Detailed Investigation of Separation Performance ... - Semantic Scholar

membranes Article

Detailed Investigation of Separation Performance of a MMM for Removal of Higher Hydrocarbons under Varying Operating Conditions Heike Mushardt † , Marcus Müller † , Sergey Shishatskiy, Jan Wind and Torsten Brinkmann * Institute of Polymer Research, Helmholtz-Zentrum Geesthacht, Geesthacht 21502, Germany; [email protected] (H.M.); [email protected] (M.M.); [email protected] (S.S.); [email protected] (J.W.) * Correspondence: [email protected]; Tel.: +49-4152-87-2400; Fax: +49-4152-87-2499 † These authors contributed equally to this work. Academic Editor: Darrell Patterson Received: 11 December 2015; Accepted: 16 February 2016; Published: 25 February 2016

Abstract: Mixed-matrix membranes (MMMs) are promising candidates to improve the competitiveness of membrane technology against energy-intensive conventional technologies. In this work, MMM composed of poly(octylmethylsiloxane) (POMS) and activated carbon (AC) were investigated with respect to separation of higher hydrocarbons (C3+ ) from permanent gas streams. Membranes were prepared as thin film composite membranes on a technical scale and characterized via scanning electron microscopy (SEM) and permeation measurements with binary mixtures of n-C4 H10 /CH4 under varying operating conditions (feed and permeate pressure, temperature, feed gas composition) to study the influence on separation performance. SEM showed good contact and absence of defects. Lower permeances but higher selectivities were found for MMM compared to pure POMS membrane. Best results were obtained at high average fugacity and activity of n-C4 H10 with the highest selectivity estimated to be 36.4 at n-C4 H10 permeance of 12 mN 3 /(m2 ¨ h¨ bar). Results were complemented by permeation of a multi-component mixture resembling a natural gas application, demonstrating the superior performance of MMM. Keywords: mixed matrix membrane; activated carbon; gas permeation; n-butane/methane separation; operating conditions; multi-component mixture

1. Introduction Today, natural gas is the fastest growing energy source and with a worldwide production of 3.5 billion mN 3 per year making it a highly promising market for separation processes [1]. Around 95% of all industrial separations are covered by the refinery and processing of fossil fuels like crude oil or natural gas or the treatment of associated effluent gas [2]. In this field, the separation of higher hydrocarbons (C3+ ) from permanent gas streams (e.g., CH4 , N2 , H2 ) is of great importance. In processing of natural or associated gas, the separation is usually required to result in a consumer-grade product, prevent formation of hydrocarbon condensates during transport that might be harmful for pipeline systems and make use of valuable by-products [3,4]. Separation is mostly performed by conventional methods such as pressure swing adsorption, cryogenic distillation, expansion or absorption, which are usually accompanied by complex process design, high energy demand and high operating costs [4]. Much research has been performed, but despite their highly energy and cost intensive character, only few technologies are competitive today.

Membranes 2016, 6, 16; doi:10.3390/membranes6010016

www.mdpi.com/journal/membranes

Membranes 2016, 6, 16

2 of 13

A promising alternative are membrane-based separations due to their low energy demand, small plant sizes, environmental friendliness and ease of operation [5]. State-of-the art materials for the separation of condensable hydrocarbons (C3+ ) are siloxane based polymers such as poly(dimethylsiloxane) (PDMS) or poly(octylmethylsiloxane) (POMS) with a highly rubbery character and a solubility controlled permeation in favour of higher hydrocarbons. They have achieved commercial significance in some areas of application today but their performance requires further improvement to ensure competitiveness. Both a high transmembrane flux and a good selectivity are required for a reasonable plant size and energy demand [6]. Further challenges are the harsh conditions regarding pressure, temperature and chemicals that the membrane material has to withstand in natural gas processing operations [7]. The choice of membrane material is crucial for separation. Whereas almost all commercial gas separations are currently based on dense polymeric membranes as they are easy to fabricate as thin film composite membranes on porous support structures, as integral asymmetric hollow fibers or flat sheet membranes while offering good separation behaviour and mechanical strength, conventional polymer chemistry seems to be exploited, and no further revolutionary improvement is expected by just modifying polymeric structure. Porous inorganic membranes made of ceramic, zeolite or carbon can provide superior separation properties, especially when molecular sieving or a selective surface flow can be achieved [8]. They are also beneficial for high temperature applications and harsh chemical conditions [5]. However, despite their superior selectivity and other advantages, widespread industrial application of inorganic membranes is hampered by serious drawbacks regarding their defect-free production, low mechanical strength, brittleness and difficulties with transfer into membrane modules, as well as their costs being up to 10 times higher compared to polymeric membranes [1]. The concept of hybrid mixed-matrix materials (MMMs) based on inorganic filler particles dispersed in a polymeric matrix has attracted much interest and research in recent years. Advantages of both materials are exploited resulting in improved separation performance, mechanical, thermal or chemical stability or specific properties such as conductivity [9]. Achieving this simple but challenging concept faces many difficulties, for example, the careful selection of appropriate materials, their merging into ideal, defect-free structures and transfer of pure material properties into the assembled mixed-matrix structure. Most important for formation of a successful MMM is the interface between continuous polymeric matrix and dispersed filler which directly affects the separation performance [10]. Ideal interface morphology is characterised by a good contact between both materials and shall not show non-selective voids, leaky interface, sieve-in-a-cage morphology, rigidified polymeric layers or blocked pores. Further difficulties are the uniform dispersion of particles within polymeric matrix without agglomeration or sedimentation, realisation of high loadings in thin separation layers and reproducibility of production [11]. So far, no industrial realisation of MMM concepts for gas separation has been reported in literature to our knowledge. Research has been focused mostly on zeolites, carbon molecular sieves (CMS), silica or metal organic frameworks (MOF) as filler materials aiming at molecular sieving and diffusion selectivity meaningful for glassy polymers as matrix for fillers [12–14]. While often investigations of MMM performance for O2 /N2 or CO2 /CH4 separations can be found in literature, only few reports deal with the preferential separation of large, condensable hydrocarbons such as n-C4 H10 from permanent CH4 gas streams.Jia et al. [15] studied the permeation of various gases through silicalite filled PDMS membranes. They observed a change of selectivity due to the sieving properties of silicalite which facilitated the permeation of small molecules and impeded the usually high permeation of large, condensable components. In further studies, mostly silica has been applied as filler material in both rubbery and glassy matrix materials for this purpose [16–18]. Khanbabaei et al. [16] reported a decreased sorption capacity for PDMS filled with fumed silica and different optimums in filler loading with respect to permeability and selectivity. The work of Nunes et al. [17] describes MMMs prepared via sol-gel approach with PDMS and tetraethyl orthosilicate (TEOS). While n-C4 H10 /CH4 selectivity could be increased 10-fold for 14% TEOS due to reduced swelling, scattered values for permeability highlighted the difficulties


3 of 13

in production of uniform layers. The authors further emphasized the importance of mixed gas experiments to account for coupling or swelling effects. Merkel et al. [18] reported an increase of Membranes 2016, 6, 16 3 of 13 n-C 4 H10 permeability for addition of silica to glassy PTMSP. No improvement of selectivity was found due to the large free volume that allows the accommodation of filler particles without remarkable effect reduced tendency for plasticization or swelling. An improvement of both permeability and on polymeric structure but with a reduced tendency for plasticization or swelling. An improvement n-C4H10/CH4 mixed gas selectivity (21 for 45 wt% silica at 25 °C) has been reported by He et al. for of both permeability and n-C4 H10 /CH4 mixed gas selectivity (21 for 45 wt% silica at 25 ˝ C) has been silica filled PMP resulting from disrupted chain packing and effective blocking of CH4 diffusion by reported by He et al. for silica filled PMP resulting from disrupted chain packing and effective blocking condensed n-C4H10 in free volume [19]. of CH4 diffusion by condensed n-C4 H10 in free volume [19]. The promising application of MMMs based on POMS and activated carbon (AC) has been The promising application of MMMs based on POMS and activated carbon (AC) has been described in our previous work[20,21]. Best performance was found for a filler content of 20 wt% described in our previous work [20,21]. Best performance was found for a filler content of 20 wt% showing an improved selectivity n-C4H10/CH4 up to 20% compared to pure POMS. Based on these showing an improved selectivity n-C4 H10 /CH4 up to 20% compared to pure POMS. Based on these results, the preparation technique has been further optimized and successfully transferred into a results, the preparation technique has been further optimized and successfully transferred into a large large scale production of up to 100 m2 membrane area. The performance of MMMs is determined by scale production of up to 100 m2 membrane area. The performance of MMMs is determined by an an interplay of structural parameters, for example the type and shape of filler particles, the filler interplay of structural parameters, for example the type and shape of filler particles, the filler content, content, the properties of the polymeric matrix, as well as process conditions [12]. Its good the properties of the polymeric matrix, as well as process conditions [12]. Its good performance shall performance shall further be highlighted in this work regarding the influence of a multitude of further be highlighted in this work regarding the influence of a multitude of operating conditions to operating conditions to identify the best suited application range. identify the best suited application range. 2. Results 2. Resultsand andDiscussion Discussion 2.1. SEM 2.1. SEM Analysis Analysis Figure 11shows showsSEM SEMimages imagesofof cross-section and surface a TFC MMM 20 AC. wt%AAC. A Figure cross-section and surface of aofTFC MMM withwith 20 wt% good good contact between polymeric and inorganic material a quite uniform dispersion of particles contact between polymeric and inorganic material and and a quite uniform dispersion of particles is is revealed. Theaddition additionofofparticles particlesresults resultsininaaslight slightincrease increaseof ofseparation separation layer layer thickness thickness giving giving revealed. The values of of 3.9 3.9 µm µm for for MMM MMM compared compared to to 3.4 3.4 µm µm of of pure pure TFC TFC POMS POMS membranes. membranes. Reported Reported values values are are values averages of of at at least least two two different different samples. samples. No No visible visible defect defect formation formation indicating indicating improper improper material material averages selection was was observed. observed. selection (a)

(b)

2 µm

(c)

5 µm

100 µm

Figure 1. 1. SEM SEMimages imagesofof pure POMS (a) and MMM (b) cross-sections of separation layers (c) Figure pure POMS (a) and MMM (b) cross-sections of separation layers and (c) and surface surface MMM composed POMS and activated 20 wt% activated carbon 50= 1.5 µm). of MMMofcomposed of POMSofand 20 wt% carbon (d = 1.5(dµm). 50

2.2. Single Gas Permeation 2.2. Single Gas Permeation Single gas measurements depict an idealized membrane performance as no coupling or Single gas measurements depict an idealized membrane performance as no coupling or competitive effects between gas mixture components occur. Nonetheless, such investigation is competitive effects between gas mixture components occur. Nonetheless, such investigation is necessary to understand fundamental differences in transport through the membrane between gas necessary to understand fundamental differences in transport through the membrane between gas components. Results for permeation of single gases are presented in Figure 2 in terms of n-C4H10 and components. Results for permeation of single gases are presented in Figure 2 in terms of n-C4 H10 and CH4permeances as function of average fugacity at 20 or 70 °C for POMS (grey) and MMM filled with CH4 permeances as function of average fugacity at 20 or 70 ˝ C for POMS (grey) and MMM filled with 20 wt% AC (black). The average fugacity is estimated as average between fugacities applied on feed 20 wt% AC (black). The average fugacity is estimated as average between fugacities applied on feed and permeate side of membrane and used as reference for the corresponding concentration of gas and permeate side of membrane and used as reference for the corresponding concentration of gas components within the membrane [22]. components within the membrane [22]. The permeance of n-C4H10 (Figure 2a) shows an increase with feed pressure or average fugacity The permeance of n-C4 H10 (Figure 2a) shows an increase with feed pressure or average fugacity in POMS and MMM with the exponential trend curve indicated by the solid lines. Higher values are in POMS and MMM with the exponential trend curve indicated by the solid lines. Higher values are found for POMS resulting from slightly lower thickness of separation layer. The tortuosity created by addition of AC is expected to slow down diffusive transport. Permeation is controlled by solubility with both solution of n-C4H10 into polymeric matrix and adsorption on AC being dependent on pressure. Dissolution of condensable n-C4H10 into polymeric matrix causes a loosening of polymeric chains known as swelling and thus higher diffusive flux [23]. This is especially pronounced at lower temperatures as permeation is highly controlled by solubility with lower


4 of 13

found for POMS resulting from slightly lower thickness of separation layer. The tortuosity created by addition of AC is expected to slow down diffusive transport. Permeation is controlled by solubility Membranes 2016, 6, 16 4 of 13 with both solution of n-C4 H10 into polymeric matrix and adsorption on AC being dependent on pressure. Dissolution of condensable n-C4 H10 into polymeric matrix causes a loosening of polymeric temperatures being beneficial for solubility and adsorption. No pressure dependency was found for chains known as swelling and thus higher diffusive flux [23]. This is especially pronounced at lower CH4 permeance in POMS or MMM (see Figure 2b). Constant values are found for each temperature temperatures as permeation is highly controlled by solubility with lower temperatures being beneficial between 20 and 70 °C with values ranging from 0.2 to 0.4 mN3/(m2·h·bar) for POMS and 0.12 to 0.23 for 3solubility and adsorption. No pressure dependency was found for CH4 permeance in POMS mN /(m2·h·bar) for MMM. Permeation of non-condensable components is dominated by diffusion or MMM (see Figure 2b). Constant values are found for each temperature between 20 and 70 ˝ C and facilitated by higher temperatures3 due2 to enhanced mobility of gas molecules. The ideal with values ranging from 0.2 to 0.4 mN /(m ¨ h¨ bar) for POMS and 0.12 to 0.23 mN 3 /(m2 ¨ h¨ bar) selectivity was calculated from n-C4H10permeances extrapolated to zero feed pressure. No for MMM. Permeation of non-condensable components is dominated by diffusion and facilitated by significant difference was found between POMS and MMM. Ideal selectivity was estimated to be 15.6 higher temperatures due to enhanced mobility of gas molecules. The ideal selectivity was calculated (POMS) and 14.6 (MMM) at 20 °C and 6.6 (POMS) and 6.5 (MMM) at 70 °C with the decrease from n-C4 H10 permeances extrapolated to zero feed pressure. No significant difference was found resulting from lowern-C4H10 permeation. between POMS and MMM. Ideal selectivity was estimated to be 15.6 (POMS) and 14.6 (MMM) at 20 ˝ C ˝ C with Figure 2. Permeance of (a) n-Cat 4H 10 and (b) CH differentresulting average fugacities in POMS and and 6.6 (POMS) and 6.5 (MMM) 70 the4 atdecrease from lower n-C4 H(grey) 10 permeation. MMM filled with 20 wt% AC (black) determined with pure gases at different temperatures (lines

represent exponential trend curves). Figure 2. Permeance of (a) n-C4 H10 and (b) CH4 at different average fugacities in POMS (grey) and MMM filled with 20 wt% AC (black) determined with pure gases at different temperatures (lines 2.3.Binary Gas Mixture Permeation and Separation represent exponential trend curves).

As stated above, single gas measurements are not sufficient to correctly evaluate separation performance to neglect of theand coupling effects present in real gas mixtures. For this reason, the 2.3. Binary Gasdue Mixture Permeation Separation performance of MMM has been investigated with binary gas mixtures n-C4H10/CH4 under a variety As stated above, single gas measurements are not sufficient to correctly evaluate separation of operating conditions [24]. performance due to neglect of the coupling effects present in real gas mixtures. For this reason, the performance of MMM has been investigated with binary gas mixtures n-C4 H10 /CH4 under a variety 2.3.1. Influence of Feed Pressure of operating conditions [24]. A significant influence of feed pressure is expected as it is directly affecting the driving force for 2.3.1. Influence of Feed Pressure permeation. Measurements have been performed with feed pressures ranging from 10 to 40 bar at a constant temperature of 20 °C of feed gas and permeate of 1.2–1.6 bar.the driving force for A significant influence of feed pressure is aexpected aspressure it is directly affecting Results Measurements for n-C4H10 andhave CH4been permeances as with a function of average fugacity permeation. performed feed pressures ranging fromn-C 104H to1040resulting bar at a from variation of feed pressure are presented in Figure 3 with lines representing exponential trend ˝ constant temperature of 20 C of feed gas and a permeate pressure of 1.2–1.6 bar. curves. For both membrane types as well as gas components, an increase of permeance is shown. Results for n-C 4 H10 and CH4 permeances as a function of average fugacity n-C4 H10 resulting While no great differences can between trends in lines both representing cases in POMS and MMM, the from variation of feed pressure be areobserved presented in Figure 3 with exponential trend increase is much more pronounced for n-C 4H10 compared to CH4. As described for single gas n-C4H10, curves. For both membrane types as well as gas components, an increase of permeance is shown. the increase is differences caused by can the be dissolution and swelling supported by the the While no great observed between trendsofinpolymeric both casesmatrix in POMS and MMM, enhanced adsorption on activated carbon in case of MMM. For example, the permeancen-C 4H10 increase is much more pronounced for n-C4 H10 compared to CH4 . As described for single gas shows an increase from 6.8 to 14.8 mN3/(m2·h·bar) in POMS and from 4.1 to 10.9 mN3/(m3·h·bar) in n-C 4 H10 , the increase is caused by the dissolution and swelling of polymeric matrix supported by MMM. While adsorption no pressureon dependency was found forofsingle gas CH 4 permeance, an increase can be the enhanced activated carbon in case MMM. For example, the permeance n-C4 H10 observed in case of binary mixture. The CH4 permeation is directly coupled to the concentration of dissolved n-C4H10 which not only facilitates CH4 diffusion by polymer swelling but also creates a more favorable environment for solution of CH4 [25]. The degree of swelling can be evaluated from the dependence of hydrocarbon permeation on the average fugacity of n-C4H10 and results are thus presented in Figure 3b. The CH4permeance increases from 0.3 to 0.6 mN3/(m2·h·bar) in POMS and


5 of 13

18

Permeance CH4 [mN³/(m² h bar)]

Permeance n -C4H10 [mN³/(m² h bar)]

shows an increase from 6.8 to 14.8 mN 3 /(m2 ¨ h¨ bar) in POMS and from 4.1 to 10.9 mN 3 /(m3 ¨ h¨ bar) in MMM. While no pressure dependency was found for single gas CH4 permeance, an increase can be observed in case of binary mixture. The CH4 permeation is directly coupled to the concentration of dissolved n-C4 H10 which not only facilitates CH4 diffusion by polymer swelling but also creates a more favorable environment for solution of CH4 [25]. The degree of swelling can be evaluated from Membranes 2016, 6, 16of hydrocarbon permeation on the average fugacity of n-C4 H10 and results are5thus of 13 the dependence presented in Figure 3b. The CH4 permeance increases from 0.3 to 0.6 mN 3 /(m2 ¨ h¨ bar) in POMS and 0.16 to to 0.36 0.36 m mN N33/(m2·h·bar) in in MMM. A lower degree of of swelling is expected forfor thethe MMM as as part of 0.16 /(m2 ¨ h¨ bar) MMM. A lower degree swelling is expected MMM part the n-C 4H10 is bound to adsorption sites of activated carbon and not available for dissolution into of the n-C4 H10 is bound to adsorption sites of activated carbon and not available for dissolution into polymeric matrix. matrix. In In addition, addition, the the presence presence of of particles particles impairs impairs the the mobility mobility of of polymeric polymeric chains chains and and polymeric thus reduces the ability to loosen up the structure [17,18]. thus reduces the ability to loosen up the structure [17,18].

(a)

16 14 12 10 8 6 4

POMS

2 0

MMM 0.0 0,0

0.2 0.4 0.6 0.8 0,2 0,4 0,6 0,8 Average fugacity n -C4H10 [bar]

1.0 1,0

0,8 0.8

(b) 0.6 0,6 0.4 0,4 0.2 0,2 0.0 0,0

POMS MMM

0.0 0,0

0.4 0.6 0.8 0.2 0,2 0,4 0,6 0,8 Average fugacity n -C4H10 [bar]

1.0 1,0

Figure and (b) Figure 3. 3. Influence Influence of ofaverage averagen-C n-C44H H10 10 fugacity fugacity on on permeance permeance of of (a) (a) n-C n-C44H H10 10 and (b) CH CH44 in in POMS POMS ˝C (grey) and MMM with 20 wt% AC (black) for a binary feed mixture with 5 mol% n-C H (grey) and MMM with 20 wt% AC (black) for a binary feed mixture with 5 mol% n-C44H10 10 at at 20 20 °C (lines (lines represent represent exponential exponential trend trend curves). curves).

35

18 16

(a)

14 12 10 8 6 4 2 0

POMS MMM

Selectivity n -C4H10 / CH4

ermeance n -C4H10 [mN³/(m² h bar)]

The influence influence of of feed feed pressure pressure on on permeance permeancen-C 10/CH 4 is illustrated in The n-C44H H1010and andselectivity selectivityn-C n-C4H 4H 10 /CH 4 is illustrated Figure 4b.4b. TheThe same trends areare observed forfor POMS andand MMM with increasing selectivity up up to 30 in Figure same trends observed POMS MMM with increasing selectivity to bar followed by a slight decrease or flattening. Nonetheless, a higher selectivity is found for MMM 30 bar followed by a slight decrease or flattening. Nonetheless, a higher selectivity is found for MMM over the thewhole wholeinvestigated investigated pressure range best results at This 30 bar. This improvement is over pressure range with with best results at 30 bar. improvement is attributed attributed the high affinity of n-C 4HAC. 10 toward AC. A high amount of n-C 4H 10 pores inside the pores to the hightoaffinity of n-C H toward A high amount of n-C H inside the might not 4 10 4 10 mightreduce not only reduce and thus minimize undesired co-permeation also evoke only swelling andswelling thus minimize undesired co-permeation but also evoke abut selective surfacea selective flow and4 blocking 4 diffusion formation layer. of a condensed The above slight flow and surface blocking of CH diffusion of viaCH formation of via a condensed The slightlayer. decrease decrease above by 30 excessive bar is caused byofexcessive of increased polymeric competition matrix andamong increased 30 bar is caused swelling polymericswelling matrix and gas competition among gas components as transport through swollen polymer is to a greater degree components as transport through swollen polymer is to a greater degree governed by diffusion which governed by diffusion which In favors smaller CH4 swollen, molecules. In case of a severely favors smaller CH4 molecules. case of a severely highly permeable matrix, swollen, this mighthighly even permeable matrix, this might even cause a by-passing of filler particles. Furthermore, a stronger cause a by-passing of filler particles. Furthermore, a stronger competition between n-C4 H10 and CH4 competition between 4H10 and CH4isfor adsorption sites in activated carbon is likely at high for adsorption sites in n-C activated carbon likely at high pressure. The thermodynamic selectivity of pressure. The thermodynamic selectivity of adsorbent is due oftentoreported to level high adsorbent materials is often reported to level off at highmaterials pressures co-adsorption ofoff theatlower pressures due to co-adsorption of decrease the lowerofaffinity componentcurves supporting thepressure. decrease of MMM affinity component supporting the MMM selectivity at higher selectivity curves at higher pressure.

30

(b)

25 20 15 10 5 0

POMS MMM

governed by diffusion which favors smaller CH4 molecules. In case of a severely swollen, highly permeable matrix, this might even cause a by-passing of filler particles. Furthermore, a stronger competition between n-C4H10 and CH4 for adsorption sites in activated carbon is likely at high pressure. The thermodynamic selectivity of adsorbent materials is often reported to level off at high pressures due to co-adsorption of the lower affinity component supporting the decrease of MMM Membranes of 13 13 Membranes 2016, 2016, 6, 6, 16 16 66 of selectivity curves at higher pressure.



Figure 4. Influence of feed pressure on (a) permeance of n-C4H10 and (b) n-C4H10/CH4 selectivity of POMS 35 18 (grey) and MMM with 20 wt% AC (black) for a binary feed mixture with 5 mol% n-C4H10 at 20 °C. (b) (a) 16 14

2.3.2. Influence of Permeate Pressure 12

30 25

16

40

14

35

12 10 8 6 4

POMS

2 0

(a) 0.0 0,0



In order to 20 and also the pressure-ratio influencing the 10 increase the driving force for permeation separation performance, membrane-based separations are often performed with vacuum on the 8 15 permeate side [5]. Thus, the influence of permeate pressure was investigated by installing a vacuum 6 10 pump and adjusting values between 0.05 and 1.5 bar. Meanwhile, the feed pressure and temperature 4 POMS POMS 5 were kept constant at 30 bar and 20 °C. Results of permeances for 2 MMM are presented in Figure 5 in terms MMM n-C4H10 (a) and0selectivity n-C4H10/CH4 (b) as function of 0 permeate pressure with lines representing 10 20 30 40 50 0 10 20 30 40 50 exponential trend0 curves. Feed pressure [bar] Again, similar trendsFeed arepressure found[bar] for POMS and MMM. Both show decreasing permeances of n-C4H 10 and CH4 with lowering of the permeate pressure. Contrary effects are evoked by variation of Figure 4. Influence of feed pressure on (a) permeance of n-C4 H10 and (b) n-C4 H10 /CH4 selectivity of permeate the with one 20 hand, a decrease ofapermeate at constant feed pressure POMSpressure. (grey) andOn MMM wt% AC (black) for binary feedpressure mixture with 5 mol% n-C 4 H10 at ˝ allows a higher pressure ratio across the membrane. The higher this ratio, the higher the driving 20 C. force for permeation which especially favors the diffusion of small molecules such as CH4. This enhancement to be superimposed by the reduced solubility at lower pressure or associated 2.3.2. Influenceseems of Permeate Pressure lower average fugacity respective to the degree of swelling. By comparing the two membrane types, In order to increase the driving force for permeation and also the pressure-ratio influencing the effect is more pronounced for POMS as indicated by the higher slope of trend curve (0.36 for the separation performance, membrane-based separations are often performed with vacuum on the POMS and 0.29 for MMM). The selectivity shows the same decreasing trend in relation to decreasing permeate side [5]. Thus, the influence of permeate pressure was investigated by installing a vacuum permeate pressure as permeances in both POMS and MMM (see Figure 5b). Results are in the range pump and adjusting values between 0.05 and 1.5 bar. Meanwhile, the feed pressure and temperature of 21.6–27.2 for POMS and 25.1–30.3˝ for MMM. Best performance could thus be achieved at higher were kept constant at 30 bar and 20 C. Results are presented in Figure 5 in terms of permeances for permeate pressure. This is quite beneficial as additional energy- or cost intensive application of n-C4 H10 (a) and selectivity n-C4 H10 /CH4 (b) as function of permeate pressure with lines representing vacuum pumps can be avoided at least in the case of feed pressure greater than 10 bar. exponential trend curves.

MMM 0.5 1.0 1.5 0,5 1,0 1,5 Permeate pressure [bar]

2.0 2,0

30 25 20 15 10

POMS

5 0

(b) 0.0 0,0

MMM 0.5 1.0 1.5 0,5 1,0 1,5 Permeate pressure [bar]

2.0 2,0

Figure 5. Influence of selectivity 4H 1010and 4H 1010/CH Figure 5. Influence of permeate permeatepressure pressureon on(a) (a)permeance permeanceofofn-C n-C 4H and(b) (b)n-C n-C 4H /CH44 selectivity of POMS (grey) and MMM with 20 wt% AC (black) for a binary feed mixture with 5 mol% n-C at of POMS (grey) and MMM with 20 wt% AC (black) for a binary feed mixture with 5 mol% n-C44H H10 10 at ˝ C (lines represent exponential trend curves). 30 bar and 20 30 bar and 20 °C (lines represent exponential trend curves).

similar trends are found for POMS and MMM. Both show decreasing permeances of 2.3.3.Again, Influence of Temperature n-C4 H10 and CH4 with lowering of the permeate pressure. Contrary effects are evoked by variation of As stated above, the temperature has a significant influence on permeation as it impacts permeate pressure. On the one hand, a decrease of permeate pressure at constant feed pressure allows solubility in POMS, adsorption on AC and diffusion. The mixed gas permeation and separation as a a higher pressure ratio across the membrane. The higher this ratio, the higher the driving force for function of temperature is presented in Figure 6 for a constant feed pressure of 30 bar. Similar trends permeation which especially favors the diffusion of small molecules such as CH4 . This enhancement are observed for POMS and MMM with both showing solubility controlled permeation over the seems to be superimposed by the reduced solubility at lower pressure or associated lower average investigated temperature range. The permeance of n-C4H10 (Figure 6a) decreases with increasing fugacity respective to the degree of swelling. By comparing the two membrane types, the effect is more temperature reflecting the lower n-C4H10 solubility respective to the lower degree of swelling of the pronounced for POMS as indicated by the higher slope of trend curve (0.36 for POMS and 0.29 for selective layer. Values for MMM are estimated to be in the range of 12–4.7 mN3/(m2·h·bar) for MMM). The selectivity shows the same decreasing trend in relation to decreasing permeate pressure as 15–35 °C. The dissolution of gas into the polymeric matrix is a two-step mechanism composed of


7 of 13

permeances in both POMS and MMM (see Figure 5b). Results are in the range of 21.6–27.2 for POMS and 25.1–30.3 for MMM. Best performance could thus be achieved at higher permeate pressure. This is quite beneficial as additional energy- or cost intensive application of vacuum pumps can be avoided at least in the case of feed pressure greater than 10 bar. 2.3.3. Influence of Temperature

16

50

(a)

14



As stated above, the temperature has a significant influence on permeation as it impacts solubility in POMS, adsorption on AC and diffusion. The mixed gas permeation and separation as a function of temperature is presented in Figure 6 for a constant feed pressure of 30 bar. Similar trends are observed for POMS and MMM with both showing solubility controlled permeation over the investigated temperature range. The permeance of n-C4 H10 (Figure 6a) decreases with increasing temperature reflecting the lower n-C4 H10 solubility respective to the lower degree of swelling of the selective layer. Values for MMM are estimated to be in the range of 12–4.7 mN 3 /(m2 ¨ h¨ bar) for 15–35 ˝ C. The dissolution of6,gas Membranes 2016, 16 into the polymeric matrix is a two-step mechanism composed of condensation 7 of 13 and mixing. At higher temperatures, n-C4 H10 exhibits a lower activity in gas phase which in turns condensation and At higher temperatures, 4H10 exhibits a lower activity in gas phase affects its readiness tomixing. condense and thus decreases n-C solubility [25]. This is in good agreement with which in turns affects its readiness to condense and thus decreases solubility [25]. This is in good behaviour of single gas permeation (see Figure 2a). For CH4 a decrease of mixed gas permeation agreement with behaviour of single gas permeation (see Figure 2a). For CH4 a decrease of mixed gas from 0.33 to 0.24 mN 3 /(m3 ¨ h¨ bar) in MMM with increasing temperature was found which is in strong permeation from 0.33 to 0.24 mN3/(m3·h·bar) in MMM with increasing temperature was found which contrast to single gas behaviour. The enhanced diffusivity seems to be compensated by decreased is in strong contrast to single gas behaviour. The enhanced diffusivity seems to be compensated by swelling in presence In n-C addition, transport through AC is still in favour for n-C H10 . 4 H10 . of decreased swellingofinn-C presence 4H10. In addition, transport through AC is still in favour for4 The influence of temperature on permeation results a decrease of mixed gas selectivity for both n-C4H10. The influence of temperature on permeationinresults in a decrease of mixed gas selectivity studied membranes as depicted in Figure 6b. The selectivity n-C H /CH decreased from 26.8 to for both studied membranes as depicted in Figure 6b. The selectivity n-C4H104/CH4 decreased from 4 10 ˝ 26.8POMS to 16.7and for from POMS31.6 andto from 19.8 for MMM in the temperature For a 16.7 for 19.831.6 for to MMM in the temperature range ofrange 20–35of 20–35 C. For°C. a successful successful of and both MMM, POMS and MMM, lower temperatures be selected ensure application of application both POMS lower temperatures shouldshould be selected thatthat ensure a high a high solubility in the polymeric matrix as well as high amount of adsorbed n-C 4H10 in the pore solubility in the polymeric matrix as well as high amount of adsorbed n-C4 H10 in the pore system of system of activated carbon. activated carbon.

12 10 8 6 4

POMS MMM

2 0 0

10

(b)

40 30 20 10

POMS MMM

0 20 30 Temperature [°C]

40

0

10

20 30 Temperature [°C]

40

Figure 6. Influence of temperature onon (a)(a) n-C and (b) (b) n-C n-C4H Figure 6. Influence of temperature n-C 4H permeance and 10/CH 4 selectivity of of 4H 1010 permeance 4H 10 /CH 4 selectivity POMS (grey) MMM with wt% AC(black) (black)for for aa binary binary feed with 5 mol% n-Cn-C 4H104at POMS (grey) and and MMM with 20 20 wt% AC feedmixture mixture with 5 mol% H10 at bar (lines represent exponential trendcurves). curves). 30 bar30(lines represent exponential trend 2.3.4. Influence of Binary Feed Composition

2.3.4. Influence of Binary Feed Composition

Industrial separations typically involve integral changes of compositions in the separation

Industrial typically involvesubstantially integral changes of compositions the separation system, i.e.,separations compositions are changing between the feed and in permeate sides ofsystem, a i.e., compositions are changing substantially between the feed and permeate sides of membrane membrane module. Furthermore, fluctuating compositions of feed mixtures to be separateda impose module. Furthermore, fluctuating compositions of feed mixtures tocomponent be separated impose a great challenge for separation process. A low concentration of desired usually makesaitgreat more for difficult to separate as low forces are available while high concentrations can cause challenge separation process. A driving low concentration of desired component usually makes it more excessive swelling or even degradation of polymeric matrix in case of corrosive components. For difficult to separate as low driving forces are available while high concentrations can cause excessive hydrocarbon mixtures, theofhigher hydrocarbons arecase usually present only in minor For amounts. To swelling or even degradation polymeric matrix in of corrosive components. hydrocarbon evaluate membrane performance, even under these unfavourable conditions, measurements were performed with binary mixtures containing 1, 2 or 5 vol-% n-C4H10. Permeance of n-C4H10 in POMS and MMM and the n-C4H10/CH4 selectivity as function of binary mixture composition are presented in Figure 7 for 20 °C and 30 bar. The permeance shows an exponential increase with feed concentration of n-C4H10. This again can be related to the average fugacity of n-C4H10 and associated degree of swelling and adsorption.


8 of 13

16

35

(a)

14



mixtures, the higher hydrocarbons are usually present only in minor amounts. To evaluate membrane performance, even under these unfavourable conditions, measurements were performed with binary mixtures containing 1, 2 or 5 vol-% n-C4 H10 . Permeance of n-C4 H10 in POMS and MMM and the n-C4 H10 /CH4 selectivity as function of binary mixture composition are presented in Figure 7 for 20 ˝ C and 30 bar. The permeance shows an exponential increase with feed concentration of n-C4 H10 . This again can be related to the average fugacity of n-C4 H10 and associated degree of swelling and adsorption. As was to be expected, more gas can dissolve or adsorb if it is present in higher concentrations in feed mixture causing an increase of average fugacity. Comparison of slopes reveals a similar increase for n-C4 H10 in POMS (0.22) and MMM (0.23) but a lower increase for CH4 in MMM (0.14) compared to POMS (0.17). It is assumed that less CH4 can permeate due to a reduced swelling of polymer as well as a dominating n-C4 H10 phase in pore system blocking pathways for CH4 diffusion. The influence of feed composition on selectivity is given in Figure 7b. At low concentrations of 1 or 2 vol-% n-C4 H10 , both membranes Membranes 2016, 6, 16 show nearly the same selectivity but the beneficial effect of AC particles is clearly 8 of 13 demonstrated in terms of the improved selectivity at higher n-C4 H10 concentrations. Feed mixtures concentrations. more than 2 vol-% H10 are thus recommended the with more than 2Feed vol-%mixtures n-C4 H10 with are thus recommended for n-C the 4application of MMM based onfor POMS application of MMM based on POMS and AC. and AC.

12 10 8 6 4

POMS MMM

2 0

(b)

30 25 20 15 10

POMS MMM

5 0

0

2 yF,n -C4H10

4 [mol %]

6

0

2 4 yF,n -C4H10 [mol %]

6

Figure 7.7. Influence Influenceofofbinary binarymixture mixturecomposition compositiononon(a)(a) n-CH4H10permeance permeanceand and(b)(b) n-CH4H10 /CH4 Figure n-C n-C 4 10 4 10 /CH4 ˝ selectivity of of POMS POMS (grey) (grey) and and MMM MMM with with 20 20 wt% wt% AC AC (black) (black) at at 20 20 °C and 30 30 bar bar (lines (lines represent represent selectivity C and exponential trend curves). exponential trend curves).

2.3.Multi-Component 2.4. Multi-Component Gas Gas Mixture Mixture Permeation Permeation and and Separation Separation The improved MMM should be be further evidenced for the The improved binary binaryseparation separationperformance performanceofofthe the MMM should further evidenced for case of a multi-component mixture resembling a typical natural gas application. The mixture is the case of a multi-component mixture resembling a typical natural gas application. The mixture is composedofofalkanes alkanes from 4 up to n-C5H12 and CO2. The total concentration of higher composed from CH4 CH up to n-C5 H12 and CO2 . The total concentration of higher hydrocarbons hydrocarbons (C 3+) was approximately 19 vol-%. Measurements have been performed at feed (C3+ ) was approximately 19 vol-%. Measurements have been performed at feed pressures between 10 pressures 10 and temperature 40 bar and a of constant and 40 barbetween and a constant 20 ˝ C. temperature of 20 °C. Permeances of of components ˝ Cand Permeances components estimated estimated at at 20 20 °C and40 40bar barfeed feedpressure pressureare aresummarized summarizedininTable Table2.1. Table 1.2. Boiling Boiling point point TTb [26], [26], permeance permeance LL and and selectivity selectivity vs. vs. CH CH4 in in multi-component multi-component mixture mixture Table 4 b permeation experiments with MMM composed of POMS and 20 wt% AC at 30 bar feed pressure and permeation experiments with MMM composed of POMS and 20 wt% AC at 30 bar feed pressure and 20 ˝°C. 20 C.

Component CH4 CO2 CH4 CO2 Tb (°C) −161.5 −78.5 ˝ C) Tb (L ´161.5 0.41 ´78.5 (mN3/(m2·h·bar)) 1.20 3 0.41 1.20 L (mN /(m2 ¨ h¨ bar)) Selectivity vs. CH 4 2.92 Component

Selectivity vs. CH4

-

2.92

C2H6 C2 H6 −88.7 ´88.7 1.64 1.64 3.98 3.98

C3H8 n-C4H10 n-C5H12 C H n-C4 H10 n-C5 H12 −42.1 3 8 −0.5 36.0 36.0 4.14´42.112.95 ´0.560.97 4.14 12.95 60.97 10.06 31.46 148.12 10.06

31.46

148.12

The order of permeance and selectivity values follows the condensability of penetrants indicated by the boiling point (see Table 2) which increases with number of carbon atoms for hydrocarbon components. The lowest permeance is thus observed for non-condensable CH4 followed by CO2, C2H6, C3H8 and n-C4H10. By far the highest permeance arises for n-C5H12 whose boiling point of 36 °C [22] indicates the great willingness to condense. All components show an increase of permeance with increasing feed pressure or, respectively, the swelling of polymeric


9 of 13

The order of permeance and selectivity values follows the condensability of penetrants indicated by the boiling point (see Table 1) which increases with number of carbon atoms for hydrocarbon components. The lowest permeance is thus observed for non-condensable CH4 followed by CO2 , C2 H6 , C3 H8 and n-C4 H10 . By far the highest permeance arises for n-C5 H12 whose boiling point of 36 ˝ C [22] indicates the great willingness to condense. All components show an increase of permeance with increasing feed pressure or, respectively, the swelling of polymeric matrix, which is illustrated for selected components in Figure 8. Membranes 2016, 6, 16 9 of 13 Membranes 2016, 6, 16

Permeance h bar)] Permeance [m[m h bar)] N³/(m² N³/(m²

9 of 13

120 120 100 100 80 80 60 60 40 40 20 20 0 0 0 0

CH4 CH4 C3H8 C3H8 n-C4H10 n-C4H10 n-C5H12 n-C5H12

10 20 30 40 10 Feed 20 pressure30[bar] 40 Feed pressure [bar]

50 50

Figure 8. Permeances Permeances of of selected selected hydrocarbon hydrocarbon components components in in multi-component multi-component mixture separation Figure 8. mixture separation Figure 8. Permeances of selected hydrocarbon components in multi-component mixture separation ˝ with MMM composed of POMS and 20 wt% AC at 20 °C (mixture composition: 1 vol-% 5HH 12, 2 with MMM composed of POMS and 20 wt% AC at 20 C (mixture composition: 1 vol-%n-C n-C 5 12,122, with MMM composed of POMS and 20 wt% AC at 20 °C (mixture composition: 1 vol-% n-C5H vol-% n-C 4H10 C3H vol-% C2H6, 0.79 vol% CH4 and 2 vol-% CO2). 2 vol-% n-C H, 6 ,vol-% 6 vol-% C8,H10 8 , 10 vol-% C2 H6 , 0.79 vol% CH4 and 2 vol-% CO2 ). vol-% n-C4H410,10 6 vol-% C3H38, 10 vol-% C2H6, 0.79 vol% CH4 and 2 vol-% CO2).

20 20

(a) (a)

16 16

Selectivityn -C n -C H10/ CH Selectivity CH 4 H410/ 4 4

Permeance n -C H10[m[m h bar)] Permeance n -C h bar)] N³/(m² 4 H410 N³/(m²

The selectivity with respect to CH4 shows the same trend and increases with increasing The selectivity with respect respect to shows the the same same trend trend and The selectivity with to CH CH4 shows and increases increases with with increasing increasing condensability. Higher selectivities for each gas component are achieved by application of MMM condensability. Higher selectivities selectivities for for each each gas gas component component are are achieved achieved by by application condensability. Higher application of of MMM MMM instead of pure POMS with greatest improvement for n-C4H10 (13%) and n-C5H12 (42%) resulting instead of pure POMS with greatest improvement for n-C H (13%) and n-C H (42%) resulting instead of pure POMS with greatest improvement for n-C4 4H1010 (13%) and n-C55H12 12 (42%) resulting from the adsorptive capacity and reduced tendency in case of activated carbon. By comparison of from from the the adsorptive adsorptive capacity capacity and and reduced reduced tendency tendency in in case case of of activated activated carbon. carbon. By By comparison comparison of of binary and multi-component mixture separation results, a greater increase of permeation (Figure 9a) binary and multi-component multi-component mixture mixture separation separation results, results, aa greater greater increase increase of of permeation permeation (Figure (Figure 9a) 9a) binary and with pressure is observed due to the presence of n-C5H12. This enhancement results in a higher with This enhancement enhancement results results in in aa higher with pressure pressure is is observed observed due due to to the the presence presence of ofn-C n-C55H H12 12.. This higher selectivity compared to values achieved with binary mixtures of same concentration n-C4H10 in feed selectivity comparedto tovalues valuesachieved achievedwith with binary mixtures of same concentration H10 in selectivity compared binary mixtures of same concentration n-C4n-C H104in feed as illustrated in Figure 9b. No negative impacts due to coupling or competitive mixture effects were feed as illustrated in Figure 9b.negative No negative impacts to coupling or competitive mixture as illustrated in Figure 9b. No impacts due todue coupling or competitive mixture effectseffects were noticed. The enhanced separation performance of MMM was clearly demonstrated even for were noticed. The enhanced separation performance of MMM wasclearly clearlydemonstrated demonstrated even even for noticed. The enhanced separation performance of MMM was for multi-component mixtures. multi-component multi-component mixtures. mixtures.

12 12 8 8 4 4 0 0

0 0

10 10

40 40 35 35 30 30 25 25 20 20 15 15 10 10 5 5 0 0

(b) (b) Multi-Mix (2%) Multi-Mix (2%) Binary (1%) Binary (1%) Binary (2%) Binary (2%) Binary (5%) Binary (5%)

20 30 40 50 0 10 20 30 40 50 20 pressure 30 [bar] 40 50 0 10 Feed 20pressure 30 [bar] 40 50 Feed Feed pressure [bar] Feed pressure [bar] Figure 9. Comparison of (a) n-C4H10 permeance and (b) n-C4H10/CH4 selectivity of MMM in Figure 9. Comparison of (a) (a) n-C n-C4H H10 permeance permeance and (b) (b) n-C n-C4H H10/CH 4 selectivity of MMM in Figure 9. Comparison MMM in 4 (2 10 vol-% n-C4H10and 4 mixture 10 /CHseparation 4 selectivity multi-component mixtureofseparation ) and binary (1,of 2 or 5 vol-% multi-component mixture separation (2 vol-% n-C 4H10) and binary mixture separation (1, 2 or 5 vol-% multi-component n-C 4H10) at 20 °C. mixture separation (2 vol-% n-C4 H10 ) and binary mixture separation (1, 2 or 5 vol-% ˝ C. n-C n-C44H H1010))atat20 20°C.

3. Experimental Section 3. Experimental Section 3.1. Materials 3.1. Materials Membranes are composed of poly(octylmethylsiloxane) (POMS) as polymeric matrix. POMS is Membranes are composed of poly(octylmethylsiloxane) (POMS) as polymeric matrix. POMS is a member of the siloxane polymer family which includes the well-known poly(dimethylsiloxane)


10 of 13

3. Experimental Section 3.1. Materials Membranes are composed of poly(octylmethylsiloxane) (POMS) as polymeric matrix. POMS is a member of the siloxane polymer family which includes the well-known poly(dimethylsiloxane) (PDMS), the industrial state-of-art material for separation of higher hydrocarbons [17]. POMS has a highly rubbery character indicated by a glass transition temperature of ´110 ˝ C giving a solubility selective permeation. A precursor solution for membrane casting was prepared in iso-octane (Merck KGaA, purity >99.5%) as solvent with a number of siloxane based cross-linking and reinforcement agents and a platinum-based catalyst to initialise cross-linking reaction. Since POMS membranes are produced and employed on commercial basis, the exact composition of precursor solution cannot be disclosed. Microporous activated carbon (AC) was used as inorganic filler phase. It was provided by Blücher GmbH as sieving fraction with mean particle size of 1.5 µm. Further characteristic properties are summarized in Table 2. The AC was selected due to its high affinity towards C3+ hydrocarbons with the adsorption isotherm following the equation of Tóth [20]. AC particles were dried at 150 ˝ C in a vacuum oven for at least 24 h prior to use. All gases were purchased at Linde (purity of 99.5%), liquid n-C5 H12 (>99.9%) at Merck KGaA and were used as received. Table 2. Characteristic properties of activated carbon. Properties

Value

d50 µm SBET m2 /g dpore Å vpore cm3 /g Porosity % Density cm3 /g

1.5 1361 18.7 0.636 57.6 0.891

3.2. Membrane Preparation To facilitate dispersion, dried AC particles were preliminarily blended with isooctane and sonicated for 40 min using an ultrasonic bath by Elmasonic S30H. MMM coating suspensions were prepared by stepwise addition of polymer precursor solution and catalyst to isooctane saturated AC particles and thoroughly mixing with a dissolver for 10–20 min at 8000 rpm (Dispermat® , VMA Getzmann GmbH) after each step. The AC content was adjusted to 20 wt% with respect to mass of polymer as identified as optimum filler content in previous work [20]. A final 2 min sonication step was applied prior to casting to ensure homogeneous dispersion and removal of trapped air. Pure POMS thin film composite (TFC) membranes have been prepared as reference material. Membranes have been prepared as TFC membranes by suspension coating on a PDMS coated support structure of microporous polyacrylonitrile (PAN) on non-woven polyester (PE). A roll-coating machine was used for TFC membrane production. An integrated oven situated immediately after the coating module allows a thermal treatment at 100 ˝ C for 5 min to induce cross-linking and solvent removal. A final PDMS layer (0.5 wt% PDMS in iso-octane) was deposited on top of the selective mixed matrix layer to eventually seal formed surface defects. 3.3. SEM Analysis MMM morphology was analysed via scanning electron microscopy (SEM) images of membrane surfaces and cross-sectional areas. Dried samples were immersed in isopropanol, freeze-fractured in liquid nitrogen and coated with a 2 nm layer of platinum. Images were taken with SEM system LEO 1550 VP by Zeiss.


11 of 13

3.4. Pure Gas Permeation Measurements Permeation measurements were performed with pure n-C4 H10 and CH4 at feed pressures up to 1 bar in the temperature range of 20–70 ˝ C. An automated set-up operating in constant volume and variable pressure mode was used to determine permeance via time dependent change of pressure in feed and permeate vessels. The set-up is described elsewhere in more detail [27,28]. 3.5. Mixed Gas Permeation Measurements Prior to characterisation, all samples were dried in a vacuum oven at 80 ˝ C and 7 mbar for at least 15 h to ensure the complete removal of residual solvent and allow activation of filler particles. The experimental mixed gas membrane characterisation set-up is depicted elsewhere [20,27]. Membrane samples of 47 mm in diameter were inserted and sealed in the test cell and the system evacuated for at least 45 min. A preliminary prepared gas mixture was filled in a feed vessel and pressurised via a compressor while a gas circulator generates a sufficient flow to avoid concentration polarization. This is ensured by a low stage-cut of less than 1%. The feed mixture was temperature controlled with a water bath before entering the membrane test cell. After equilibrating, the pressure and flow rates of feed, permeate and retentate were measured and the gas composition on feed, permeate and retentate sides analysed with a gas chromatograph (Varian 3400) equipped with a packed metal column (Chromosorb 107, carrier gas argon or helium). Mixed gas measurements were performed with previously prepared binary mixtures of n-C4 H10 in CH4 with 1–5 vol-% n-C4 H10 in feed under varying operating conditions of 15–35 ˝ C, 10–40 bar feed pressure and 0.05–1.6 bar permeate pressure [24]. A multi-component mixture with composition resembling natural gas has been prepared with 1 vol-% n-C5 H12 , 2 vol-% n-C4 H10 , 6 vol-% C3 H8 , 10 vol-% C2 H6 , 0.79 vol% CH4 and 2 vol-% CO2 and investigated at 20 ˝ C, feed pressure 10–40 bar and permeate pressure 1.2–1.5 bar. Results are presented as average values for at least two different samples with errors estimated according to the t-distribution. Since condensable gases like n-C4 H10 or n-C5 H12 show significant gas phase non-idealities, permeances were calculated based on fugacities (fi ) to account for the real gas behaviour. The fugacity coefficients were estimated by the Soave-Redlich-Kwong (SRK) equation of state. The permeance (Li ) of a component i was calculated as the permeate flow rate (Vi ) at standard conditions (1.01325 bar, 0 ˝ C) divided by its driving force, namely the fugacity difference between feed and permeate sides, and membrane area (Am ). .

Li “

Vi ´ ¯ Am ¨ f i,F ´ f i,Pq

(1)

The selectivity given by the ratio of permeances of two components i and j quantifies the separation efficiency. It is always determined with respect to the less permeable component. αi{j “

Li Lj

(2)

4. Conclusions In this work, the performance of a MMM composed of rubbery POMS and 20 wt% AC has been evaluated with respect to separation of higher hydrocarbon from permanent gas streams such as n-C4 H10 /CH4 . A solubility selective permeation was found similar to pure POMS. The detailed study of separation performance for binary feed mixtures under varying operating conditions has revealed the superior performance of MMM as well as some guidelines regarding the selection of appropriate application. Best performance was achieved at highest average fugacity of n-C4 H10 caused by increasing feed pressure, high permeate pressure and high concentration of condensable hydrocarbon component in feed mixture. Further, low temperatures provide a high activity of n-C4 H10 . The highest


12 of 13

selectivity was estimated to be 36.4 at n-C4 H10 permeance of 12 mN 3 /(m2 ¨ h¨ bar). The superior performance of MMM was also confirmed in experiments on separation of a multi-component mixture similar to natural gas with even higher selectivity n-C4 H10 /CH4 compared to binary mixture separation. Acknowledgments: This work was performed within the framework of the research initiative “Technologies for Sustainability and Climate Protection—Chemical Processes and Material Use of CO2” by the German Federal Ministry of Education and Research (BMBF, reference number 033RC1018, “Mixed Matrix Membranes for Gas Separation”). The authors would like to thank the BMBF and its commissioned project management agency (DLR) and Project Management Jülich (PtJ) for financial support and their project partners at Technical University of Berlin, Blücher GmbH and Sterling SIHI for the good cooperation. We highly appreciate the work of Sofia Dami on SEM images (Helmholtz-Zentrum-Geesthacht) and the advice of Irina Smirnova (Hamburg University of Technology). Author Contributions: The membrane samples were prepared by Heike Mushardt and Jan Wind. Single gas permeation were performed and analysed by Heike Mushardt. The mixed gas experiments were conceived by Torsten Brinkmann and Heike Mushardt and performed and analysed by Marcus Mueller under guidance and supervision of Torsten Brinkmann, Sergey Shishatskiy and Heike Mushardt. The paper was written by Heike Mushardt. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

Baker, R.W. Membrane Technology and Applications, 2nd ed.; John Wiley: Chichester, UK; New York, NY, USA, 2004. Henley, E.J.; Seader, J.D.; Roper, D.K. Separation Process Principles, 3rd ed.; Wiley: Hoboken, NJ, USA, 2011. Thompson, S.M.; Robertson, G.; Johnson, E. Liquefied Petroleum Gas. In Ullmann's Encyclopedia of Industrial Chemistry; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2000. Baker, R.W.; Lokhandwala, K. Natural Gas Processing with Membranes: An Overview. Ind. Eng. Chem. Res. 2008, 47, 2109–2121. [CrossRef] Melin, T. Membranverfahren: Grundlagen der Modul-und Anlagenauslegung, 3rd ed.; Springer: Berlin, Germany, 2007. Baker, R.W. Future Directions of Membrane Gas Separation Technology. Ind. Eng. Chem. Res. 2002, 41, 1393–1411. [CrossRef] Ohlrogge, K., Ebert, K., Eds.; Membranen: Grundlagen, Verfahren Und Industrielle Anwendungen; Wiley-VCH: Weinheim, Germany, 2005. Rao, M.; Sircar, S. Nanoporous carbon membranes for separation of gas mixtures by selective surface flow. J. Membr. Sci. 1993, 85, 253–264. [CrossRef] Kulprathipanja, S. Zeolites in Industrial Separation and Catalysis; Wiley-VCH: Weinheim, Germany, 2010. Moore, T.T.; Koros, W.J. Non-ideal effects in organic-inorganic materials for gas separation membranes. J. Mol. Struct. 2005, 739, 87–98. [CrossRef] Cong, H.; Radosz, M.; Towler, B.; Shen, Y. Polymer-inorganic nanocomposite membranes for gas separation. Sep. Purif. Technol. 2007, 55, 281–291. [CrossRef] Chung, T.; Jiang, L.; Li, Y.; Kulprathipanja, S. Mixed matrix membranes (MMMs) comprising organic polymers with dispersed inorganic fillers for gas separation. Prog. Polym. Sci. 2007, 32, 483–507. [CrossRef] Dong, G.; Li, H.; Chen, V. Challenges and opportunities for mixed-matrix membranes for gas separation. J. Mater. Chem. A 2013, 1, 4610–4630. [CrossRef] Goh, P.S.; Ismail, A.F.; Sanip, S.M.; Ng, B.C.; Aziz, M. Recent advances of inorganic fillers in mixed matrix membrane for gas separation. Sep. Purif. Technol. 2011, 81, 243–264. [CrossRef] Jia, M.; Peinemann, K.-V.; Behling, R.-D. Molecular sieving effect of the zeolite-filled silicone rubber membranes in gas permeation. J. Membr. Sci. 1991, 57, 289–296. [CrossRef] Khanbabaei, G.; Vasheghani-Farahani, E.; Rahmatpour, A. Pure and mixed CH4 and n-C4 H10 permation in PDMS-fumed silica nanocomposite membranes. Chem. Eng. J. 2012, 191, 369–377. [CrossRef] Nunes, S.; Schultz, J.; Peinemann, K.-V. Silicone membranes with silica nanoparticles. J. Mater. Sci. Lett. 1996, 15, 1139–1141. [CrossRef]


18. 19. 20. 21.

22.

23. 24. 25. 26. 27.

28.

13 of 13

Merkel, T.C.; He, Z.; Pinnau, I.; Freeman, B.D.; Meakin, P.; Hill, A.J. Effect of Nanoparticles on Gas Sorption and Transport in Poly(1-trimethylsilyl-1-propyne). Macromolecules 2003, 36, 6844–6855. [CrossRef] He, Z.; Pinnau, I.; Morisato, A. Nanostructured poly(4-methyl-2-pentyne)/silica hybrid membranes for gas separation. Desalination 2002, 146, 11–15. [CrossRef] Mushardt, H.; Kramer, V.; Hülagü, D.; Brinkmann, T.; Kraume, M. Development of Solubility Selective Mixed Matrix Membranes for Gas Separation. Chem. Ing. Tech. 2014, 86, 83–91. [CrossRef] Löffler, V.; Kraume, M. Development of Mixed-Matrix Membranes for separation of gaseous hydrocarbons. In Proceedings of the 18th International Conference Process Engineering and Chemical Plant Design, Berlin, Germany, 19–23 September 2011; Wozny, G., Ed.; Univ.-Verl. der TU: Berlin, Germany, 2011; pp. 207–216. Ohlrogge, K.; Wind, J.; Brinkmann, T. Membranes for recovery of volatile organic compounds. In Comprehensive Membrane Science and Engineering, 1st ed.; Drioli, E., Giorno, L., Eds.; Elsevier Science: Amsterdam, The Netherlands; London, UK, 2010; pp. 213–240. Raharjo, R.D.; Freeman, B.D.; Paul, D.R.; Sarti, G.C.; Sanders, E.S. Pure and mixed gas CH4 and n-C4H10 permeability and diffusivity in poly(dimethylsiloxane). J. Membr. Sci. 2007, 306, 75–92. [CrossRef] Müller, M. Permeationsverhalten von Mixed-Matrix-Membranen zur Abtrennung von Gasförmigen, Höheren Kohlenwasserstoffen; Masterarbeit: Hamburg, Germany, 2015. Raharjo, R.; Freeman, B.D.; Sanders, E. Pure and mixed gas CH4 and n-C4H10 sorption and dilation in poly(dimethylsiloxane). J. Membr. Sci. 2007, 292, 45–61. [CrossRef] Reid, R.C.; Prausnitz, J.M.; Sherwood, T.K. The Properties of Gases and Liquids, 3rd ed.; McGraw-Hill: New York, NY, USA, 1977. Car, A.; Stropnik, C.; Yave, W.; Peinemann, K.-V. Pebax® /polyethylene glycol blend thin film composite membranes for CO2 separation: Performance with mixed gases. Sep. Purif. Technol. 2008, 62, 110–117. [CrossRef] Brinkmann, T.; Pohlmann, J.; Withalm, U.; Wind, J.; Wolff, T. Theoretical and Experimental Investigations of Flat Sheet Membrane Module Types for High Capacity Gas Separation Applications. Chem. Ing. Tech. 2013, 85, 1210–1220. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons by Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).