Puxu Liu ,Yong Wang ,Yang Chen ,Xiaoqing Wang ,Jiangfeng Yang ,Libo Li,2,*,Jinping Li

1 College of Chemistry and Chemical Engineering,Shanxi Key Laboratory of Gas Energy Efficient and Clean Utilization,Taiyuan University of Technology,Taiyuan 030024,China

2 Key Laboratory of Coal Science and Technology,Taiyuan University of Technology,Taiyuan 030024,China

Keywords:Titanium metal-organic framework Adsorption Separation Ethylene purification Strong binding affinity Molecular simulation

ABSTRACT Direct separation of high purity ethylene(C2H4)from an ethane(C2H6)/ethylene mixture is a critical and challenging task owing to the very similar molecular size and physical properties of the two components.While some studies have attempted this separation,there is a lack of excellent porous materials with strong binding affinity for C2H6-selective adsorption via an energy-efficient adsorptive separation process.Herein,we report a titanium metal-organic framework with strong binding affinity and excellent stability for the highly efficient removal of C2H6 from C2H6/C2H4 mixtures.Single component adsorption isotherms demonstrated a larger amount of adsorbed ethane(1.16 mmol·g-1 under 1 kPa)and high C2H6/C2H4 selectivity (2.7) for equimolar C2H6/C2H4 mixtures,especially in the low-pressure range,which is further confirmed by the results of grand canonical Monte Carlo simulations for C2H6 adsorption in this framework.The experimental breakthrough curves showed that C2H4 with a high purity was collected directly from both 1:9 and 1:15 C2H6/C2H4 (volume ratio) mixtures at 298 K and 100 kPa.Moreover,the unchanged adsorption and separation performance after cycling experiments confirmed the promising applicability of this material in future.

1.Introduction

Ethylene (C2H4) is the main raw material for manufacturing a variety of chemicals,such as polyethylene and ethylene glycol[1,2],and it is produced by methods such as catalytic cracking and dehydrogenation of ethane[3-5].To meet the criterion of poly grade C2H4(＞99.95%),the inevitable byproduct ethane (C2H6)obtained during ethylene separation needs to be removed [6-8].However,the similar physical properties and molecular size of these two components (C2H6:(3.81×10-10) × (4.08×10-10) ×(4.82×10-10) m3;C2H4(3.28×10-10) × (4.18×10-10) ×(4.84×10-10) m3) make their separation quite a challenging task[9-17].The efficient separation of these two compounds is regarded as one of the seven chemical separations that will have significant global impact [18].

Nowadays,the separation of C2H6and C2H4is still performed mostly by employing conventional cryogenic distillation,operated at a low temperature and high pressure over 100 trays,which accounts for nearly half of the energy consumption in the entire chemical separation process [18].Thus,the development of a portable and energy-efficient method to separate these homologues is urgently needed [7,19].Therefore,adsorption separation using porous materials has been developed and has become a focus by many researchers due to advantages such as facile operation requirements and low energy consumption [19-29].

For C2H6/C2H4separation,the conventional adsorbents were almost C2H4-selective adsorbents due to the π-complexation between the open metal sites (OMS) with the unsaturated olefin.High purity C2H4could only be obtained by regenerating the C2H4-selective adsorbents to collect the desire product after multiple adsorption and desorption cycles.It thus would be more portable to fabricate C2H6-selective adsorbents straightforwardly in views of a little impurity C2H6in the mixture during the real applications.Because energy consumption and operation cost will be reduced substantially[30-35].Generally,the mechanisms for preferential C2H6adsorption include the gate-opening effect,the hydrogen-binding affinity,and the difference in the van der Waals force between the framework and gas molecules.The first reported C2H6adsorption material was a type of zeolitic imidazolate framework (ZIF),synthesized by Gasconet al.in 2010 [36].They confirmed that the microporous material ZIF-7 could efficiently and selectively adsorb C2H6over C2H4through the gate-opening effect at a low pressure,but it exhibited a poor separation performance at ambient conditions.Thereafter,in 2015,Chenet al.[8] created a microporous metal-organic framework(MOF),MAF-49,that exhibited 1D zigzag channels decorated with a high density of polar nitrogen atoms with the diameter of about 0.3×10-9m,it preferentially captured C2H6viamultiple hydrogen bonding and electrostatic interactions resulting from its special structure.However,the limited space in this microporous material resulted in a low C2H6adsorption capacity.Subsequently,Liet al.[37,38]conducted adsorption and dynamic breakthrough experiments on Ni(bdc)(ted)0.5and PCN-250;they revealed that the unique pore environment and the van der Waals force might account for the large amount of adsorbed C2H6.In 2018,Chenet al.[39] reported improved C2H6/C2H4separation performance by using a microporous MOF Cu(Qc)2;they demonstrated that the optimized pore structure in Cu(Qc)2could maximize the number of weak hostguest interactions within the MOF,so that the C2H6molecules were well-accommodated in the suitable pores,resulting in improved C2H6adsorption.Further,recently,our group reported the benchmark material Fe2(O2)(dobdc) that showed distinctive high C2H6/C2H4selectivity and excellent separation performance[40].However,the heat of adsorption and the moisture sensitivity of this special material were the highest of all reported materials for C2H6/C2H4separation,which limited its application.Although the use of C2H6-selective adsorbents is favorable for this separation because of low energy consumption,the lack of sufficient stability and strong binding affinity toward C2H6has impeded the practical application of these materials.To the best of our knowledge,few examples exist of MOFs with good stability and the ability to efficiently remove C2H6from C2H6/C2H4mixtures.Given the high coordination number and robustness of the scaffold assembled with the highly charged metal node with the organic linkers [41-45],we hypothesized that a titanium-based MOF material with a nonpolar surface might have great potential for paraffin-selective adsorption with high stability.Therefore,we constructed and reported the adsorption separation of C2H6and C2H4on a titanium MOF ZSTU-1 with high stability [46].The material showed strong binding affinity toward C2H6and adsorbed a higher amount of C2H6than C2H4,especially at low pressures.In addition,the C2H6distribution simulated by the grand canonical Monte Carlo method(GCMC) also confirmed the stronger binding affinity toward C2H6of this scaffold.Moreover,the dynamic breakthrough experimental results suggested that ZSTU-1 could efficiently remove C2H6from C2H6/C2H4mixtures (1:9 and 1:15,volume ratio) to afford highpurity C2H4(＞99.95%).Moreover,the maintained structure and unchanged sorption performance of ZSTU-1 after cycling measurements and various examinations for stability indicated that these materials have great potential for C2H6/C2H4separation on an industrial scale.

2.Experimental

2.1.Synthesis of ZSTU-1

ZSTU-1 was synthesized according to the approach reported previously,with minor modifications.All the chemical reagents were used as received without further purification.4,4′,4′′-nitrilo tribenzoic acid (H3TCA) (0.377 g,1 mmol) was added toN,Ndimethylformamide(DMF,5 ml)and fully dissolved by continuous stirring at room temperature in a 25 ml Teflon vessel.Then,0.31 ml(1 mmol) titanium(IV) isopropoxide [Ti(OiPr)4] was introduced into the mixture,and a bright yellow solution was immediately formed.Thereafter,the Teflon vessel was sealed and transferred into an autoclave maintained at 478 K for 24 h.After the mixture was naturally cooling down to the room temperature,the solution was centrifuged and washed thrice with DMF and thrice with methanol,and the desired yellow sample was obtained after drying overnight in a vacuum oven at 358 K.For gas sorption measurements,the guest-free sample was obtained after activation of the yellow powder at 393 K until no further mass loss was observed.For structural stability investigations,0.02 g sample was soaked in 4 ml vial filled with corresponding solvent to determine the solvent and acid/base stability of ZSTU-1.For moisture stability,the 0.02 g sample was exposed to water vapor in a sealed vial containing different concentration NaCl solution.Ethane (C2H6,99.99%),ethylene (C2H4,99.99%),helium (He,99.999%),nitrogen (N2,99.99%),and C2H6/C2H4mixed gases (1:1,1:9,and 1:15,volume ratio) were purchased from Beijing Special Gas Co.Ltd,China.

2.2.Characterization

Powder X-ray diffraction (PXRD) patterns were recorded on a Bruker D8 ADVANCE X-ray diffractometer by employing Cu-Kα radiation (λ=1.5418 × 10-10m) at 30 kV over the 2θ range of 5°-40°at a scanning rate of 1(°)·min-1.Further,scanning electron microscopy (SEM) was performed to observe the morphology of the sample using a Hitachi SU8010 scanning electron microscope.Thermogravimetric analysis(TGA)was conducted under N2conditions with a NETZSCH STA 449F5 thermal analyzer at a heating rate of 5 K·min-1from room temperature to 1073 K.

2.3.Gas sorption measurements

An Intelligent Gravimetric Analyzer (IGA 001,Hiden,UK) was employed to precisely record the C2H4and C2H6adsorption isotherms using gravimetric techniques at equilibrium conditions.Before each experiment,the solvent-exchanged sample was degassed at high vacuum (＞10-3kPa) until no further mass loss was observed.The N2sorption isotherms of ZSTU-1 at 77 K were obtained using a Micromeritics ASAP 2020 analyzer after the sample was fully activated.The samples were activated at 393 K over 4 h to remove the guest molecules prior to each measurement.

2.4.Dynamic breakthrough experiments

The dynamic breakthrough experiments were performed using the homemade set-up composed of two columns for the separation analysis.One of the columns loaded with the activated MOF powder was used for adsorption,while the other was employed to stabilize the flow rate of the gas mixture.The flow rate at the inlet and outlet of the adsorption column was controlled simultaneously by a pressure gauge and a mass flow meter.During the experiment,the effluent gas from the adsorption column was monitored by using a gas chromatography system (GC-2014C Shimadzu)equipped with a thermal conductivity detector.For the cycling breakthrough experiments,the adsorption column was purgedin situwith helium at a flow rate of 30 ml·min-1for 30 min.All gases used were of high purity (99.99%).

2.5.Simulation methods

To evaluate the mechanism of C2H6-slective adsorption in the framework,GCMC simulations were performed to determine the adsorption sites of C2H6.These simulations were performed using the sorption module with the COMPASS force field in Materials Studio.The Lennard-Jones interactions were calculated using a cutoff radius of 12.0 × 10-10m,while the Coulombic and van der Waals interactions were handled with the Ewald summation method.The MOF structure was treated as rigid with the building atoms frozen at their original crystallographic positions during the simulations.For each state point,2.0 × 107steps were used to guarantee adsorption equilibrium.

3.Results and Discussion

3.1.Characterization of ZSTU-1

As shown in the Fig.1(a),six Ti atoms were linked by the organic linkers to form the Ti6(μ3-O)6(COO)6cluster,and these Ti6clusters were further connected to form a 1D infinite nanorod through the μ2-OH groups;next,viathe connection of six of the Ti6clusters,these nano-rods were extended by the TCA linkers and joined to form the 3D porous structure with 1D hexagonal channels along thecaxis,as shown in Fig.1(b)(depicted as the yellow hexagonal prism).This special structure was similar to that of the typical MOF-74 and MIL-177-HT [25,47] as demonstrated in Fig.S1 (in Supplementary Material).The recorded PXRD patterns of ZSTU-1 was consistent with those of the simulated one,confirming the successful synthesis and good crystallinity of the sample(Fig.S2).The morphology of this MOF was hexagonal,as seen in the SEM images in Fig.S3.TG curves showed two-step mass loss during heating:the first one might be caused by the removal of guest solvents and the-OH groups,while the second one is attributed to the disintegration of the structure.Moreover,the remained PXRD patterns at 598 K indicating the excellent thermal stability of ZSTU-1 (Fig.S4).The N2sorption isotherm at 77 K was used to determine the permanent porosity of ZSTU-1.As shown in Fig.S5,ZSTU-1 exhibited a microporous structure with the pore size distribution and BET surface area of 0.7 × 10-10m and 563 m2·g-1(Langmuir surface area:1559 m2·g-1),respectively.

3.2.C2H6/C2H4 adsorption isotherms and selectivity of ZSTU-1

The adsorption of C2H6and C2H4on the prepared ZSTU-1 with 1D non-polar surface channels was investigated.The structure exhibited the inversed adsorption performance.And the amount of C2H6adsorbed (26.3 cm3·g-1) was larger than C2H4(15.8 cm3-·g-1)especially at the low pressure 1 kPa(Fig.2(a))at ambient conditions;this performance is superior to that of most C2H6-selective adsorbents,such as ZIF-7(0.259 cm3·g-1),PCN-250(5.10 cm3·g-1),JNU-2 (3.67 cm3·g-1),and Ni(bdc)(ted)0.5(1.94 cm3·g-1) (Fig.2b).Further,this exceptional adsorption performance encouraged us to investigate the strong binding affinity toward C2H6in detail.We calculated the IAST selectivity for the C2H6/C2H4mixture based on the dual-sites Langmuir-Freundlich fitting parameters of the adsorption isotherms (Fig.S6):the IAST selectivity in the case of the 1:1 (volume ratio) C2H6/C2H4mixture was 2.3 at 298 K and 100 kPa.Then,the isosteric heat of adsorption for both components was determined using the Clapeyron-Clausius equation,on the basis of the virial fitting parameters for C2H6and C2H4adsorption isotherms recorded at 298 K,288 K,and 273 K (Fig.S7).The obtained isosteric heat of adsorption for C2H6was 43 kJ·mol-1larger than that for C2H4(32 kJ·mol-1) under the zero-coverage region,which was consistent with the adsorption isotherms(Fig.2 (c)-(e)).Therefore,the selectivity and adsorbed amount of C2H6at 1 kPa of the C2H6-selective adsorbents were compared,and the performance of ZSTU-1 was comparable to that of the benchmark material MAF-49 (Fig.2(f)).

3.3.Dynamic breakthrough experiments and GCMC simulations

To further evaluate the separation performance of ZSTU-1,dynamic breakthrough experiments were carried out online using a homemade set-up (Fig.S8) for actual separation using three C2H6/C2H4mixture compositions (1:1,1:9,and 1:15,volume ratio).As shown in Fig.3(a)-(c),C2H4was first eluted from the adsorption column and reached the equilibrium concentration,while C2H6was preferentially adsorbed and remained the adsorption column until adsorption saturation was reached.The breakthrough time intervals for the three C2H6/C2H4mixtures were 3 min,5 min,and 10 min,respectively;during these time intervals,C2H4with a high purity could be collected directly at the outlet of the column without much energy consumption.In addition,at room temperature,the diffusion rate for C2H6was higher than that for C2H4(Fig.S9),and this excellent separation performance in dynamic conditions was consistent with the single component adsorption isotherms for two components.We also calculated the dynamic adsorption capacity for C2H6and C2H4at mixed gas condition,and their capture amounts were 34.86 cm3·g-1and 24 cm3·g-1for equimolar C2H6/C2H4mixture respectively (Fig.S10).Thus,the promising applicability of ZSTU-1 in practical separation was confirmed.

To further elucidate the adsorbent’s strong binding affinity toward C2H6,the GCMC simulations were conducted to get insight into the interactions and identify the potential adsorption sites at both 1 kPa and 100 kPa at 298 K.Fig.4(a)-(b)shows the adsorbed C2H6density distributions after the C2H6molecules were introduced into the structure and adsorption equilibrium was achieved.The area for C2H6adsorption at both pressures was located at the corner part adjacent to uncoordinated carbonyl oxygen of Ti-O nano-rod in the hexagon oval channels,and the adsorbed amount increased with the pressure,as indicated by the higher concentration of red dots at 100 kPa.Due to the non-planar ligands,the three benzene rings rotated and formed a one-dimensional zigzag channel along thecaxis with the non-polar surface to construct the ZSTU-1 scaffold.Furthermore,some terminal ligands like water and the -OH group grafted at the Ti-O nano-rod avoided the OMS,so the π-complexation between the unsaturated olefin and OMS would be suppressed,thus creating a suitable pore environment for C2H6to be adsorbedviamultiple hydrogen bonds and C-H···π interactions.Moreover,the measured shortest distance of the hydrogen bond(1.89×10-10m)indicated the strong affinity between the framework and the molecules,which was in line with the calculated heat of C2H6adsorption.

Fig.1.A schematic illustration of ZSTU-1 synthesized for C2H6/C2H4 adsorptive separation along c (a) and b (b) direction,and hydrogen atoms were omitted for clarity.

Fig.2.C2H6(red)and C2H4(blue)adsorption(solid)and desorption(open)isotherms at 298 K and 100 kPa(a)and a comparison among other C2H6-selective adsorbents for C2H6 adsorption amount under low pressure at 298 K(b).IAST selectivity of ZSTU-1 for C2H6/C2H4(1/1,volume ratio)at 298 K(c).Isosteric heat for C2H6 and C2H4 of ZSTU-1(d)-(e).A comparison combined with the IAST selectivity (1/1,volume ratio) and C2H6 adsorption capacity under 298 K and 1 kPa (f).

Fig.3.Experimental breakthrough curves for 1/1 (a),1/9 (b) and 1/15 (c) C2H6/C2H4 (volume ratio) mixture at a gas flow rate of 2.0 ml·min-1 at 298 K and 100 kPa.

Fig.4.The C2H6 adsorption distribution among the framework simulated by the GCMC method at 1 kPa(a),(b)and 100 kPa(c),(d)and 298 K respectively.(C2H6,red dots and green).

Fig.5.Cycling adsorption isotherms for C2H6 and multiple breakthrough experiments for C2H6/C2H4 mixture (a)-(b).And associated PXRD patterns and N2 sorption isotherms at 77 K showed the good structural stability of ZSTU-1 (c)-(d).

3.4.Structural stability investigation

The strong binding affinity toward C2H6and the modest C2H6/C2H4separation performance of ZSTU-1 encouraged us to investigate the stability of ZSTU-1 in detail.Thus,the cycling adsorption of C2H6and multiple breakthrough experiments for the 1:15 (volume ratio) C2H6/C2H4mixture were conducted at ambient conditions (Fig.5(a)-(b)).There were no noticeable changes in the adsorption isotherms and breakthrough curves,which indicates the good cyclability of ZSTU-1.Furthermore,the intact structure was also confirmed from the PXRD patterns after adsorption and breakthrough experiments.Moreover,the moisture and solvent stability of ZSTU-1 were further explored by exposing or soaking the sample in a 4 ml sealed vial for 2 days;the PXRD pattern and almost unchanged N2sorption isotherms (Fig.5(c)-(d)) confirmed the moisture stability and solvent stability of this material.

4.Conclusions

In summary,because of the lack of efficient C2H6-selective adsorbents with a strong binding affinity toward C2H6for direct C2H4purification,we constructed a titanium metal-organic framework,involving ZSTU-1 for the C2H6/C2H4separation.The synthesized ZSTU-1 exhibited strong selective C2H6adsorption over C2H4in single component adsorption isotherms,especially in the low-pressure range,and this excellent performance was also validated by the results of GCMC simulations and calculation of isosteric heat of adsorption for the two components.Moreover,the conducted dynamic breakthrough experiments demonstrated that ZSTU-1 could preferentially remove trace C2H6from the C2H6/C2H4mixtures(1:9 and 1:15,volume ratio)to give C2H4with a high purity in dynamic conditions.Furthermore,the distinctive good stability of this material was also confirmedviacycling adsorption and breakthrough experiments.This work not only presented a novel MOF for C2H6-selective adsorption but also paved the way for the construction of stable adsorbents with a high binding affinity toward target species.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We gratefully acknowledge the financial support from the National Natural Science Foundation of China (21922810,21908153,21908155)and program of Innovative Talents of Higher Education Institutions of Shanxi.We are grateful for the supported by Cultivate Scientific Research Excellence Programs of Higher Education Institutions in Shanxi (CSREP).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.07.027.