNO3/ 2MEAN.
the order of the measured densities followed the trend: 22HEEAN > 2MEAN > PAN, which indicated the closer ionic package and stronger ionic interactions in the 22HEEAN. The change tendency of viscosity was the same as that of density. The viscosity is dependent on the ionion interactions, and the increase in the length of the hydrocarbon 80
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 9. The gas permeability (a–c) and C2H4/C2H6 selectivity (d) of the PIL based membranes. (Conditions: 298.15 K, 60 mL/min (50:50 vol% C2H4/C2H6) feed stream and 20 mL/min sweep gas; it should be noted that the corresponding partial pressures of C2H4 and C2H6 were both 0.55, 0.6, 0.65, 0.7, 0.75 bar).
Fig. 10. The gas separation performances of PIL-FTMs: (a) permeability, (b) C2H4/C2H6 selectivity. (Conditions: 1.5 mol/L silver salt concentration, 0.1 bar transmembrane pressure, 298.15 K, 60 mL/min (50:50 vol% C2H4/C2H6) feed stream, 20 mL/min sweep gas; it should be noted that the corresponding partial pressures of C2H4 and C2H6 were both 0.55 bar).
formation mechanism, which leads to the up-field shift of ethylene hydrogens. The area of these peaks qualitatively indicated the amount of ethylene absorbed, which revealed that the C2H4 solubility was mainly dominated by the chemical absorption.
of that more carriers were available to coordinate with C2H4. However, AgNO3 content had a complicated effect on the solubility of C2H6. The solubility of C2H6 decreased as the AgNO3 concentration increased from 0 to 2 mol/L, and then increased slightly with the AgNO3 concentration increasing up to 3 mol/L, which was probably ascribed to the structural variation of AgNO3/PIL (Fig. 7b). The dissolving behavior of C2H4 in the AgNO3/2MEAN was further explored by 1H NMR. As shown in Fig. 8, two new peaks appeared at 5.396 and 4.890 ppm with adding C2H4 to the AgNO3/2MEAN, which were assigned to physical and chemical absorption of C2H4, respectively. The C2H4 forms complexes with Ag+ via a π-bond complex
3.3. Gas permeation of pure PIL based membranes The gas separation performances of pure PIL based membranes are plotted in Fig. 9a-d. It is well known that three pure PILs interact only physically with C2H4 and C2H6, thus gas permeation in these PILs is based on a simple solution-diffusion mechanism [32]. As expected, the 81
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 11. (a) Gas permeabilities, (b) solubilities, (c) diffusivities, and (d) C2H4/C2H6 selectivities of the 2MEAN based PIL-FTMs with different silver salt concentrations. (Conditions: 0.1 bar transmembrane pressure, 298.15 K, 60 mL/min (50:50 vol% C2H4/C2H6) feed stream, 20 mL/min sweep gas; it should be noted that the corresponding partial pressures of C2H4 and C2H6 were both 0.55 bar).
presence of functional groups such as ether group and hydroxyl group of the cations of PILs significantly enhanced the facilitated C2H4 transport.
permeability of C2H4 and C2H6 almost remained unchanged with the transmembrane pressure changing from 0.1 to 0.5 bar (Fig. 9a–c). Therefore, the C2H4/C2H6 selectivity remained unchanged in that pressure range, which indicated the facilitated transport was not taking place. This behavior has already been reported by Urtiaga et al. in the CO/N2 separation using copper(I)-containing FTMs [33]. Moreover, the C2H4 and C2H6 permeabilities followed the sequence of PAN > 2MEAN > 22HEEAN. The permeabilities were opposite to the viscosities of PILs, indicating that diffusion played a vital role in gas permeation. It also should be noted that the C2H4/C2H6 selectivity of the three PILs based membranes varied from 2 to 3.3, which revealed the poor C2H4/C2H6 separation abilities of pure PILs (Fig. 9d). Therefore, the transport carrier was needed to enhance the C2H4/C2H6 selectivity of the membranes.
3.4.2. Effect of silver salt concentration on separation performances of PILFTMs Fig. 11 shows the gas separation performances of the PIL-FTMs with different silver salt concentrations. As expected, the C2H4 permeability increased from 11 to 172 Barrers with the silver salt concentration increasing from 0 to 3 mol/L, which could be clearly attributed to sharp increase in the gas solubility (Fig. 11a-b). With the silver salt concentration increasing, more carriers are available to coordinate with C2H4 and provided higher solubility, thus increasing permeability of C2H4 through the PIL-FTMs. The C2H6 permeability decreased initially and then increased as silver salt concentration changes from 0 to 3 mol/ L, obtaining a minimum at the silver salt concentration of 2 mol/L. The same trend was observed for C2H6 solubility. As seen from Fig. 11c, the C2H6 diffusivity exhibited a similar trend with its solubility and permeability, the C2H6 diffusivity at the silver salt concentrations of 1 mol/ L and 3 mol/L increased by 21% and 42% compared with that when the silver salt concentration was 1 mol/L, respectively. It should be also noted that the C2H4 diffusivity was higher than that of C2H6 through the membranes without carrier. However, upon introducing the AgNO3 into PILs, the C2H4 diffusivity was always smaller than that of C2H6, which was probably due to the larger size of [Ag(C2H4)n]+ complex (Fig. 11c). It should be also noted that the gas diffusivity changed slightly as the silver salt concentration changed, which was in accordance with generally accepted behavior for silver salt/IL membranes [17]. This behavior could be explained by high compatibility of ILs and silver salts at high concentrations due to the same anion used [34]. Moreover, the data on permeability selectivity, solubility selectivity, and diffusivity selectivity are also summarized in Fig. 11d, and it was apparent that the C2H4/C2H6 selectivity obtained a maximum at silver
3.4. Gas permeation through PIL-FTMs 3.4.1. Effect of PIL structure on the separation performances of PIL-FTMs The structure-performance relationships of PIL-FTMs were investigated intensively, which offered insight into smart selection of PILs for better performance. As shown Fig. 10, the cations of PILs had a great effect on the gas separation performances of PIL-FTMs. For all three PILs, the C2H4 permeability was always almost one order of magnitude higher than that of C2H6 because of different transport mechanisms. In addition to normal Fick transport, C2H4 can reversibly coordinate with the carrier inside FTMs, thus facilitating the transport of C2H4. The permeabilities of C2H4 through PAN, 2MEAN and 22HEEAN based PILFTMs at 0.1 bar were 178, 55 and 11 Barrers, respectively (Fig. 10a). The permeability of C2H6 also decreased in the following order: PAN > 2MEAN > 22HEEAN. As also seen from Fig. 10b, the C2H4/ C2H6 selectivity varied obviously with different PILs, e.g. 12 for PAN, 24 for 2MEAN and 42 for 22HEEAN. It indicated that the C2H4/C2H6 selectivity could be tuned by the structural variation of PILs and the 82
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 12. The effect of transmembrane pressure (a) and operating temperature (b, c) on the gas permeability and C2H4/C2H6 selectivity of 2MEAN based PIL-FTMs. The logarithm of the permeability against the inverse of the temperature (d). (Conditions: 2 mol/L silver salt concentration, 0.1 bar transmembrane pressure or 298.15 K, 60 mL/min (50:50 vol% C2H4/C2H6) feed stream, 20 mL/min sweep gas; it should be noted that the corresponding partial pressures of C2H4 and C2H6 were both 0.55, 0.6, 0.65, 0.7, 0.75 bar).
C2H4/C2H6 selectivity gradually decreased with the increasing temperature. It is known that the C2H4/C2H6 selectivity is the combination of solubility selectivity with diffusivity selectivity [37,38]. As the temperature increased, the reaction between carrier and C2H4 was weakened and the solubility of C2H4 was reduced dramatically as well. Whereas C2H6 solubility nearly remained unchanged since it was a physical absorption [5]. Therefore, the decrease of C2H4/C2H6 selectivity was originated from the decrease of solubility selectivity. The logarithm of the permeability followed a linear relationship with the inverse of the temperature as shown in Fig. 12d. Therefore, the temperature influence on the gas permeability is well described in terms of an Arrhenius type relationship [39]. The activation energies of permeations were 9.07 kJ/mol and 15.18 kJ/mol for C2H4 and C2H6, respectively, which illustrated the effect of temperature on gas permeability quantitatively.
salt concentration of 2 mol/L. Comparing the solubility selectivity with the diffusivity selectivity, it could be concluded that the C2H4/C2H6 selectivity of the PIL-FTMs was essentially dominated by the solubility selectivity.
3.4.3. Effects of operation conditions on separation performances of PILFTMs The effect of transmembrane pressure on the gas separation performances of PIL-FTMs is demonstrated in Fig. 12a. Intriguingly, the C2H4 permeability dramatically decreased from 80.15 to 58.53 Barrers as the transmembrane pressure difference increased from 0.1 to 0.5 bar. This is a typical feature of the facilitated transport mechanism, since the carrier tends to be more easily reacted with C2H4 and gets saturated at high C2H4 pressure [35,36]. The variation trend of C2H4 permeability could be also explained by C2H4 absorption isotherm (Fig. 7), for example, when the C2H4 partial pressures were at 0.535 and 0.781bar0.535 to 0.781 the C2H4 solubilities were 0.5387 and 0.6050 mol/L, respectively. In other words, the solubility/C2H4 partial pressure decreased by 23.2% (from 1.0069 to 0.7749), which was well in agreement with C2H4 Permeability. On the other hand, the permeability of C2H6 almost did not change with the transmembrane pressure difference, which implied that the facilitated transport was absent in this case and was in accordance with our previous research [6]. Therefore, the C2H4/C2H6 selectivity decreased by 27% (from 42.3 to 30.6) with the transmembrane pressure increasing. The effect of operating temperature was also investigated (Fig. 12bc). It was found that the permeabilities of C2H4 and C2H6 increased with the increasing temperature, probably due to the fact that the AgNO3/ PIL viscosity decreased with the temperature increasing and the gas diffusion was enhanced as a result. It was noted that the enhancement of C2H4 permeability was smaller than that of C2H6. As a result, the
3.5. Long-term stability of PIL-FTMs The long-term stability of PIL-FTMs were confirmed by the C2H4/ C2H6 separation performance over time, the color change of membranes and the ATR-FTIR spectroscopy. As shown in Fig. 13A, the gas permeability and selectivity remained almost constant during long-time run for about 170 h, which was due to the useful protic acidic properties of PILs [12,17,22]. As seen from Fig. 13B, the appearance of PIL-FTMs did not changed at all after stored for 180 days, indicating the PIL-FTMs were highly stable. Furthermore, as suggested by the ATR-FTIR spectra, no noticeable changes of chemical structures of PIL-FTMs could be detected during 180 days, which further proved the structural stability. The reusability of the PIL-FTMs was also tested. The PIL-FTMs were stored without any protection for six months and the gas separation performances were measured monthly using the same membranes 83
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 13. (A) The long-term stability of PIL-FTMs; (B) the digital photographs and FTIR spectra of PIL-FTMs at different periods; (C, D) the reusability of the PIL-FTMs. (Conditions: 2 mol/L silver concentration, 298 K, 0.1 bar transmembrane pressure, 60 mL/min (50:50 vol% C2H4/C2H6) feed stream, 20 mL/min sweep gas; it should be noted that the corresponding partial pressures of C2H4 and C2H6 were both 0.55 bar).
Fig. 14. Comparison of ethylene separation performance utilizing PIL-FTMs and other membranes in the literatures. Data are plotted on a log-log scale.
throughout the experiment. Fig. 13C-D clearly showed that the gas permeability and selectivity were almost unchanged during six months. The above results clearly indicated that the prepared membranes had great potential for long-time operation.
FTMs measured in this study were much higher than most of results in the literature. Even compared to the recent reported AgCF3SO3-acetamide deep eutectic solvents (DESs) based membranes [5], our membranes still performed comparable C2H4 permeability and high C2H4/ C2H6 selectivity. It is worthy of noting that AgNO3 is much cheaper and more available than other common used silver salts such as AgCF3SO3 and AgBF4 [5,13,40].
3.6. Comparison with other studies Our results were also compared with those reported in recent studies [5–7,31,40], which are summarized in Fig. 14. Not surprisingly, the permeability of C2H4 as well as C2H4/C2H6 selectivity through PIL84
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 15. The investigation of AgNO3/PILs structure: the FTIR spectra (a) and the FT-Raman spectra (b). The mechanism of enhancement of facilitated C2H4 transport: the 1H NMR spectra (c), the FTIR spectra (d).
“structure-breaking” in 2MEAN and increase the free volume of the AgNO3/2MEAN, thus increasing the C2H6 solubility and permeability [43]. The enhancement of facilitated C2H4 transport by the ether group and hydroxyl group of PILs was further investigated by the 1H NMR and FTIR spectroscopy. As shown in Fig. 15c, the N‒H of cations shifted down-field with the addition of AgNO3, e. g. from 7.827 to 7.848 ppm for PAN, from 7.900 to 7.962 ppm for 2MEAN and from 7.879 to 7.995 ppm for 22HEEAN, respectively, which confirmed the intermolecular hydrogen bonds between cation of PIL and the NO3- of AgNO3. The hydrogen bonds interactions pulled up the NO3- from the Ag+ and weakened interactions between Ag+ and NO3-, which was conducive to the disassociation of silver salt, thus increasing the number of effective carrier and enhancing the facilitated C2H4 transport [44]. The red shifts of N‒H in FTIR spectra further verified the hydrogen bonds interactions (Fig. 15d). It also should be noted that the down-field shift of O‒H from 4.185 to 4.357 ppm and red shift from 3396 to 3375 cm−1 of O‒H stretching vibration for 22HEEAN suggested the coordinative interactions between carrier and hydroxyl groups. The coordinative interactions weakened the interaction between silver cation and the counter anion (NO3-), thus increasing the interaction between the silver salt and C2H4 and further enhancing the facilitated C2H4 transport [45,46]. Therefore, the 22HEEAN based PILFTMs obtained the highest C2H4/C2H6 selectivity among three PILs.
3.7. Separation mechanisms As discussed above, the gas separation performances of PIL-FTMs were greatly affected by the liquid structure of AgNO3/PILs, which were further determined by the interactions between AgNO3 and PILs. Therefore, FTIR, FT-Raman and 1H NMR spectra were used to obtain a profound understanding of interactions between AgNO3 and PILs. As shown in Fig. 15a, the N-H asymmetric and symmetric stretching vibrations appeared at 3428 and 3049 cm−1 at the silver salt concentration of 2 mol/L, which was the lowest among three different concentrations and suggested the occurrence of the strongest hydrogen bond interactions. As seen from Fig. 15b, for pure 2MEAN, two peaks at 1117 and 966 cm−1 were attributed to the asymmetric and symmetric stretching vibrations of C‒O. Upon the mixing of AgNO3 with 2MEAN, these two peaks exhibited red shifts, which indicated the interactions between Ag+ and C‒O. Among different concentrations, AgNO3/ 2MEAN at silver salt concentration of 2 mol/L exhibited the largest red shifts of the C‒O vibrations, from 1117 to 1108 cm−1 and from 966 to 961 cm−1, which meant that the strongest coordinative interactions between Ag+ and C‒O occurred. The strongest hydrogen bond and coordination interactions resulted in the structural compactness AgNO3/2MEAN, which could clearly explain why the solubility and diffusivity of C2H4 obtained minimum values at the silver salt concentration of 2 mol/L. Therefore, at low concentrations, the dissolved AgNO3 is probably “structure-making” in 2MEAN and leads to structural compactness of AgNO3/2MEAN, thus resulting in that the C2H6 solubility and permeability decreased and a salting-out effect was observed [41,42]. However, at high concentration, AgNO3 may act as
4. Conclusions Novel PIL-FTMs were fabricated for the separation of C2H4/C2H6 for 85
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
the first time. The PILs not only exhibited good dissolving capacity of carrier but also could stabilize and activate carrier due to their special solvent properties and useful protic acidity. The complexation ability of silver ions with C2H4 could be significantly enhanced by the presence of functional groups such as ether group and hydroxyl group of the cations of PILs, thus increasing the C2H4/C2H6 selectivity. It was also found that the strongest hydrogen bond and coordination interactions between AgNO3 and 2MEAN were achieved at the silver salt concentration of 2 mol/L, endowing PIL-FTMs with the highest C2H4/C2H6 selectivity. In addition, the increase of transmembrane pressure decreased the permeability of C2H4 and selectivity of C2H4/C2H6, while the increase of temperature increased the permeability of C2H4 but decreased the selectivity of C2H4/C2H4. Together with low cost of PILs and AgNO3, the excellent separation performance and good long-term stability would make this new PIL-FTMs move a step towards industrialization.
[19]
[20]
[21]
[22]
[23]
[24]
[25]
Acknowledgments [26] [27]
We are grateful for financial support from the National Key R&D Program of China (Nos. 2016YFC0400406 and 2017B0602702).
[28]
References
[29]
[1] B. Li, Y. Zhang, R. Krishna, K. Yao, Y. Han, Z. Wu, D. Ma, Z. Shi, T. Pham, B. Space, J. Liu, P.K. Thallapally, J. Liu, M. Chrzanowski, S. Ma, Introduction of pi-complexation into porous aromatic framework for highly selective adsorption of ethylene over ethane, J. Am. Chem. Soc. 136 (2014) 8654–8660. [2] C. Ma, X. Wang, X. Wang, B. Yuan, Y. Wu, Z. Li, Novel glucose-based adsorbents (Glc-As) with preferential adsorption of ethane over ethylene and high capacity, Chem. Eng. Sci. 172 (2017) 612–621. [3] A. Luna-Triguero, J.M. Vicent-Luna, P. Gomez-Alvarez, S. Calero, Olefin/paraffin separation in open metal site Cu-BTC metal-organic framework, J. Phys. Chem. C. 121 (2017) 3126–3132. [4] X. Ma, S. Williams, X. Wei, J. Kniep, Y.S. Lin, Propylene/propane mixture separation characteristics and stability of carbon molecular sieve membranes, Ind. Eng. Chem. Res. 54 (2015) 9824–9831. [5] B. Jiang, H. Dou, B. Wang, Y. Sun, Z. Huang, H. Bi, L. Zhang, H. Yang, Silver-based deep eutectic solvents as separation media: supported liquid membranes for facilitated olefin transport, ACS Sustain. Chem. Eng. 5 (2017) 6873–6882. [6] B. Jiang, H. Dou, L. Zhang, B. Wang, Y. Sun, H. Yang, Z. Huang, H. Bi, Novel supported liquid membranes based on deep eutectic solvents for olefin-paraffin separation via facilitated transport, J. Membr. Sci. 536 (2017) 123–132. [7] Y. Sun, H. Bi, H. Dou, H. Yang, Z. Huang, B. Wang, R. Deng, L. Zhang, A novel copper(I)-based supported ionic liquid membrane with high permeability for ethylene/ethane separation, Ind. Eng. Chem. Res. 56 (2017) 741–749. [8] L.C. Tome, D. Mecerreyes, C.S.R. Freire, L.P.N. Rebelo, I.M. Marrucho, Polymeric ionic liquid membranes containing IL-Ag+ for ethylene/ethane separation via olefin-facilitated transport, J. Mater. Chem. A 2 (2014) 5631–5639. [9] C. Zhang, Y. Dai, J.R. Johnson, O. Karvan, W.J. Koros, High performance ZIF-8/ 6FDA-DAM mixed matrix membrane for propylene/propane separations, J. Membr. Sci. 389 (2012) 34–42. [10] A. Cadiau, K. Adil, P.M. Bhatt, Y. Belmabkhout, M. Eddaoudi, A metal-organic framework-based splitter for separating propylene from propane, Science 353 (2016) 137–140. [11] J. Liu, Y. Xiao, T.-S. Chung, Flexible thermally treated 3D PIM-CD molecular sieve membranes exceeding the upper bound line for propylene/propane separation, J. Mater. Chem. A 5 (2017) 4583–4595. [12] T.C. Merkel, R. Blanc, I. Ciobanu, B. Firat, A. Suwarlim, J. Zeid, Silver salt facilitated transport membranes for olefin/paraffin separations: carrier instability and a novel regeneration method, J. Membr. Sci. 447 (2013) 177–189. [13] L.M. Galán. Sánchez, G.W. Meindersma, A.B. Haan, Potential of silver-based roomtemperature ionic liquids for ethylene/ethane separation, Ind. Eng. Chem. Res. 48 (2009) 10650–10656. [14] Y. Li, S. Wang, G. He, H. Wu, F. Pan, Z. Jiang, Facilitated transport of small molecules and ions for energy-efficient membranes, Chem. Soc. Rev. 44 (2015) 103–118. [15] J.H. Kim, B.R. Min, J. Won, Y.S. Kang, Complexation mechanism of olefin with silver ions dissolved in a polymer matrix and its effect on facilitated olefin transport, Chem.-Eur. J. 8 (2002) 650–654. [16] R. Zarca, A. Ortiz, D. Gorri, I. Ortiz, A practical approach to fixed-site-carrier facilitated transport modeling for the separation of propylene/propane mixtures through silver-containing polymeric membranes, Sep. Purif. Technol. 180 (2017) 82–89. [17] M. Fallanza, A. Ortiz, D. Gorri, I. Ortiz, Experimental study of the separation of propane/propylene mixtures by supported ionic liquid membranes containing Ag+–RTILs as carrier, Sep. Purif. Technol. 97 (2012) 83–89. [18] S. Kasahara, E. Kamio, R. Minami, H. Matsuyama, A facilitated transport ion-gel
[30]
[31]
[32]
[33] [34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42] [43]
[44] [45]
[46]
86
membrane for propylene/propane separation using silver ion as a carrier, J. Membr. Sci. 431 (2013) 121–130. Y. Sung Park, Y. Soo Kang, S. Wook Kang, Cost-effective facilitated olefin transport membranes consisting of polymer/AgCF3SO3/Al(NO3)3 with long-term stability, J. Membr. Sci. 495 (2015) 61–64. B. Jose, J.H. Ryu, B.G. Lee, H. Lee, Y.S. Kang, H.S. Kim, Effect of phthalates on the stability and performance of AgBF4-PVP membranes for olefin/paraffin separation, Chem. Commun. (2001) 2046–2047. S.W. Kang, J.H. Kim, J. Won, Y.S. Kang, Suppression of silver ion reduction by Al (NO3)3 complex and its application to highly stabilized olefin transport membranes, J. Membr. Sci. 445 (2013) 156–159. J.H. Kim, B.R. Min, H.S. Kim, J. Won, Y.S. Kang, Facilitated transport of ethylene across polymer membranes containing silver salt: effect of HBF4 on the photoreduction of silver ions, J. Membr. Sci. 212 (2003) 283–288. S. Jeong, S.W. Kang, Effect of Ag2O nanoparticles on long-term stable polymer/ AgBF4/Al(NO3)3 complex membranes for olefin/paraffin separation, Chem. Eng. J. 327 (2017) 500–504. S.W. Kang, K. Char, J.H. Kim, C.K. Kim, Y.S. Kang, Control of ionic interactions in silver salt-polymer complexes with ionic liquids: implications for facilitated olefin transport, Chem. Mater. 18 (2006) 1789–1794. S.W. Kang, J.H. Kim, K. Char, Y.S. Kang, Chemical activation of AgNO3 to form olefin complexes induced by strong coordinative interactions with phthalate oxygens of poly(ethylene phthalate), Ind. Eng. Chem. Res. 45 (2006) 4011–4014. A.S. Amarasekara, Acidic ionic liquids, Chem. Rev. 116 (2016) 6133–6183. J. Reid, F. Agapito, C.E.S. Bernardes, F. Martins, A.J. Walker, S. Shimizu, M.E.M. da Piedade, Structure-property relationships in protic ionic liquids: a thermochemical study, Phys. Chem. Chem. Phys. 19 (2017) 19928–19936. L. Zhang, L. He, C.-B. Hong, S. Qin, G.-H. Tao, Brønsted acidity of bio-protic ionic liquids: the acidic scale of [AA]X amino acid ionic liquids, Green. Chem. 17 (2015) 5154–5163. C. Chiappe, A. Mezzetta, C.S. Pomelli, G. Iaquaniello, A. Gentile, B. Masciocchi, Development of cost-effective biodiesel from microalgae using protic ionic liquids, Green. Chem. 18 (2016) 4982–4989. B. Jiang, H. Yang, L. Zhang, R. Zhang, Y. Sun, Y. Huang, Efficient oxidative desulfurization of diesel fuel using amide-based ionic liquids, Chem. Eng. J. 283 (2016) 89–96. R. Deng, Y. Sun, H. Bi, H. Dou, H. Yang, B. Wang, W. Tao, B. Jiang, Deep eutectic solvents as tuning media dissolving Cu+ used in facilitated transport supported liquid membrane for ethylene/ethane separation, Energy Fuel 31 (2017) 11146–11155. X. Zhang, Z. Tu, H. Li, K. Huang, X. Hu, Y. Wu, D.R. MacFarlane, Selective separation of H2S and CO2 from CH4 by supported ionic liquid membranes, J. Membr. Sci. 543 (2017) 282–287. G. Zarca, I. Ortiz, A. Urtiaga, Copper(I)-containing supported ionic liquid membranes for carbon monoxide/nitrogen separation, J. Membr. Sci. 438 (2013) 38–45. C. Chiappe, M. Malvaldi, B. Melai, S. Fantini, U. Bardi, S. Caporali, An unusual common ion effect promotes dissolution of metal salts in room-temperature ionic liquids: a strategy to obtain ionic liquids having organic–inorganic mixed cations, Green Chem. 12 (2010) 77–80. S. Kasahara, E. Kamio, A.R. Shaikh, T. Matsuki, H. Matsuyama, Effect of the aminogroup densities of functionalized ionic liquids on the facilitated transport properties for CO 2 separation, J. Membr. Sci. 503 (2016) 148–157. K. Huang, X.-M. Zhang, Y.-X. Li, Y.-T. Wu, X.-B. Hu, Facilitated separation of CO2 and SO2 through supported liquid membranes using carboxylate-based ionic liquids, J. Membr. Sci. 471 (2014) 227–236. A. Ilyas, N. Muhammad, M.A. Gilani, K. Ayub, I.F.J. Vankelecom, A.L. Khan, Supported protic ionic liquid membrane based on 3-(trimethoxysilyl)propan-1aminium acetate for the highly selective separation of CO2, J. Membr. Sci. 543 (2017) 301–309. X.-M. Zhang, Z.-H. Tu, H. Li, L. Li, Y.-T. Wu, X.-B. Hu, Supported protic-ionic-liquid membranes with facilitated transport mechanism for the selective separation of CO2, J. Membr. Sci. 527 (2017) 60–67. E. Santos, J. Albo, A. Irabien, Acetate based Supported Ionic Liquid Membranes (SILMs) for CO2 separation: influence of the temperature, J. Membr. Sci. 452 (2014) 277–283. L.C. Tomé, D. Mecerreyes, C.S.R. Freire, L.P.N. Rebelo, I.M. Marrucho, Polymeric ionic liquid membranes containing IL–Ag+ for ethylene/ethane separation via olefin-facilitated transport, J. Mater. Chem. A 2 (2014) 5631–5639. T. Mendez-Morales, J. Carrete, J.R. Rodriguez, O. Cabeza, L.J. Gallego, O. Russina, L.M. Varela, Nanostructure of mixtures of protic ionic liquids and lithium salts: effect of alkyl chain length, Phys. Chem. Chem. Phys. 17 (2015) 5298–5307. R. Hayes, S.A. Bernard, S. Imberti, G.G. Warr, R. Atkin, Solvation of inorganic nitrate salts in protic ionic liquids, J. Phys. Chem. C. 118 (2014) 21215–21225. T. Murphy, S.K. Callear, G.G. Warr, R. Atkin, Dissolved chloride markedly changes the nanostructure of the protic ionic liquids propylammonium and ethanolammonium nitrate, Phys. Chem. Chem. Phys. 18 (2016) 17169–17182. R. Faiz, K. Li, Olefin/paraffin separation using membrane based facilitated transport/chemical absorption techniques, Chem. Eng. Sci. 73 (2012) 261–284. J.H. Kim, B.R. Min, K.B. Lee, J. Won, Y.S. Kang, Coordination structure of various ligands in crosslinked PVA to silver ions for facilitated olefin transport, Chem. Commun. (2002) 2732–2733. S.W. Kang, J.H. Kim, J. Won, K. Char, Y.S. Kang, Effect of amino acids in polymer/ silver salt complex membranes on facilitated olefin transport, J. Membr. Sci. 248 (2005) 201–206.
Contents lists available at ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
Ultra-stable and cost-efficient protic ionic liquid based facilitated transport membranes for highly selective olefin/paraffin separation
T
Haozhen Doua,1, Bin Jianga, Xiaoming Xiaoa, Mi Xua,1, Baoyu Wangb, Li Haoa, Yongli Suna, ⁎ Luhong Zhanga, a b
School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China School of Chemical Engineering and Food Science, Zhengzhou Institute of Technology, Zhengzhou 450000, China
A R T I C LE I N FO
A B S T R A C T
Keywords: Facilitated transport membranes Protic ionic liquids, olefin/paraffin separation Long-term stability
Facilitated transport membranes (FTMs) for olefin/paraffin separations have failed to achieve commercial success due to the instability of carriers although great efforts have been made. In this work, ultra-stable and cost-efficient protic ionic liquid based FTMs (PIL-FTMs) were firstly prepared by utilizing the Brönsted acidic property of PILs to stabilize the carrier. The gas solubility in the carrier/PILs was measured and the separation performances of PIL-FTMs were evaluated systemically. The results indicated that the structure of PILs affected the C2H4 permeability and the presence of ether group and hydroxyl group in PILs significantly enhanced the C2H4/C2H6 selectivity. The carrier concentration led to structural variation of PIL-FTMs, thus manipulating the gas separation performances of PIL-FTMs. The increase of transmembrane pressure decreased C2H4 permeability and C2H4/C2H6 selectivity, indicating a typical feature of FTMs. The increase of temperature increased the C2H4 permeability but decreased C2H4/C2H6 selectivity. The separation performances of PIL-FTMs were much higher than other results in the literature. Furthermore, the PIL-FTMs exhibited excellent stability during the long-term experiments carried out for six months. Finally, the investigation of separation mechanism revealed that the hydrogen-bonding and coordinative interactions between PILs and carrier accounted for the high separation efficiency of PIL-FTMs. In all, the excellent long-term stability, outstanding separation performances and economic feasibility of PIL-FTMs could play an important role in moving these membranes toward industrial application.
1. Introduction Short-chain olefins are among the most important feedstocks, and widely used to produce plastics, rubber, films and other chemicals [1,2]. Industrial olefin/paraffin separations heavily rely upon the extremely cost and energy intensive cryogenic distillation technology [3]. The development of economically viable olefin/paraffin separation processes is becoming increasingly important [4]. Many novel materials and alternative techniques have been investigated for olefin/paraffin separations, such as deep eutectic solvent based membranes [5,6], ionic liquid based membranes [7], polymeric ionic liquid membranes [8], mixed matrix membranes [9], metal-organic frameworks [10] and so on [11]. Facilitated transport membranes (FTMs) have gained prominence as research targets for olefin/paraffin separations because of their potential to perform separation more efficiently than the passive polymeric
⁎
1
Corresponding author. E-mail address: [email protected] (L. Zhang). These authors contributed equally to this work.
https://doi.org/10.1016/j.memsci.2018.04.015 Received 16 December 2017; Received in revised form 6 April 2018; Accepted 10 April 2018 Available online 12 April 2018 0376-7388/ © 2018 Elsevier B.V. All rights reserved.
membranes [12]. In the FTMs, dispersed transition metal cations, such as silver and copper (I) ions, can coordinate with olefins based on a πbond complex formation mechanism, thus significantly increasing the gas separation performances of FTMs [13]. Usually, according to the mobility of carriers, FTMs can be classified into two types: fixed-site carrier membranes and mobile carrier membranes [14]. Polymeric electrolyte membranes, as a typical example of fixed-site carrier FTMs, have been investigated in great detail over the past decades [15,16]. It was found that silver salts added in the polymer electrolytes effectively facilitated olefin transport. Moreover, mobile carrier FTMs can be obtained by adding silver salts into aprotic ionic liquids (AILs). Ortiz et al. firstly reported the separation of propane/propylene mixture by mobile carrier FTMs containing AgBF4 as carrier [17]. Matsuyama et al. designed mobile carrier facilitated transport ion-gel membranes for propylene/propane separation by introducing gelator into AILs [18]. Despite the intensive research of FTMs outlined above, the
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
instability of carrier is still the most challenge for the industrial application of FTMs [12]. The silver ions can be easily reduced to silver particles even under dark conditions, resulting in deterioration of separation performances of FTMs [19]. However, there have been few studies of carrier stability and little is known about how to stabilize the silver ions. Kim et al. firstly reported the reduction of silver ions could be restrained by phthalates but after one week silver particles still appeared [20]. Subsequently, Kang et al. reported the stability of FTMs could be improved by additives such as HBF4 and Al(NO3)3, which could be kept for 100–200 h [21,22], but these additives resulted in relatively low separation performances of FTMs [23]. Very recently, the instability of FTMs was solved by a regeneration method that used a peroxide/acid liquid or vapor phase treatment to oxidize reduced silver carriers within the FTMs [12]. However, this method suffers from many operating drawbacks and challenges which consequently increase the high capital and operating costs. Besides the challenge of carrier instability, economic feasibility of carrier is another decisive factor for the large-scale application of FTMs. Among the many silver salts used as carriers, AgBF4 and AgCF3SO3 have been commonly used because of their highly favorable ability to facilitate olefin transport [17,18,24]. However, their high cost has created another barrier for the large-scale application of FTMs. This barrier can be easily overcome by AgNO3, which has advantages of low cost and facile availability [25]. Unfortunately, AgNO3 is rather inactive in facilitated transport of olefin due to its high lattice energy [25]. Therefore, exploitation of greener and more sustainable solvents to stabilize and activate the carrier of AgNO3 is in high urgency for designing novel and promising mobile carrier FTMs. Protic ionic liquids (PILs) have aroused our great interest due to their useful Brönsted acidic property, which is expected to stabilize silver ions [26]. PILs are a subset of ILs formed by proton transfer from the Brønsted acid to Brønsted base [27]. Their straightforward preparation reduces or eliminates additional work-up stages, such as solvent separation and ion-exchange reactions, which makes them significantly less expensive than common AILs [28]. Moreover, PILs are often characterized by a reduced environmental impacts, such as lower toxicity and higher biodegradability with respect to the corresponding AILs [29]. And most importantly, the structural designability of PILs endows their ability to tune the activity of AgNO3. Compared with AILs, which are widely applied in the olefin/paraffin separations [17,18], PILs have never been reported for olefin/paraffin separations. In this study, a series of PIL-FTMs were designed and fabricated for the C2H4/C2H6 separation. To the best of our knowledge, this is the first report on the fabrication of PIL-FTMs exhibiting excellent long-term stability, C2H4 permeability and C2H4/C2H6 selectivity. The gas solubility in AgNO3/PILs was measured and the gas permeability as well as the selectivity of PIL-FTMs were investigated. The structure-property relationship of PIL-FTMs has been highlighted, the main affected factors of the separation process were optimized systemically and the longterm stability of PIL-FTMs was evaluated intensively. Finally, the facilitated transport mechanism was analyzed by various characterization techniques, which shed insight into the smart molecule design of novel FTMs for the C2H4/C2H6 separation.
Fig. 1. A synthetic route to prepare PILs through the proton-transfer reaction (a); the chemical structure of PILs studied in this work (b).
obtained from Haining Zhongli Filtering equipment Corporation (China). 2.2. Synthesis of PILs All the PILs were prepared according to a similar procedure in the literature [30]. As a typical example, the aqueous solution of nitric acid (0.2 mol) was added dropwise into 2-methoxyethylamine (0.2 mol) using methanol as solvent at 0 °C or below. Then, the reaction was carried out under stirring vigorously at RT for 12 h (Fig. 1). After the reaction was completed, the product was purified by rotary evaporating at 60 °C and then dried under vacuum at 80 °C for 48 h to remove the trace of water and obtain colorless and transparent PIL. The water contents of the PILs were determined to be 0.25–0.58 wt% by Karl-Fischertitration method (DL37 KF Coulometer, Mettler Toledo). 2.3. PIL-FTMs preparation PIL-FTMs were prepared using the similar procedure reported by our group previously [5,6]. Initially, AgNO3 was dissolved in the PIL under magnetic stirring at RT to obtain reactive AgNO3/PIL mixture. Then, after the residual water and other trace volatile compounds inside the pores of nylon membranes were removed under vacuum (< 2 mbar) at 60 °C for 1 h, 3 mL of AgNO3/PIL mixture was spread onto the top surface of the treated membrane support and impregnated into membrane pores by using nitrogen gas with a transmembrane pressure difference of 1 bar in the permeation cell for 10 min. In order to ensure the completed impregnation, the above procedure was carried out several times until getting a thin and visible layer liquid on the bottom surface of the membrane support. Finally, the excessive AgNO3/PIL mixture was removed from the membrane surface using tissue paper. As shown in Fig. 2, the nylon membrane was a highly porous supported material with a typical sponge-like structure. After the impregnation, the pores of the nylon membrane were homogeneously and completely saturated by AgNO3/PIL mixture, which confirmed the successful fabrication of PIL-FTMs.
2. Experimental 2.1. Materials
2.4. Determination of physical properties Ethylene and ethane gases (99.9 mol%) were purchased from Tianjin Tang Dynasty Gas Co., Ltd (China). Propylamine (99 wt%), 2methoxyethylamine (99.5 wt%), 2-(2-Aminoethoxy) ethanol (99 wt%) and nitric acid (65–68 wt%) were all supplied from Shanghai Aladdin Biochem Technology Co., Ltd. (China) and used as received. The microporous membrane support was hydrophilic nylon flat sheet membrane with porosity of 70%, the average thickness of 100 µm, a nominal pore size of 0.1 µm and effective area of 19.625 cm2, which was
The density measurements were conducted at 25 °C with an Anton Paar DMA 5000 type automatic densimeter with a precision of 0.0001 g/cm3, which was calibrated with distilled water prior to the measurements. The viscosities were measured on a Brookfield LVDV-II +Pro viscometer with an uncertainty of ± 1% in relation to full scale. The melting points of PILs were determined by a DSC 204 differential scanning calorimeter. 77
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 2. SEM images of the surface for Nylon membrane (a) and PIL-FTMs (b).
gas at different equilibrium states (P0, P1 and P2), which can be calculated by the Eq. (3):
2.5. Characterization The 1H NMR spectra were recorded on the VARIA INOVA 500 MHz spectrometer using d6-DMSO as the solvent with TMS as the internal standard. The attenuated total reflection infrared spectra (ATR-IR) were recorded on a Bio-Rad FTS 6000 FT-IR spectrometer. Raman measurements were carried out by a Bruker RFS 100 Fourier transform Raman spectrometer. The microscopic morphology of the PIL-FTMs was characterized by the Hitachi S-4800 field emission scanning electron microscope (SEM). Thermogravimetric measurements were conducted on a Netzsch TG 209 thermal gravimetric analyser.
Zi = 1+
Ci =
ni VPILs
BC 2H 4 = 8.7776·10−6.0061·10 4 / T +1.2857·106/ T 2−1.0861·109/ T 3
(4)
BC 2H 6 = 1.0773·102−8.2548·10 4 / T +5.2387·106/ T 2−1.9764·109/ T 3
(5)
Where BC2H4 and BC2H4 are the second-order virial coefficient of C2H4 and C2H6, respectively, cm3/mol. T is the temperature (298.15 K).
The gas solubility was determined in light of pressure drop method described in our previous publication [5]. Briefly, the apparatus is illustrated in Fig. 3. The apparatus consisted of an equilibrium vessel (7) with volume of 133.4 cm3 equipped with a magnetic stirrer, and a gas storage vessel (6) for the introduction of known amounts of 467.5 cm3 gas. The pressures in these two vessels were monitored using pressure transducers (Model CYYZ11, China) with an accuracy of 0.001 bar. The temperature was maintained constant by a DC-5060 thermostatic air bath (10) with an accuracy of ± 0.1 K. The solubility of pure gas in PIL or AgNO3/PIL is determined by Eqs. (1) and (2):
P (V − VPILs ) P0 Vr PV − 1 s − 2 r Z0 RT Z1 RT Z2 RT
(3)
Where the subscripts of “0”, “1”, and “2” represent different equilibrium states (P0, P1 and P2). B is the second-order virial coefficient of ethylene or ethane, and its correlation formula for different gases is illustrated by the Eq. (4) and Eq. (5) [31]:
2.6. Gas absorption
ni = ninitial − nfinal =
Pi⋅B (i = 0,1,2) R·T
2.7. Gas permeation Mixed-gas permeation experiments were performed as reported by our group previously [6], and the setup is shown in Fig. 4. In brief, the operation temperature and transmembrane pressure were regulated by a micrometer value and an air blowing thermostatic oven according to the experimental requirements, respectively. The flow rates of individual gases (ethylene, ethane, sweep gas and outlet gas) were controlled and measured by using four mass flow controllers and the composition of permeated gases was analyzed by the online gas chromatograph. Three parallel experiments were carried out to obtain the reliability values. For mixed gas experiments, the permeability coefficient (Pi ) of a particular gas is defined as the permeation flux (Ji ) normalized to the pressure difference across the membrane ( ∆Pi ), as well as the membrane thickness (δ ), which is described by the Eq. (6):
(1) (2)
Where P0 is the pressure of the total gas introduced into the gas storage vessel while P1 and P2 are the pressure of equilibration vessel and gas storage vessel at equilibrium. Vr, Vs, and VPILs are the volume of the equilibration vessel, gas storage vessel, and the absorbents in equilibration vessel. Z0, Z1, and Z2 are the compressibility factors for the pure
Pi = Ji
δ ∆Pi
(6)
The gas transport through FTMs can be interpreted by the solutiondiffusion theory [8,17]. Therefore, the permeability can further be represented as the product of thermodynamic solubility coefficient (S), and kinetic effective diffusion coefficient (D).
Pi = Si⋅Di
(7)
The gas solubility coefficient can be determined by the gas solubility (ni ) . Once determined the gas permeability through the membrane and gas solubility in the PILs, the effective diffusion coefficient can be derived from Eq. (7). The permeability selectivity (αi, j ) is obtained by dividing the permeability of the more permeable specie i to the permeability of the less permeable specie j for the case where the downstream pressure is negligible relative to the upstream feed pressure. As shown in Eq. (8), the αi, j can be also expressed as the product of the diffusivity selectivity and the solubility selectivity.
Fig. 3. Diagram of gas absorption experimental setup. 1-gas bottles, 2-micrometric values, 3-three way valves, 4-fourway valves, 5-vacuum pump, 6-gas storage vessel, 7-gas equilibrium vessel, 8-pressure transducers, 9-magnetic stirring, 10-constant temperature circulating water bath.
αij = 78
Pi S D = i⋅ i Pj Sj Dj
(8)
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 4. Diagram of gas permeation experimental setup.1-gas bottles, 2-buffer tanks, 3-mass flow controllers, 4-two-way valves, 5-static mixer, 6-three-way valves, 7permeation cell, 8-air blowing thermostatic oven, 9-mass flow meter, 10-pressure transducers, 11-micrometric values, 12-gas chromatograph.
Fig. 5. The PILs characterizations: the 1H NMR spectra (a), the FTIR spectra (b).
respectively. For 2MEAN, the chemical shifts of C1H, C2H and C3H groups occurred at 3.005, 3.512 and 3.293 ppm, respectively. In the case of 22HEEAN, the chemical shift at 3.013 ppm was assigned to C3H group, while the peaks around 3.595 ppm were ascribed to C2H groups. As also seen from Fig. 5a, the chemical shifts at 7.827, 7.909 and 7.897 ppm were assigned to the N‒H proton of PAN, 2MEAN and 22HEEAN, respectively. The chemical shift at 4.185 ppm was attributed to the O-H of 22HEEAN. As shown in Fig. 5b, the broad peaks between 3600 and 3000 cm−1 confirmed the formation of hydrogen bonds between the cation and anion of PIL, which maybe existed as N‒H···O and O‒H···O. However, the peaks for three PILs varied slightly, which was attributed to different structural arrangement of hydrogen bonds and electrostatic interactions. For example, the symmetric N‒H stretching vibrations appeared at 3041, 2996, 3058 cm−1 for PAN, 2MEAN and 22HEEAN, respectively. The peaks at 1350, 1382, 1380 cm−1 were ascribed to the N‒O asymmetric stretching vibrations of PAN, 2MEAN and 22HEEAN. It also should be noted that the broad peak at 3395 cm−1 was nominated as the O‒H stretching vibration of 22HEEAN. The physical properties of PILs are particularly important in view of the use of PILs in FTMs and the main properties of PILs are collected in Table 1. As shown in Table 1, the PILs exhibited low densities and viscosities, which favored the gas permeation through FTMs. As visible,
Table 1 The water content, glass transition temperature (Tg), melting points (Tm), densities and viscosities of PILs investigated in this study. PILs
Water content (wt%)
Tg (°C)
Tm (°C)
ρ (g/cm3)
η (cp, 25 °C)
PAN 2MEAN 22HEEAN
0.58 0.73 0.65
– − 83.8 − 73.3
4.0 – –
1.193 1.251 1.275
64.5 104.2 194.5
Note that the “-” represents that the data cannot be detected.
3. Results and discussion 3.1. Chemical structure and physical properties of PILs The chemical structure of PILs was characterized by the 1H NMR and FTIR spectra, which confirmed the successful proton transfer from the Brønsted acid to Brønsted base. As shown in Fig. 5a, the ratio of integrations of the peaks in the same PIL was in agreement with the ratio of the number of hydrogen atoms. The chemical shifts of CH2 and CH3 groups in three PILs located in the range from 0.901 to 3.512 ppm. As for the 1H NMR spectrum of PAN, the chemical shifts at 0.901, 1.550 and 2.767 ppm were correspond to the C1H, C2H and C3H groups, 79
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 6. The viscosities of PILs at different temperatures (a) and the TGA curves of PILs (b).
Fig. 7. The absorption isotherms of C2H4 (a) and C2H6 (b) at 298.15 K (Note: the dashed line is fitted curve).
chain leads to an increase in the viscosity. It was worth noting that the presence of one extra ether group (in 2MEAN) or hydroxyl group (in 22HEEAN) led to an increase in viscosity probably due to the establishment of extra hydrogen bonds. The effect of temperature on the viscosity was also highlighted. As expected, the viscosities of three PILs decreased with the temperature increasing (Fig. 6a). Furthermore, the viscosity decreased with addition of AgNO3, which was probably because of the destruction of cation-anion interactions. Fig. 6b confirmed the satisfactory thermal stability of the investigated PILs. The decomposition temperatures of PILs and AgNO3/PIL mixture were higher than 150 °C, which assured the thermal stability of PIL-FTMs. 3.2. Gas absorption in the PIL and AgNO3/PIL The gas permeability through PIL-FTMs depends on both thermodynamic and kinetic mechanisms, the gas absorption and the gas diffusion, respectively. Therefore, it is reasonable and necessary to measure the gas solubility in the AgNO3/PIL. Fig. 7 clearly showed the solubility of C2H4 in such AgNO3/PIL was much higher than that of C2H6. The gas pressure presented a positive effect on the solubilities of both gases. The C2H4 equilibrium data clearly demonstrated the combination of chemical and physical absorption (Fig. 7a). In contrast, the C2H6 solubility exhibited a linear increase with the pressure, which could be simply described by Henry's law and implied a physical absorption (Fig. 7b). As also shown in Fig. 7a, the C2H4 solubility was enhanced with the AgNO3 concentration increasing, which was because
Fig. 8. The 1H NMR investigation of dissolving behavior of C2H4 in the AgNO3/ 2MEAN.
the order of the measured densities followed the trend: 22HEEAN > 2MEAN > PAN, which indicated the closer ionic package and stronger ionic interactions in the 22HEEAN. The change tendency of viscosity was the same as that of density. The viscosity is dependent on the ionion interactions, and the increase in the length of the hydrocarbon 80
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 9. The gas permeability (a–c) and C2H4/C2H6 selectivity (d) of the PIL based membranes. (Conditions: 298.15 K, 60 mL/min (50:50 vol% C2H4/C2H6) feed stream and 20 mL/min sweep gas; it should be noted that the corresponding partial pressures of C2H4 and C2H6 were both 0.55, 0.6, 0.65, 0.7, 0.75 bar).
Fig. 10. The gas separation performances of PIL-FTMs: (a) permeability, (b) C2H4/C2H6 selectivity. (Conditions: 1.5 mol/L silver salt concentration, 0.1 bar transmembrane pressure, 298.15 K, 60 mL/min (50:50 vol% C2H4/C2H6) feed stream, 20 mL/min sweep gas; it should be noted that the corresponding partial pressures of C2H4 and C2H6 were both 0.55 bar).
formation mechanism, which leads to the up-field shift of ethylene hydrogens. The area of these peaks qualitatively indicated the amount of ethylene absorbed, which revealed that the C2H4 solubility was mainly dominated by the chemical absorption.
of that more carriers were available to coordinate with C2H4. However, AgNO3 content had a complicated effect on the solubility of C2H6. The solubility of C2H6 decreased as the AgNO3 concentration increased from 0 to 2 mol/L, and then increased slightly with the AgNO3 concentration increasing up to 3 mol/L, which was probably ascribed to the structural variation of AgNO3/PIL (Fig. 7b). The dissolving behavior of C2H4 in the AgNO3/2MEAN was further explored by 1H NMR. As shown in Fig. 8, two new peaks appeared at 5.396 and 4.890 ppm with adding C2H4 to the AgNO3/2MEAN, which were assigned to physical and chemical absorption of C2H4, respectively. The C2H4 forms complexes with Ag+ via a π-bond complex
3.3. Gas permeation of pure PIL based membranes The gas separation performances of pure PIL based membranes are plotted in Fig. 9a-d. It is well known that three pure PILs interact only physically with C2H4 and C2H6, thus gas permeation in these PILs is based on a simple solution-diffusion mechanism [32]. As expected, the 81
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 11. (a) Gas permeabilities, (b) solubilities, (c) diffusivities, and (d) C2H4/C2H6 selectivities of the 2MEAN based PIL-FTMs with different silver salt concentrations. (Conditions: 0.1 bar transmembrane pressure, 298.15 K, 60 mL/min (50:50 vol% C2H4/C2H6) feed stream, 20 mL/min sweep gas; it should be noted that the corresponding partial pressures of C2H4 and C2H6 were both 0.55 bar).
presence of functional groups such as ether group and hydroxyl group of the cations of PILs significantly enhanced the facilitated C2H4 transport.
permeability of C2H4 and C2H6 almost remained unchanged with the transmembrane pressure changing from 0.1 to 0.5 bar (Fig. 9a–c). Therefore, the C2H4/C2H6 selectivity remained unchanged in that pressure range, which indicated the facilitated transport was not taking place. This behavior has already been reported by Urtiaga et al. in the CO/N2 separation using copper(I)-containing FTMs [33]. Moreover, the C2H4 and C2H6 permeabilities followed the sequence of PAN > 2MEAN > 22HEEAN. The permeabilities were opposite to the viscosities of PILs, indicating that diffusion played a vital role in gas permeation. It also should be noted that the C2H4/C2H6 selectivity of the three PILs based membranes varied from 2 to 3.3, which revealed the poor C2H4/C2H6 separation abilities of pure PILs (Fig. 9d). Therefore, the transport carrier was needed to enhance the C2H4/C2H6 selectivity of the membranes.
3.4.2. Effect of silver salt concentration on separation performances of PILFTMs Fig. 11 shows the gas separation performances of the PIL-FTMs with different silver salt concentrations. As expected, the C2H4 permeability increased from 11 to 172 Barrers with the silver salt concentration increasing from 0 to 3 mol/L, which could be clearly attributed to sharp increase in the gas solubility (Fig. 11a-b). With the silver salt concentration increasing, more carriers are available to coordinate with C2H4 and provided higher solubility, thus increasing permeability of C2H4 through the PIL-FTMs. The C2H6 permeability decreased initially and then increased as silver salt concentration changes from 0 to 3 mol/ L, obtaining a minimum at the silver salt concentration of 2 mol/L. The same trend was observed for C2H6 solubility. As seen from Fig. 11c, the C2H6 diffusivity exhibited a similar trend with its solubility and permeability, the C2H6 diffusivity at the silver salt concentrations of 1 mol/ L and 3 mol/L increased by 21% and 42% compared with that when the silver salt concentration was 1 mol/L, respectively. It should be also noted that the C2H4 diffusivity was higher than that of C2H6 through the membranes without carrier. However, upon introducing the AgNO3 into PILs, the C2H4 diffusivity was always smaller than that of C2H6, which was probably due to the larger size of [Ag(C2H4)n]+ complex (Fig. 11c). It should be also noted that the gas diffusivity changed slightly as the silver salt concentration changed, which was in accordance with generally accepted behavior for silver salt/IL membranes [17]. This behavior could be explained by high compatibility of ILs and silver salts at high concentrations due to the same anion used [34]. Moreover, the data on permeability selectivity, solubility selectivity, and diffusivity selectivity are also summarized in Fig. 11d, and it was apparent that the C2H4/C2H6 selectivity obtained a maximum at silver
3.4. Gas permeation through PIL-FTMs 3.4.1. Effect of PIL structure on the separation performances of PIL-FTMs The structure-performance relationships of PIL-FTMs were investigated intensively, which offered insight into smart selection of PILs for better performance. As shown Fig. 10, the cations of PILs had a great effect on the gas separation performances of PIL-FTMs. For all three PILs, the C2H4 permeability was always almost one order of magnitude higher than that of C2H6 because of different transport mechanisms. In addition to normal Fick transport, C2H4 can reversibly coordinate with the carrier inside FTMs, thus facilitating the transport of C2H4. The permeabilities of C2H4 through PAN, 2MEAN and 22HEEAN based PILFTMs at 0.1 bar were 178, 55 and 11 Barrers, respectively (Fig. 10a). The permeability of C2H6 also decreased in the following order: PAN > 2MEAN > 22HEEAN. As also seen from Fig. 10b, the C2H4/ C2H6 selectivity varied obviously with different PILs, e.g. 12 for PAN, 24 for 2MEAN and 42 for 22HEEAN. It indicated that the C2H4/C2H6 selectivity could be tuned by the structural variation of PILs and the 82
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 12. The effect of transmembrane pressure (a) and operating temperature (b, c) on the gas permeability and C2H4/C2H6 selectivity of 2MEAN based PIL-FTMs. The logarithm of the permeability against the inverse of the temperature (d). (Conditions: 2 mol/L silver salt concentration, 0.1 bar transmembrane pressure or 298.15 K, 60 mL/min (50:50 vol% C2H4/C2H6) feed stream, 20 mL/min sweep gas; it should be noted that the corresponding partial pressures of C2H4 and C2H6 were both 0.55, 0.6, 0.65, 0.7, 0.75 bar).
C2H4/C2H6 selectivity gradually decreased with the increasing temperature. It is known that the C2H4/C2H6 selectivity is the combination of solubility selectivity with diffusivity selectivity [37,38]. As the temperature increased, the reaction between carrier and C2H4 was weakened and the solubility of C2H4 was reduced dramatically as well. Whereas C2H6 solubility nearly remained unchanged since it was a physical absorption [5]. Therefore, the decrease of C2H4/C2H6 selectivity was originated from the decrease of solubility selectivity. The logarithm of the permeability followed a linear relationship with the inverse of the temperature as shown in Fig. 12d. Therefore, the temperature influence on the gas permeability is well described in terms of an Arrhenius type relationship [39]. The activation energies of permeations were 9.07 kJ/mol and 15.18 kJ/mol for C2H4 and C2H6, respectively, which illustrated the effect of temperature on gas permeability quantitatively.
salt concentration of 2 mol/L. Comparing the solubility selectivity with the diffusivity selectivity, it could be concluded that the C2H4/C2H6 selectivity of the PIL-FTMs was essentially dominated by the solubility selectivity.
3.4.3. Effects of operation conditions on separation performances of PILFTMs The effect of transmembrane pressure on the gas separation performances of PIL-FTMs is demonstrated in Fig. 12a. Intriguingly, the C2H4 permeability dramatically decreased from 80.15 to 58.53 Barrers as the transmembrane pressure difference increased from 0.1 to 0.5 bar. This is a typical feature of the facilitated transport mechanism, since the carrier tends to be more easily reacted with C2H4 and gets saturated at high C2H4 pressure [35,36]. The variation trend of C2H4 permeability could be also explained by C2H4 absorption isotherm (Fig. 7), for example, when the C2H4 partial pressures were at 0.535 and 0.781bar0.535 to 0.781 the C2H4 solubilities were 0.5387 and 0.6050 mol/L, respectively. In other words, the solubility/C2H4 partial pressure decreased by 23.2% (from 1.0069 to 0.7749), which was well in agreement with C2H4 Permeability. On the other hand, the permeability of C2H6 almost did not change with the transmembrane pressure difference, which implied that the facilitated transport was absent in this case and was in accordance with our previous research [6]. Therefore, the C2H4/C2H6 selectivity decreased by 27% (from 42.3 to 30.6) with the transmembrane pressure increasing. The effect of operating temperature was also investigated (Fig. 12bc). It was found that the permeabilities of C2H4 and C2H6 increased with the increasing temperature, probably due to the fact that the AgNO3/ PIL viscosity decreased with the temperature increasing and the gas diffusion was enhanced as a result. It was noted that the enhancement of C2H4 permeability was smaller than that of C2H6. As a result, the
3.5. Long-term stability of PIL-FTMs The long-term stability of PIL-FTMs were confirmed by the C2H4/ C2H6 separation performance over time, the color change of membranes and the ATR-FTIR spectroscopy. As shown in Fig. 13A, the gas permeability and selectivity remained almost constant during long-time run for about 170 h, which was due to the useful protic acidic properties of PILs [12,17,22]. As seen from Fig. 13B, the appearance of PIL-FTMs did not changed at all after stored for 180 days, indicating the PIL-FTMs were highly stable. Furthermore, as suggested by the ATR-FTIR spectra, no noticeable changes of chemical structures of PIL-FTMs could be detected during 180 days, which further proved the structural stability. The reusability of the PIL-FTMs was also tested. The PIL-FTMs were stored without any protection for six months and the gas separation performances were measured monthly using the same membranes 83
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 13. (A) The long-term stability of PIL-FTMs; (B) the digital photographs and FTIR spectra of PIL-FTMs at different periods; (C, D) the reusability of the PIL-FTMs. (Conditions: 2 mol/L silver concentration, 298 K, 0.1 bar transmembrane pressure, 60 mL/min (50:50 vol% C2H4/C2H6) feed stream, 20 mL/min sweep gas; it should be noted that the corresponding partial pressures of C2H4 and C2H6 were both 0.55 bar).
Fig. 14. Comparison of ethylene separation performance utilizing PIL-FTMs and other membranes in the literatures. Data are plotted on a log-log scale.
throughout the experiment. Fig. 13C-D clearly showed that the gas permeability and selectivity were almost unchanged during six months. The above results clearly indicated that the prepared membranes had great potential for long-time operation.
FTMs measured in this study were much higher than most of results in the literature. Even compared to the recent reported AgCF3SO3-acetamide deep eutectic solvents (DESs) based membranes [5], our membranes still performed comparable C2H4 permeability and high C2H4/ C2H6 selectivity. It is worthy of noting that AgNO3 is much cheaper and more available than other common used silver salts such as AgCF3SO3 and AgBF4 [5,13,40].
3.6. Comparison with other studies Our results were also compared with those reported in recent studies [5–7,31,40], which are summarized in Fig. 14. Not surprisingly, the permeability of C2H4 as well as C2H4/C2H6 selectivity through PIL84
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
Fig. 15. The investigation of AgNO3/PILs structure: the FTIR spectra (a) and the FT-Raman spectra (b). The mechanism of enhancement of facilitated C2H4 transport: the 1H NMR spectra (c), the FTIR spectra (d).
“structure-breaking” in 2MEAN and increase the free volume of the AgNO3/2MEAN, thus increasing the C2H6 solubility and permeability [43]. The enhancement of facilitated C2H4 transport by the ether group and hydroxyl group of PILs was further investigated by the 1H NMR and FTIR spectroscopy. As shown in Fig. 15c, the N‒H of cations shifted down-field with the addition of AgNO3, e. g. from 7.827 to 7.848 ppm for PAN, from 7.900 to 7.962 ppm for 2MEAN and from 7.879 to 7.995 ppm for 22HEEAN, respectively, which confirmed the intermolecular hydrogen bonds between cation of PIL and the NO3- of AgNO3. The hydrogen bonds interactions pulled up the NO3- from the Ag+ and weakened interactions between Ag+ and NO3-, which was conducive to the disassociation of silver salt, thus increasing the number of effective carrier and enhancing the facilitated C2H4 transport [44]. The red shifts of N‒H in FTIR spectra further verified the hydrogen bonds interactions (Fig. 15d). It also should be noted that the down-field shift of O‒H from 4.185 to 4.357 ppm and red shift from 3396 to 3375 cm−1 of O‒H stretching vibration for 22HEEAN suggested the coordinative interactions between carrier and hydroxyl groups. The coordinative interactions weakened the interaction between silver cation and the counter anion (NO3-), thus increasing the interaction between the silver salt and C2H4 and further enhancing the facilitated C2H4 transport [45,46]. Therefore, the 22HEEAN based PILFTMs obtained the highest C2H4/C2H6 selectivity among three PILs.
3.7. Separation mechanisms As discussed above, the gas separation performances of PIL-FTMs were greatly affected by the liquid structure of AgNO3/PILs, which were further determined by the interactions between AgNO3 and PILs. Therefore, FTIR, FT-Raman and 1H NMR spectra were used to obtain a profound understanding of interactions between AgNO3 and PILs. As shown in Fig. 15a, the N-H asymmetric and symmetric stretching vibrations appeared at 3428 and 3049 cm−1 at the silver salt concentration of 2 mol/L, which was the lowest among three different concentrations and suggested the occurrence of the strongest hydrogen bond interactions. As seen from Fig. 15b, for pure 2MEAN, two peaks at 1117 and 966 cm−1 were attributed to the asymmetric and symmetric stretching vibrations of C‒O. Upon the mixing of AgNO3 with 2MEAN, these two peaks exhibited red shifts, which indicated the interactions between Ag+ and C‒O. Among different concentrations, AgNO3/ 2MEAN at silver salt concentration of 2 mol/L exhibited the largest red shifts of the C‒O vibrations, from 1117 to 1108 cm−1 and from 966 to 961 cm−1, which meant that the strongest coordinative interactions between Ag+ and C‒O occurred. The strongest hydrogen bond and coordination interactions resulted in the structural compactness AgNO3/2MEAN, which could clearly explain why the solubility and diffusivity of C2H4 obtained minimum values at the silver salt concentration of 2 mol/L. Therefore, at low concentrations, the dissolved AgNO3 is probably “structure-making” in 2MEAN and leads to structural compactness of AgNO3/2MEAN, thus resulting in that the C2H6 solubility and permeability decreased and a salting-out effect was observed [41,42]. However, at high concentration, AgNO3 may act as
4. Conclusions Novel PIL-FTMs were fabricated for the separation of C2H4/C2H6 for 85
Journal of Membrane Science 557 (2018) 76–86
H. Dou et al.
the first time. The PILs not only exhibited good dissolving capacity of carrier but also could stabilize and activate carrier due to their special solvent properties and useful protic acidity. The complexation ability of silver ions with C2H4 could be significantly enhanced by the presence of functional groups such as ether group and hydroxyl group of the cations of PILs, thus increasing the C2H4/C2H6 selectivity. It was also found that the strongest hydrogen bond and coordination interactions between AgNO3 and 2MEAN were achieved at the silver salt concentration of 2 mol/L, endowing PIL-FTMs with the highest C2H4/C2H6 selectivity. In addition, the increase of transmembrane pressure decreased the permeability of C2H4 and selectivity of C2H4/C2H6, while the increase of temperature increased the permeability of C2H4 but decreased the selectivity of C2H4/C2H4. Together with low cost of PILs and AgNO3, the excellent separation performance and good long-term stability would make this new PIL-FTMs move a step towards industrialization.
[19]
[20]
[21]
[22]
[23]
[24]
[25]
Acknowledgments [26] [27]
We are grateful for financial support from the National Key R&D Program of China (Nos. 2016YFC0400406 and 2017B0602702).
[28]
References
[29]
[1] B. Li, Y. Zhang, R. Krishna, K. Yao, Y. Han, Z. Wu, D. Ma, Z. Shi, T. Pham, B. Space, J. Liu, P.K. Thallapally, J. Liu, M. Chrzanowski, S. Ma, Introduction of pi-complexation into porous aromatic framework for highly selective adsorption of ethylene over ethane, J. Am. Chem. Soc. 136 (2014) 8654–8660. [2] C. Ma, X. Wang, X. Wang, B. Yuan, Y. Wu, Z. Li, Novel glucose-based adsorbents (Glc-As) with preferential adsorption of ethane over ethylene and high capacity, Chem. Eng. Sci. 172 (2017) 612–621. [3] A. Luna-Triguero, J.M. Vicent-Luna, P. Gomez-Alvarez, S. Calero, Olefin/paraffin separation in open metal site Cu-BTC metal-organic framework, J. Phys. Chem. C. 121 (2017) 3126–3132. [4] X. Ma, S. Williams, X. Wei, J. Kniep, Y.S. Lin, Propylene/propane mixture separation characteristics and stability of carbon molecular sieve membranes, Ind. Eng. Chem. Res. 54 (2015) 9824–9831. [5] B. Jiang, H. Dou, B. Wang, Y. Sun, Z. Huang, H. Bi, L. Zhang, H. Yang, Silver-based deep eutectic solvents as separation media: supported liquid membranes for facilitated olefin transport, ACS Sustain. Chem. Eng. 5 (2017) 6873–6882. [6] B. Jiang, H. Dou, L. Zhang, B. Wang, Y. Sun, H. Yang, Z. Huang, H. Bi, Novel supported liquid membranes based on deep eutectic solvents for olefin-paraffin separation via facilitated transport, J. Membr. Sci. 536 (2017) 123–132. [7] Y. Sun, H. Bi, H. Dou, H. Yang, Z. Huang, B. Wang, R. Deng, L. Zhang, A novel copper(I)-based supported ionic liquid membrane with high permeability for ethylene/ethane separation, Ind. Eng. Chem. Res. 56 (2017) 741–749. [8] L.C. Tome, D. Mecerreyes, C.S.R. Freire, L.P.N. Rebelo, I.M. Marrucho, Polymeric ionic liquid membranes containing IL-Ag+ for ethylene/ethane separation via olefin-facilitated transport, J. Mater. Chem. A 2 (2014) 5631–5639. [9] C. Zhang, Y. Dai, J.R. Johnson, O. Karvan, W.J. Koros, High performance ZIF-8/ 6FDA-DAM mixed matrix membrane for propylene/propane separations, J. Membr. Sci. 389 (2012) 34–42. [10] A. Cadiau, K. Adil, P.M. Bhatt, Y. Belmabkhout, M. Eddaoudi, A metal-organic framework-based splitter for separating propylene from propane, Science 353 (2016) 137–140. [11] J. Liu, Y. Xiao, T.-S. Chung, Flexible thermally treated 3D PIM-CD molecular sieve membranes exceeding the upper bound line for propylene/propane separation, J. Mater. Chem. A 5 (2017) 4583–4595. [12] T.C. Merkel, R. Blanc, I. Ciobanu, B. Firat, A. Suwarlim, J. Zeid, Silver salt facilitated transport membranes for olefin/paraffin separations: carrier instability and a novel regeneration method, J. Membr. Sci. 447 (2013) 177–189. [13] L.M. Galán. Sánchez, G.W. Meindersma, A.B. Haan, Potential of silver-based roomtemperature ionic liquids for ethylene/ethane separation, Ind. Eng. Chem. Res. 48 (2009) 10650–10656. [14] Y. Li, S. Wang, G. He, H. Wu, F. Pan, Z. Jiang, Facilitated transport of small molecules and ions for energy-efficient membranes, Chem. Soc. Rev. 44 (2015) 103–118. [15] J.H. Kim, B.R. Min, J. Won, Y.S. Kang, Complexation mechanism of olefin with silver ions dissolved in a polymer matrix and its effect on facilitated olefin transport, Chem.-Eur. J. 8 (2002) 650–654. [16] R. Zarca, A. Ortiz, D. Gorri, I. Ortiz, A practical approach to fixed-site-carrier facilitated transport modeling for the separation of propylene/propane mixtures through silver-containing polymeric membranes, Sep. Purif. Technol. 180 (2017) 82–89. [17] M. Fallanza, A. Ortiz, D. Gorri, I. Ortiz, Experimental study of the separation of propane/propylene mixtures by supported ionic liquid membranes containing Ag+–RTILs as carrier, Sep. Purif. Technol. 97 (2012) 83–89. [18] S. Kasahara, E. Kamio, R. Minami, H. Matsuyama, A facilitated transport ion-gel
[30]
[31]
[32]
[33] [34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42] [43]
[44] [45]
[46]
86
membrane for propylene/propane separation using silver ion as a carrier, J. Membr. Sci. 431 (2013) 121–130. Y. Sung Park, Y. Soo Kang, S. Wook Kang, Cost-effective facilitated olefin transport membranes consisting of polymer/AgCF3SO3/Al(NO3)3 with long-term stability, J. Membr. Sci. 495 (2015) 61–64. B. Jose, J.H. Ryu, B.G. Lee, H. Lee, Y.S. Kang, H.S. Kim, Effect of phthalates on the stability and performance of AgBF4-PVP membranes for olefin/paraffin separation, Chem. Commun. (2001) 2046–2047. S.W. Kang, J.H. Kim, J. Won, Y.S. Kang, Suppression of silver ion reduction by Al (NO3)3 complex and its application to highly stabilized olefin transport membranes, J. Membr. Sci. 445 (2013) 156–159. J.H. Kim, B.R. Min, H.S. Kim, J. Won, Y.S. Kang, Facilitated transport of ethylene across polymer membranes containing silver salt: effect of HBF4 on the photoreduction of silver ions, J. Membr. Sci. 212 (2003) 283–288. S. Jeong, S.W. Kang, Effect of Ag2O nanoparticles on long-term stable polymer/ AgBF4/Al(NO3)3 complex membranes for olefin/paraffin separation, Chem. Eng. J. 327 (2017) 500–504. S.W. Kang, K. Char, J.H. Kim, C.K. Kim, Y.S. Kang, Control of ionic interactions in silver salt-polymer complexes with ionic liquids: implications for facilitated olefin transport, Chem. Mater. 18 (2006) 1789–1794. S.W. Kang, J.H. Kim, K. Char, Y.S. Kang, Chemical activation of AgNO3 to form olefin complexes induced by strong coordinative interactions with phthalate oxygens of poly(ethylene phthalate), Ind. Eng. Chem. Res. 45 (2006) 4011–4014. A.S. Amarasekara, Acidic ionic liquids, Chem. Rev. 116 (2016) 6133–6183. J. Reid, F. Agapito, C.E.S. Bernardes, F. Martins, A.J. Walker, S. Shimizu, M.E.M. da Piedade, Structure-property relationships in protic ionic liquids: a thermochemical study, Phys. Chem. Chem. Phys. 19 (2017) 19928–19936. L. Zhang, L. He, C.-B. Hong, S. Qin, G.-H. Tao, Brønsted acidity of bio-protic ionic liquids: the acidic scale of [AA]X amino acid ionic liquids, Green. Chem. 17 (2015) 5154–5163. C. Chiappe, A. Mezzetta, C.S. Pomelli, G. Iaquaniello, A. Gentile, B. Masciocchi, Development of cost-effective biodiesel from microalgae using protic ionic liquids, Green. Chem. 18 (2016) 4982–4989. B. Jiang, H. Yang, L. Zhang, R. Zhang, Y. Sun, Y. Huang, Efficient oxidative desulfurization of diesel fuel using amide-based ionic liquids, Chem. Eng. J. 283 (2016) 89–96. R. Deng, Y. Sun, H. Bi, H. Dou, H. Yang, B. Wang, W. Tao, B. Jiang, Deep eutectic solvents as tuning media dissolving Cu+ used in facilitated transport supported liquid membrane for ethylene/ethane separation, Energy Fuel 31 (2017) 11146–11155. X. Zhang, Z. Tu, H. Li, K. Huang, X. Hu, Y. Wu, D.R. MacFarlane, Selective separation of H2S and CO2 from CH4 by supported ionic liquid membranes, J. Membr. Sci. 543 (2017) 282–287. G. Zarca, I. Ortiz, A. Urtiaga, Copper(I)-containing supported ionic liquid membranes for carbon monoxide/nitrogen separation, J. Membr. Sci. 438 (2013) 38–45. C. Chiappe, M. Malvaldi, B. Melai, S. Fantini, U. Bardi, S. Caporali, An unusual common ion effect promotes dissolution of metal salts in room-temperature ionic liquids: a strategy to obtain ionic liquids having organic–inorganic mixed cations, Green Chem. 12 (2010) 77–80. S. Kasahara, E. Kamio, A.R. Shaikh, T. Matsuki, H. Matsuyama, Effect of the aminogroup densities of functionalized ionic liquids on the facilitated transport properties for CO 2 separation, J. Membr. Sci. 503 (2016) 148–157. K. Huang, X.-M. Zhang, Y.-X. Li, Y.-T. Wu, X.-B. Hu, Facilitated separation of CO2 and SO2 through supported liquid membranes using carboxylate-based ionic liquids, J. Membr. Sci. 471 (2014) 227–236. A. Ilyas, N. Muhammad, M.A. Gilani, K. Ayub, I.F.J. Vankelecom, A.L. Khan, Supported protic ionic liquid membrane based on 3-(trimethoxysilyl)propan-1aminium acetate for the highly selective separation of CO2, J. Membr. Sci. 543 (2017) 301–309. X.-M. Zhang, Z.-H. Tu, H. Li, L. Li, Y.-T. Wu, X.-B. Hu, Supported protic-ionic-liquid membranes with facilitated transport mechanism for the selective separation of CO2, J. Membr. Sci. 527 (2017) 60–67. E. Santos, J. Albo, A. Irabien, Acetate based Supported Ionic Liquid Membranes (SILMs) for CO2 separation: influence of the temperature, J. Membr. Sci. 452 (2014) 277–283. L.C. Tomé, D. Mecerreyes, C.S.R. Freire, L.P.N. Rebelo, I.M. Marrucho, Polymeric ionic liquid membranes containing IL–Ag+ for ethylene/ethane separation via olefin-facilitated transport, J. Mater. Chem. A 2 (2014) 5631–5639. T. Mendez-Morales, J. Carrete, J.R. Rodriguez, O. Cabeza, L.J. Gallego, O. Russina, L.M. Varela, Nanostructure of mixtures of protic ionic liquids and lithium salts: effect of alkyl chain length, Phys. Chem. Chem. Phys. 17 (2015) 5298–5307. R. Hayes, S.A. Bernard, S. Imberti, G.G. Warr, R. Atkin, Solvation of inorganic nitrate salts in protic ionic liquids, J. Phys. Chem. C. 118 (2014) 21215–21225. T. Murphy, S.K. Callear, G.G. Warr, R. Atkin, Dissolved chloride markedly changes the nanostructure of the protic ionic liquids propylammonium and ethanolammonium nitrate, Phys. Chem. Chem. Phys. 18 (2016) 17169–17182. R. Faiz, K. Li, Olefin/paraffin separation using membrane based facilitated transport/chemical absorption techniques, Chem. Eng. Sci. 73 (2012) 261–284. J.H. Kim, B.R. Min, K.B. Lee, J. Won, Y.S. Kang, Coordination structure of various ligands in crosslinked PVA to silver ions for facilitated olefin transport, Chem. Commun. (2002) 2732–2733. S.W. Kang, J.H. Kim, J. Won, K. Char, Y.S. Kang, Effect of amino acids in polymer/ silver salt complex membranes on facilitated olefin transport, J. Membr. Sci. 248 (2005) 201–206.