# Fileset

[High Hexane Sorption Capacity of Loosely Crosslinked PDMS Rubbers at Low Temperatures Macromolecular and Physicochemical Elucidation for VOC Recovery.pdf](https://mdr.nims.go.jp/filesets/6819f287-e1ed-48e9-a11b-246ae1ce399e/download)

## Creator

[M. Inoue](https://orcid.org/0000-0003-3098-5448), [E. Ismail](https://orcid.org/0000-0003-1031-6562), [S. Samitsu](https://orcid.org/0000-0002-4139-1656), H. Kanoh, [I. Ichinose](https://orcid.org/0000-0002-2236-0942)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[High Hexane Sorption Capacity of Loosely Crosslinked PDMS Rubbers at Low Temperatures: Macromolecular and Physicochemical Elucidation for VOC Recovery](https://mdr.nims.go.jp/datasets/a01299bd-0b66-4f57-afd8-a30bb3e97ba1)

## Fulltext

Microsoft Word - Sept. PuriHigh Hexane Sorption Capacity of Loosely Crosslinked PDMS Rubbers at Low Temperatures: Macromolecular and Physicochemical Elucidation for VOC Recovery M. Inouea, E. Ismaila, S. Samitsub, H. Kanohc, I. Ichinosea* a Research Center for Functional Materials, National Institute for Materials Science, 1−1 Namiki, Tsukuba 305−0044, Japan b Research and Services Division of Materials Data and Integrated System, National Institute for Materials Science, 1−2−1 Sengen, Tsukuba 305−0047, Japan c Graduate School of Science, Chiba University, 1−33 Yayoi-cho, Inage, Chiba 263−8522, Japan KEYWORDS: soft PDMS, elastic modulus, hexane gas sorption, swelling, low temperature  Corresponding Author Prof. Dr. Izumi Ichinose – Research Center for Functional Materials, National Institute for Materials Science, 1−1 Namiki, Tsukuba 305−0044, Japan; orcid.org/0000-0002-2236-0942; Email: ICHINOSE.Izumi@nims.go.jp   ABSTRACT: A series of soft polydimethylsiloxane (PDMS) rubbers was obtained by diluting reaction mixtures of a PDMS macromer and a crosslinker with an organic solvent. The degree of swelling of the softest rubber was 600 wt% when immersed in pure hexane at room temperature. From the Flory–Huggins equations, the free energy change for hexane sorption per 1 g rubber was calculated to be −7.4 J g−1 and the entropy contribution of the polymer chains was 1.6 J g−1 larger than that of the conventional PDMS rubber. The loosely crosslinked PDMS rubber could sorb 776 mg g−1 of hexane at 298 K and 606 mg g−1 at 195 K. The latter value was 4.4 times larger than that of the conventional PDMS rubber. At a low pressure of 23 Pa, the PDMS rubber sorbed a 3500-fold greater amount of hexane at 195 K, as compared with the value at 298 K. This high sorption amount is because the expansion energy of such soft PDMS rubber is small. The sorption heat of the loosely crosslinked PDMS rubber was 30.3 kJ mol−1 when calculated from the Clausius–Clapeyron equation. This value is substantially smaller than that of commercial activated carbon (75 kJ mol−1). Unlike activated carbon, PDMS rubber is insensitive to water and selectively sorbs VOCs even under a saturated humidity. Because of the large swelling potential and small sorption/desorption heat, the loosely crosslinked PDMS rubber will be an indispensable sorbent for the recovery of flammable gasses such as alkane and iso-alkane VOCs in the chemical and oil and gas industries. 1. Introduction 1 An energy-efficient technology for the recovery of volatile organic compounds (VOCs) [1,2] is strongly 2 demanded to address the problems of air pollution and global climate change [3,4]. Recovery has been an 3 urgent issue, especially for the chemical, printing, and semiconductor industries, as well as for the 4 development of oil, gas, and other fossil fuels. The present major recovery methods are cryogenic 5 condensation, pressure-swing adsorption, liquid-phase absorption, membrane separation, and their 6 combinations [5–8]. Membranes have been used for the separation of VOCs with a high partial pressure. 7 For example, polydimethylsiloxane (PDMS) membranes have been applied for the recovery of VOCs 8 dissolved in water [9]. Strong sorbents are often required for removing trace concentrations of VOCs. 9 Simpsons et al. designed a Friedel–Crafts-modified polystyrene with improved sorption ability 10 comparable to that of activated carbon (AC). Their sorbent was insensitive to humidity and easy to 11 regenerate [10]. 12 Adsorbents such as AC and silica gel efficiently adsorb VOCs; however, the desorption of the VOCs 13 (regeneration of adsorbents) requires a large amount of energy. In general, AC is regenerated at 14 temperatures greater than 300 ℃ under vacuum conditions. By contrast, regeneration of polymeric 15 sorbents, such as crosslinked polystyrene and acrylic resin, carried out in vacuum at temperatures less than 16 100 ℃ in order to prevent the deterioration. These sorbents are also critical for removing VOCs with a 17 low vapor pressure [11]. Recently, porous organic crystals have attracted attention because of their 18 excellent size-selective sorption properties. For example, Yang et al. reported that a hydrogen-bonded 19 organic framework composed of 3,3′,6,6′-tetracyano-9,9′-bicarbazole showed high C2H4/C2H6 selectivity 20 [12]. Such strict molecular size/shape separations are expected to reduce the cost of chemical industrial 21 processes substantially, as Sholl et al. have advocated [13]. Interaction between adsorbates and adsorbents 22 such as carbon nanotubes is commonly discussed on the basis of the Lennard–Jones potential. However, 23 in the case of porous polymeric sorbents, sorption behaviors are analyzed by the Brunauer–Emmett–Teller 24 (BET) method or the Kelvin’s equation [14,15]. Soft, nonporous polymeric materials also sorb VOCs by 25 swelling, and this behavior is characterized using the χ interaction parameter of the Flory–Huggins theory 26 [16]. However, the sorption energy of polymeric sorbents at room temperature is weak because of the 27 flexibility of the polymer chains. In the case of polymeric sorbents, the sorption of VOCs generally follows 28 Henry’s law; that is, the sorption amount increases with increasing VOC vapor pressure. Sanders et al. 29 explained that gas solubility in separation membranes in general depends weakly on the membranes’ free 30 volume [17]. This behavior must be true for hard membranes. However, we found that soft PDMS rubbers 31 sorb substantial amounts of gasses and discontinuously swell with increasing vapor pressure, and the gas 32 solubility then increases with the expansion of their volumes even at low temperatures. The last finding 33 has not been reported elsewhere for conventional PDMS rubbers. 34 PDMS is commonly studied because of its chemical and thermal stabilities as well as its high 35 hydrophobicity resulting from its low surface tension. There are many reports on the separation 36 performance of organic solvents using PDMS membranes, in which porosity and free volume of the 37 membranes have been studied in detail [18-22]. The highly flexible polymer chains show high diffusion 38 coefficients of various gasses [23,24]. The membranes have often been used for CO2/N2 separation, as 39 reported by Liu and coworkers [25]. However, no detailed studies of the sorption behaviors have been 40 conducted through physicochemical measurements such as evaluation of sorption heats and their 41 temperature dependency. This information gap stems from the network structure of PDMS chains 42 drastically changing depending on the sorption amount, and the sorption amount will increase if the 43 network structure is loose. 44 In the present study, we evaluated the relationship of the elasticity of PDMS rubbers and the sorption 45 ability of hexane from the viewpoints of macromolecular characterization (polymer science) and 46 thermodynamics of the sorption process. PDMS rubbers with different mechanical properties were 47 prepared by diluting PDMS macromers in a good solvent before crosslinking; the polymer chains were 48 fixed as they expanded. We then found that the swelling and sorption properties of soft PDMS rubbers 49 could be improved as much as 400–600% over the swelling and sorption properties of hard ones. This 50 improvement is attributable to a memory effect of the three-dimensional (3D) configuration of the PDMS 51 chains. Many papers on molecular imprinting have been reported for highly crosslinked polymeric 52 materials [26,27]. However, the entanglement-based memory effect in loosely crosslinked PDMS rubbers 53 is, to our best knowledge, the first report in literature. 54  55 2. Experimental 56 2.1. Preparation of loosely crosslinked PDMS rubbers 57 PDMS rubbers were prepared using a SYLGARD™ 184 kit. Its solution A and solution B are 58 corresponding to PDMS macromer and crosslinker, respectively. The average molecular weights of the 59 PDMS macromer and crosslinker were 57,000 and 3700, as estimated from viscosity measurements 60 (Figure S1). The synthesis scheme of the PDMS rubbers is shown in Figure 1. This crosslinking reaction 61 is based on hydrosilylation between a vinyl group in the macromer and Si-H group in the crosslinker by 62 means of platinum catalyst. The reaction yields were almost 100 % and polymer elusion was not observed. 63 The crosslinking density is about 0.25 % for all samples, in terms of the ratio of vinyl-modified siloxane 64 units in total siloxane units. A series of PDMS rubbers with different elastic moduli were obtained from 65 1 g PDMS macromers diluted with 1, 2, 3, 4, or 5 g of hexane containing 0.1 g of crosslinker, followed 66 by curing at 30 ℃. The curing completes in one day for PDMS rubber without hexane dilution and in four 67 days for PDMS rubber diluted with 5 g of hexane. Decane and dodecane were also used as a solvent 68 instead of hexane. All solvents were special-grade products of FUJIFILM Wako Pure Chemical. Curing 69 conditions for diluted reaction mixtures with decane and dodecane were 2 days at 100 ℃. Hexane and 70 decane in solidified reaction mixtures were removed by air/vacuum drying. PDMS rubbers containing 71 dodecane were characterized in their as-prepared state. As shown in Figure 1, the polymer chains were 72 strongly entangled when prepared without a solvent. However, overlap among polymer chains decreased 73 when the chains were crosslinked under the diluted condition with a solvent. The former yielded a hard 74 gel, whereas the latter resulted in a soft gel. The names of the samples are denoted with the initial letter 75 of the solvent and the amount. For example, PDMS_h2 is the rubber obtained from 2 g of hexane, 1 g of 76 macromer, and 0.1 g of crosslinker; PDMS_de1.5 and PDMS_do5 are the rubbers prepared with 1.5 g of 77 decane and 5 g of dodecane, respectively, using 1 g of macromer and 0.1 g of crosslinker; and PDMS_0 78 is the rubber prepared without any solvent. 79  80 2.2. Analysis of thermal and mechanical properties 81 Viscosity measurements of the polymer solutions were conducted with an AR-G2 (TA Instruments). A 82 tensile tester (Autograph AGS-X, Shimadzu) was used for measuring the elastic moduli of PDMS rubbers. 83 A 2 × 2 × 8 mm3 sample was pulled at 2.5 mm min−1, and its elastic modulus was calculated when the 84 strain was 3–8%. The elastic modulus in the swollen state was obtained from the samples with a fixed 85 amount of decane sorbed. The experimental error was less than 0.5%. The Poisson’s ratio was calculated 86 from the horizontal and vertical deformation observed during a tensile test using a video camera. 87 Dynamic mechanical analysis (DMA) was conducted using an RSA-G2 (TA Instruments). The storage 88 modulus was obtained in the temperature range from −103 to 25 ℃ (170–298 K) at a frequency of 1 Hz. 89 The experimental error was less than 1%. Differential scanning calorimetry (DSC) measurements were 90 carried out using a Q2000 (TA Instruments) for dodecane sorbed in PDMS rubbers; samples were 91 analyzed in the temperature range from −90 to 20 ℃ at a heating rate of +0.5 ℃ min−1. 92  93  94 Figure 1. Preparation of PDMS rubbers and a schematic of the PDMS dispersions with different concentrations. The sample names are shown in the yellow box. 2.3. Sorption experiments 95 The hexane sorption isotherms of PDMS rubbers were obtained for granular samples using a 96 BELSORP-max (MicrotracBEL). The granular samples were prepared by cryogenic grinding 97 (Pulverisette 14, Fritsch) with 80–500 μm sieve rings. The samples were then heated at 50 ℃ for 24 h 98 under vacuum before sorption experiments at 298, 263, and 195 K. Equilibrium criteria in repeated 99 sorption/desorption cycles at 298 and 263 K were set as a pressure deviation of less than 0.3% for 1000 s. 100 At 195 K, the criterion was set as a pressure deviation of less than 0.15% for 4500 s. The error of measured 101 pressures is less than 0.5 %. 102  103 3. Results and discussion 104 3.1. Mechanical properties of PDMS rubbers 105 The mechanical properties of PDMS rubbers vary substantially with the amount of hexane used in the 106 crosslinking process, that is, the swelling of macromer in crosslinking process largely reduces the elastic 107 modulus of the resultant rubber. Table 1 shows the elastic moduli of dried PDMS rubbers and their degrees 108 of swelling in hexane. The first sample (PDMS_0) was prepared without using hexane, and the others 109 were prepared using a one- to five-fold weight of hexane relative to the weight of PDMS macromer during 110 the crosslinking process. The elastic modulus of PDMS_0 was 2.0 MPa, and this rubber sorbed 1.0 g g−1 111 of hexane. The elastic modulus of the PDMS_h5 rubber, which had five times more hexane than macromer, 112 decreased to 0.04 MPa, and the rubber sorbed 5.9 g g−1 of hexane. The degree of swelling increased in 113 proportion to the amount of hexane used in the crosslinking process. The elastic moduli of these samples 114 exponentially decreased in the series from PDMS_0 to PDMS_h5. In this paper, the rubber with an elastic 115 modulus lower than 0.5 MPa is addressed as loosely crosslinked PDMS rubber. 116  117  118 Figure 2 shows the crosslinking structures of the PDMS_0 and PDMS_h5 in reaction mixtures. From 119 the molecular weight of the PDMS macromer (57,000), the contour length of the PDMS chain is calculated 120 to be 126 nm. The average distances between crosslinkers are 4.1 nm for PDMS_0 and 8.1 nm for 121 Sample Weight ratio Hexane / resin (g g−1) Degree of overlapping 𝐶 𝐶⁄  Dry density (g cm−1) †Elastic modulus (MPa) Degree of swelling in hexane (g g−1) Macromer Crosslinker Hexane PDMS_0 1 0.1 0 0 15.6 1.05 2.0 1.0 PDMS_h1 1 0.1 1 0.91 9.6 1.00 0.28 2.2 PDMS_h2 1 0.1 2 1.81 7.0 1.00 0.13 3.5 PDMS_h3 1 0.1 3 2.73 5.5 1.01 0.08 4.2 PDMS_h4 1 0.1 4 3.64 4.5 1.01 0.04 5.3 PDMS_h5 1 0.1 5 4.55 3.8 1.01 0.04 5.9 †The values were obtained for samples in the dried state. Table 1. Composition of reaction mixture, degree of overlapping concentration, and material properties of the obtained PDMS rubbers.  PDMS_h5, as determined from the concentration of the crosslinker. Using the results of Arrighi et al. [28] 122 and Higgins et al. [29], we have estimated PDMS macromer gyration radii (𝑟 ) values of 7.3 nm for 123 PDMS_0 and 9.1 nm for PDMS_h5. The former value was obtained from the literature results of small-124 angle neutron scattering (SANS) experiments, and the latter is based on the swelling effect by a good 125 solvent (refer to the Supporting Information for details). When a five-fold greater amount of hexane 126 relative to the amount of PDMS macromer was added, the volume of the gyration sphere became twice 127 that of PDMS_0. For PDMS_0, the concentration of a certain polymer chain is high at the central part of 128 this sphere, and entanglements with other polymer chains increase at the boundary of this sphere. By 129 contrast, the polymer chains of PDMS_h5 spread out because of the penetration of hexane molecules, and 130 their entanglements are relatively reduced, which results in a very loose crosslinked structure. 131 We focused on the degree of overlapping (𝐶 𝐶⁄ ) to evaluate the extent of entanglements of the PDMS 132 chains diluted in a good solvent. Here, 𝐶  and 𝐶  are the mass concentration of the polymer and the 133 critical concentration, respectively. 𝐶  is defined by the following equation [30,31], and the unit is kg 134 m−3. 135 𝐶 =𝑀 𝑁⁄43 𝜋𝑟 136 Here, 𝑀 is molecular weight (kg mol−1), 𝑁  is Avogadro’s number (6.02×1023 mol−1), and 𝑟  is the 137 gyration radius (m). At the concentration of 𝐶 , one polymer molecule is present in a gyration sphere, and 138 overlapping is negligible when 𝐶 𝐶⁄ < 1, whereas overlapping becomes substantial when 𝐶 𝐶⁄ > 1. 139 Table 1 shows the degrees of overlapping calculated from the critical concentration, where 𝑟  is 140 expected to change in proportion to the concentration of added hexane [29]. The value for PDMS_0 is 141 Figure 2. Presumed crosslinking structures of PDMS_0 (left) and PDMS_h5 with solvent molecules (right). 15.6, which means that more than 15 polymer molecules are present within a gyration sphere. The degree 142 of overlapping for PDMS_h5 was 3.8, i.e., one-fourth of the value for PDMS_0. Entanglements still 143 occurred even when hexane was added five-fold relative to the amount of PDMS macromer. The 144 difference in the degrees of overlapping substantially changes both the dry-state elastic moduli and the 145 degrees of swelling. 146 Figure 3a shows the stress–strain curves of PDMS_hX obtained from tensile tests. PDMS_0, which was 147 crosslinked without hexane, exhibited a rupture stress of 3.0 MPa (inset of Figure 3a). In sharp contrast, 148 the rupture stress was 0.05 MPa for PDMS_h5, i.e., 1/60 of the value for PDMS_0, even at the same 149 crosslinking density. A further increase in dilution increased the elongation rate and reduced the rupture 150 stress. That is, the entanglement of PDMS chains is reduced. The rupture stress was roughly similar from 151 PDMS_h2 to PDMS_h5, indicating the formation of easily breakable defects within the rubbers. However, 152 the breaking elongation gradually increased to 370%. Figure 3b shows pictures of PDMS_0 and 153 Figure 3. (a) Stress–strain curves for PDMS_h1, PDMS_h2, PDMS_h3, PDMS_h4, and PDMS_h5. The inset is the stress–strain curves for PDMS_0 and PDMS_h1. (b) Pictures of PDMS_0 and PDMS_h5 immediately before breaking and illustrations of their crosslinked polymer chains. PDMS_h5 immediately before breaking. The latter’s elongation rate was slightly less than three times that 154 of the former. However, these samples returned to their original shape when the tensile force was removed, 155 which means that both PDMS_0 and PDMS_h5 exhibited entropic elasticity. 156 We calculated the elastic moduli (Young’s moduli) of these PDMS rubbers in the dried state (Figure 157 S2). Interestingly, the elastic moduli of these rubbers decreased exponentially with increasing hexane 158 amount. When the hexane amount was 75–80 wt%, the elastic modulus was 1/50 of that of PDMS_0. In 159 this range, the rubbers became very soft and the data showed some deviations; however, the rubbers were 160 still sufficiently robust to be evaluated by the usual tensile test. The temperature dependence of the elastic 161 moduli in DMA and a tensile test of the PDMS rubber swollen by decane are described in the next section. 162  163 3.2. Swelling properties of PDMS rubbers 164 PDMS rubber sorbs a large amount of hexane. By contrast, the sorption amount of ethanol is very small. 165 Their swelling property is generally explained by crosslinking density of PDMS chains [32–36]. However, 166 our samples have a constant crosslinking density (0.25 %), and only volumes of added hexane are different. 167 Figure 4a shows the degree of swelling in hexane–ethanol mixtures plotted against the molar fraction of 168 hexane. The sorption amounts of pure hexane linearly increase with decreasing elastic moduli of the 169 PDMS rubbers, as shown on the right side of Figure 4a. Increase of sorption amount is due to the 170 contribution of polymer chain entanglements [37]. The sorption amounts of ethanol are almost zero for all 171 of the samples. At a hexane molar fraction of 0.4, 1 g of PDMS_h5 sorbed 2.2 g of the mixed solvent; 172 when the molar fraction was greater than 0.8, it sorbed 6.0 g of the mixed solvent. The change in sorption 173 amounts did not increase linearly but showed sigmoid-type sorption as a function of the hexane ratio. 174 Kappert et al. reported that crosslinked PDMS membranes sorbed 1.8 g g−1 of hexane [38]. Stafie et al. 175 observed the sorption amount increased up to 2.5 g g−1 with decreasing the crosslinking density [32]. By 176 contrast, the sorption amounts of PDMS_h2 to PDMS_h5 far exceed these values. 177 We also evaluated the hexane–ethanol compositions before and after the sorption experiments by 178 measuring the refractive indexes of the immersion solutions (Figure S3). In this case, a substantial change 179 in the composition was not observed. That is, hexane was not selectively sorbed at any fraction. 180 At molar fractions greater than 0.6, the sorption amounts increase in proportion to the hexane amount 181 added during the crosslinking reaction. The 3D configuration of the PDMS chains is memorized during 182 the crosslinking and is maintained even after drying, and the polymer chains return to their original 183 position when the PDMS is immersed in a good solvent. We refer to this phenomenon of polymer chains 184 returning to their original conformation as the "swelling memory effect." The same memory effect was 185 confirmed for hexane gas sorption, as described in the next section. Figure 4b shows images of PDMS_0 186 and PDMS_h5 in the dried state and after being swollen by hexane. PDMS_h5 exhibits a large volume 187 increase because of the memory effect; however, there was no change in shape because of the isotropic 188 expansion. Both samples are transparent, which means the rubbers are homogenous with respect to 189 refractive index. 190 Changes in free energy by the mixing of solvent molecules and PDMS chains were estimated using the 191 Flory–Huggins equation [32,33,36,39–47]. This equation is suitable for the thermodynamic evaluation of 192 PDMS and good solvent, especially when the crosslinking density is low. On the other hand, the Flory–193 Huggins equation cannot predict the distance between crosslinking points, which is usually evaluated by 194 the Flory–Rehner equation [33,34,36,37,46,48–52]. Using the Flory–Huggins equation, the mixing 195 enthalpy change (Δ𝐻 ) was calculated from the molecular weight of the polymer, interaction parameter 196 χ, and the volume fraction of the solvent (refer to the Supporting Information for details). The change in 197 mixing entropy (Δ𝑆 ) was calculated from the volume fraction. The results are shown in Figure 5a. The 198 energy plotted on the y-axis is normalized against 1 g of PDMS rubber. Figure 5b shows the mixing free 199 energy (Δ𝐺 ) calculated from Δ𝐻  and Δ𝑆 . In the case of PDMS_0 swollen by hexane at 298 K, 200 Δ𝐻  and 𝑇Δ𝑆  were 8.3 J g−1 and 14.2 J g−1, respectively. Note that the former is endothermic 201 (destabilization), whereas the latter is exothermic (stabilization). The Δ𝐺  for PDMS_h5 was stabilized 202 by 1.6 J g−1 compared with that of PDMS_0. This value is corresponding to 23 J mol−1 when converted to 203 the per-molar value of sorbed hexane. In case of ethanol sorption to PDMS_h5, the Δ𝐻 , 𝑇Δ𝑆 , and 204 Δ𝐺  were calculated to be 4.7, 8.6, and −3.8 J g−1, respectively. The value of Δ𝐺  was almost 205 constant for the series from PDMS_0 to PDMS_h5. 206 Figure 4. (a) Degree of swelling of PDMS_0, PDMS_h1, PDMS_h2, PDMS_h3, PDMS_h4, and PDMS_h5 rubbers in different hexane–ethanol mixtures. Swelling weights were measured using sealed containers to reduce the experimental error. (b) Volume increase of PDMS_0 and PDMS_h5 by hexane. Figure 5c shows the fluctuation of PDMS chains and retained hexane molecules in PDMS_0 and 207 PDMS_h5. The retention volume is small for PDMS_0 because the segment number is topologically 208 decreased by entanglement. By contrast, the retention volume for PDMS_h5 is large because of the small 209 degree of entanglement. The increase in the number of states of PDMS chains results in entropic 210 stabilization, consistent with PDMS_h5 exhibiting the largest 𝑇Δ𝑆  value among the investigated 211 samples (Figure 5a). 212 DSC measurements were conducted to evaluate the crystallization behaviors of the polymer and solvent 213 molecules. First, each sample was cooled to −90 ℃ and the DSC thermogram was acquired at a heating 214 rate of 0.5 ℃ min−1 (Figure S4). The thermogram of dry-state PDMS_0 showed a very small peak (3.2 J 215 g−1) at −46.8 ℃, which is attributed to the melting of PDMS crystals [53]. Table 2 shows the results of 216 DSC measurements of sorbed dodecane in PDMS rubbers. Pure dodecane was also measured for reference 217 and exhibited a melting point of −8.6 ℃, with a melting enthalpy of 203 J g−1 (34.6 kJ mol−1). These 218 values are approximately the same as the reported values [54,55]. In the case of PDMS_do1, a broad 219 endothermic peak was observed at −11.9 ℃, along with a comparatively sharp peak of dodecane 220 nanocrystals at −8.6 ℃. The former peak is attributed to the dodecane molecules perturbed by flexible 221 Figure 5. (a) Flory–Huggins’ Δ𝐻  (■) and 𝑇Δ𝑆  (♦) values of hexane/PDMS mixtures, and Δ𝐻  (■) and 𝑇Δ𝑆  (♦) values of ethanol–PDMS mixtures at 298 K. (b) Δ𝐺  of hexane–PDMS (●) and ethanol–PDMS (●) systems. (c) Illustration of the fluctuation of PDMS chains and their retention of hexane.  PDMS chains, and the melting enthalpy was twice as large as the latter one. However, a peak or shoulder 222 was not clearly observed near −11.9 ℃ in the thermogram of PDMS_do3 or PDMS_do5; their 223 thermograms showed only a single peak near the melting peak of pure dodecane. Their FWHM values 224 increased four-fold compared with those of pure dodecane, and the end of the endothermic peak exceeded 225 the melting point of pure dodecane. We speculate that two effects influence the stability of the dodecane 226 nanocrystals: destabilization by the fluctuation of PDMS chains, and stabilization by the network structure 227 of crosslinked PDMS chains. The molar melting enthalpy of dodecane in PDMS rubbers was 10–12% 228 lower than that of pure dodecane. 229  230 3.3. Hexane gas sorption 231 For the hexane gas sorption experiments, we prepared submillimeter-sized PDMS particles. A small 232 particle size is important for achieving a short equilibrium time in the sorption experiments. Figure 6a 233 shows the optical images of PDMS_0 and PDMS_de4 obtained by the cryogenic grinding method. By 234 selecting an appropriate mesh size of the sieve rings, we obtained 80 μm powders for hard PDMS_0 rubber 235 and 500 μm powders for soft PDMS_de4 rubber. Figure 6b shows the hexane sorption isotherms at 298 236 K (see also Figure S5 for the data at 263 K). The relative pressure, 𝑃/𝑃  (𝑃 : saturated vapor pressure) 237 is plotted on the abscissa. The sorption amount linearly increased for all samples in the 𝑃/𝑃  range from 238 0 to 0.6. The sorption amount at 𝑃/𝑃  = 0.95 (hexane pressure of 19.2 kPa) was 551 mg g−1 (PDMS_0), 239 641 mg g−1 (PDMS_de0.5), and 776 mg g−1 (PDMS_de4). The latter values of soft PDMS rubbers are not 240 surprising increases. The adsorption amount of hexane in conventional AC has been reported to be 410 241 mg g−1 at room temperature [11]. This adsorption amount is 55% of the sorption amount of PDMS_de4. 242 In the 𝑃/𝑃  range lower than 0.6, the hexane sorption obeys Henry’s law because the sorbed hexane 243 molecules are isolated among PDMS chains. However, when 𝑃/𝑃  is high, PDMS chains are partially 244 solvated by hexane molecules. The increase in the sorption amount of soft PDMS rubbers is a phenomenon 245 similar to the increase in the degree of swelling of the soft PDMS rubbers in liquid hexane. 246 The sorption amounts were 136 mg g−1 for PDMS_0 and 606 mg g−1 for PDMS_de4 at 195 K and 𝑃/𝑃  247 = 0.95 (23.8 Pa) (Figure 6c). PDMS rubber usually exhibits a glass-transition temperature (𝑇 ) of ~150 248 K; thus, both samples should exhibit rubber elasticity. However, the elastic modulus of PDMS_0 was six 249 Sample Weight ratio Dodecane / resin (g g－1) Peak position (℃) †FWHM (℃) ††Melting enthalpy, Δ𝐻  (kJ mol－1) †††Decrease in molar enthalpy (kJ mol－1) Macromer Crosslinker Dodecane PDMS_do1 1 0.1 1 0.91 −11.9 −8.6 3.2 0.6 20.0 10.3 −4.3 PDMS_do3 1 0.1 3 2.73 −8.6 0.8 31.3 −3.3 PDMS_do5 1 0.1 5 4.55 −8.5 0.7 31.2 −3.4 Dodecane - - - - −8.6 0.2 34.6 0 †Full-width of half-maximum. ††Molar enthalpy of dodecane. †††Difference from Δ𝐻  of pure dodecane. Table 2. Preparation conditions for PDMS rubbers, along with their dodecane content and DSC data. times higher than that of PDMS_de4 at 195 K, and this high elasticity led to decrease in sorption amount. 250 Figure 6d shows a comparison of the hexane sorption isotherms for PDMS_de4 at 298 K and 195 K, 251 plotted as a function of the absolute pressure. The inset shows isotherms in the low-pressure range. At 298 252 K and at 1.09 kPa, PDMS_de4 captures 7.60 mg g–1 of hexane. In the very low vapor pressure range, the 253 sorption amount obeys Henry’s law, and then the hexane sorption at 0.023 kPa is expected to be 0.16 mg 254 g–1. As marked by the red arrow at 23 Pa, the sorption amount at low temperature (195 K) was more than 255 3500 times greater than the sorption amount at room temperature; that is, PDMS_de4 could sorb large 256 amounts of hexane gas at a vapor pressure of a few tens of pascals (0.01–0.02 mol%, 300–600 ppm). For 257 comparison, the maximum hexane emission regulated by the National Institute for Occupational Safety 258 and Health (NIOSH) is a partial pressure of 1.5 Pa. To meet such a strict limit, trace amounts of hexane 259 need to be removed, for which a strong adsorbent such as AC is indispensable. However, the combined 260 Figure 6. (a) Optical microscope images of PDMS_0 and PDMS_de4 powders prepared by cryogenic grinding. (b) Sorption isotherms of PDMS_0 (●), PDMS_de0.5 (●), and PDMS_de4 (●) at 298 K. (c) Sorption isotherms of PDMS_0 and PDMS_de4 at 195 K. (d) Comparison of the sorption isotherms of PDMS_de4 at 298 and 195 K. The x-axis is shown as the absolute pressure. The inset shows magnified plots in the low-pressure range. use of PDMS sorbents could substantially reduce the burden on AC and make it possible to design an 261 environmentally conscious and low-cost VOC recovery system. 262 Figure 7a shows the sorption heats at 298 and 195 K, as calculated from the Clausius–Clapeyron 263 equation (see Supporting Information). The sorption heats at 298 K for PDMS_0 and PDMS_de4 at the 264 sorption amount of 80 mg g−1 were 34.5 and 32.9 kJ mol−1, respectively. These values exceed the 265 condensation heat of hexane of 28.9 kJ mol−1. As the sorption amount decreased from 20 to 0 mg g−1, the 266 sorption heat increased. This trend has often been observed for microporous adsorbents [56], implying the 267 formation of strong sorption sites in PDMS rubbers when small amounts of hexane are sorbed. At the 268 sorption amount of 80 mg g−1 at 195 K, the sorption heats for PDMS_0 and PDMS_de4 were 27.4 and 269 30.3 kJ mol−1, respectively. The condensation heat of hexane increases with decreasing temperature. We 270 calculated the value at 195 K to be 33.6 kJ mol−1 using Watson’s equation (see Supporting Information) 271 [57]. If compared with this value, the observed sorption heat is much smaller than the condensation heat. 272 At low temperatures, the elastic moduli of PDMS rubbers increase and the rubbers do not readily expand. 273 Some part of the exothermic heat acquired by the sorption of hexane must be consumed for the volume 274 expansion. The heat loss due to the volume expansion was 12 times larger at 195 K than at 298 K, as 275 Figure 7. (a) Comparison of the condensation heat (Δ𝐻 ) and the sorption heat (Δ𝐻 ) of PDMS_0 (●) and PDMS_de4 (●) calculated by the Clausius–Clapeyron equation at 298 and 195 K. (b) The energy level diagrams for hexane sorption in PDMS_de4 at 298 and 195 K. Δ𝐻 : mixing enthalpy change of hexane and PDMS; Expansion: energy loss by volume expansion of PDMS rubber; Stabilization: decrease in total energy mainly due to entropy change. calculated from the bulk moduli of PDMS_0. Small heat loss was confirmed even for PDMS_de4. These 276 results are discussed later in this section. As evident in Figure 7a (right), the sorption heat of PDMS_0 277 was 2.7–3.7 kJ mol−1 smaller than that of PDMS_de4 at 195 K. Interestingly, the sorption heat gradually 278 decreased as the sorption amount decreased from 80 to 0 mg g−1. This trend was opposite to the increase 279 in the sorption heat at 298 K at small sorption amounts between 20 and 0 mg g−1. The beginning of hexane 280 sorption at low temperatures incurs a large energy loss because the rubber’s elastic modulus is relatively 281 high. 282 Figure 7b shows the energy diagrams of hexane sorption into PDMS_de4 at 298 and 195 K. Both 283 temperatures are greater than 𝑇 ; thus, the PDMS chains exhibit liquid-like motion. Gaseous hexane is 284 captured by such polymer chains. The total exothermic energy consists of the gas-to-liquid condensation 285 and stabilization energies, mainly because of an increase in the mixing entropy of PDMS chains and 286 hexane molecules. For example, the total energy of the PDMS_de4 system at 298 K decreased by 28.9 kJ 287 mol−1 through condensation of hexane and was then further stabilized by the mixing entropy, although the 288 accurate value is unclear. However, some part of the exothermic heat was consumed by the endothermic 289 swelling (expansion) process and the Flory–Huggins’ mixing enthalpy (Δ𝐻 ). The sum of exothermic 290 and endothermic heats is consistent with the sorption heat calculated from the Clausius–Clapeyron 291 equation. PDMS rubbers undergo the same exothermic and endothermic processes at 195 K; however, 292 their energy loss due to volume expansion is much larger than that at 298 K because of the high elastic 293 modulus. The sorption heat thus becomes smaller than the condensation heat.  294 Changes of the elastic moduli and Poisson’s ratios of the PDMS rubbers after swelling with small 295 amounts of solvent were evaluated using decane, which exhibits low volatility at room temperature. 296 Figure 8a and 8b show the results. Dried PDMS_0 exhibited an elastic modulus of 2.0 MPa. The moduli 297 decreased quasi-linearly in the initial sorption range (0–8.0 wt%). The modulus for PDMS_0 became 1.1 298 MPa at 8 wt%. As shown in the inset, the elastic modulus of PDMS_de4 was 0.048 MPa in the absence 299 of decane and became 0.043 MPa when the sorption amount reached 7.8 wt%. Figure 8b shows the 300 changes in the Poisson’s ratio for PDMS_0 and PDMS_de4 as functions of the sorption amount (see also 301 Figure S6 for their logarithmic plots). The original value for PDMS_0 was 0.49, and the value became 302 0.33 when the sorption amount was 6.2 wt%. Although not shown in the figure, the Poisson’s ratio was 303 further reduced to 0.30 at 11.5 wt%. By contrast, the Poisson’s ratio of PDMS_de4 was 0.48 without 304 decane and 0.38 when the sorption amount was 6.3 wt%. In general, when rubbers are elongated, the 305 horizontal thinning is approximately one-half of the vertical elongation, giving a Poisson’s ratio of 0.50. 306 The PDMS_0 and PDMS_de4 have ratios close to 0.50. However, the Poisson's ratio of the PDMS rubbers 307 swollen by decane is far smaller than this value. Presumably, this low ratio is attributable to the solvent 308 molecules in swollen PDMS rubbers readily migrating to the elongated part and preventing the horizontal 309 thinning [58]. 310 Bulk modulus is important for characterizing swelling phenomena. The value was calculated from 311 elastic modulus and Poisson's ratio using the following equation: 312 𝐾 =𝐸3(1 − 2𝜈) 313 , where 𝐸 is the elastic modulus (Pa) and 𝜈 is the Poisson’s ratio. Figure 8c shows the decreases of the 314 bulk moduli of PDMS_0 and PDMS_de4 with increasing amount of sorbed decane. The bulk modulus of 315 the former rubber (42 MPa) rapidly decreased with increasing amount of sorbed decane. By contrast, the 316 latter rubber exhibited a small bulk modulus of 0.33 MPa from the beginning; this value is less than 1/130 317 of the value for PDMS_0. The inset shows enlarged plots of the decrease of the bulk modulus of 318 PDMS_de4. The value became 0.059 MPa when the sorption amount was 7.8 wt%. This modulus is 319 approximately 1/16 of that of PDMS_0 at the same sorption amount. Figure 8d shows the changes in 320 energy loss due to volume expansion (see Supporting Information). This value corresponds to the 321 Figure 8. The changes in (a) elastic modulus, (b) Poisson’s ratio, (c) bulk modulus, and (d) energy loss due to volume expansion with increasing sorption amount of decane at 298 K. PDMS rubbers: PDMS_0 (●) and PDMS_de4 (●). (e) The relationship between storage modulus and temperature for PDMS_0 and PDMS_de4 in the dried state. (f) Energy loss at 195 K for PDMS_0 and PDMS_de4. The insets in (a), (c), (d), and (f) are the magnified figures for PDMS_de4. integration of the bulk moduli in Figure 8c, and the unit of the calculated energy is joules per 1 g of dried 322 PDMS rubber, which was converted to kilojoules per 1 mol of decane in Figure 8d. The energy loss at 323 the beginning of sorption was 4.9 kJ mol−1 for PDMS_0 and 0.04 kJ mol−1 for PDMS_de4. The latter was 324 less than 1/120 of that of the former. As shown in the inset, the energy loss for PDMS_de4 at the sorption 325 amount of 7.8 wt% was 0.012 kJ mol−1, or approximately 1/33 of the energy loss for PDMS_0. 326 The elastic moduli at low temperatures were evaluated by DMA. Figure 8e shows the temperature 327 dependency of the storage moduli at 1 Hz. At 195 K, the storage modulus was 24.8 MPa for PDMS_0 and 328 4.0 MPa for PDMS_de4. Figure 8f shows the energy loss due to volume expansion at 195 K. We assumed 329 that the Poisson’s ratio did not change in the range 298–195 K. The energy loss for PDMS_0 at the 330 beginning of sorption was then 59 kJ mol−1. By contrast, the energy loss for PDMS_de4 was 3.6 kJ mol−1, 331 which was less than 1/16 of that for PDMS_0. 332 Notably, the energy loss due to the volume expansion at 195 K is more than ten times greater than that 333 at 298 K. The former value (59 kJ mol−1) far exceeds the condensation heat at 195 K (33.6 kJ mol−1), 334 which is why the sorption of hexane gas at low temperatures is suppressed. However, when the 𝑃/𝑃  335 value increases to greater than 0.5, PDMS_0 sorbs a certain amount of hexane because the bulk modulus 336 of this rubber (energy loss due to expansion) decreases with increasing amount of sorbed hexane. 337 Importantly, compared with the condensation heat of hexane, the expansion energy loss of PDMS_de4 338 (3.6 kJ mol−1) was very small from the beginning. Therefore, this rubber acts as a powerful sorbent for 339 hexane even at very low temperatures. 340 Figure 9 shows sorption isotherms of hexane gas for conventional AC at 283 K and PDMS_de4 at 195 341 K. The sorption isotherm of PDMS_de4 clearly differs from that of the AC. The maximum sorption 342 amount of PDMS_de4 exceeds that of the AC in the high 𝑃/𝑃  range. The sorption heat of AC at room 343 temperature has been reported to be 75 kJ mol−1, which is twice the sorption heat of PDMS_de4 (Table 344 S1). A high temperature of 882 K or greater is then required for desorption from AC at ambient pressure, 345 Figure 9. Comparison of hexane sorption isotherms of activated carbon [11] at 283 K and PDMS_de4 at 195 K. as calculated from Trouton’s rule [11]. In sharp contrast, a decrease in pressure is sufficient for PDMS_de4 346 to desorb hexane even at low temperatures. 347 Conventional adsorbents such as activated carbon, zeolites, and mesoporous materials have nanometer 348 to subnanometer pores that do not change size during the adsorption of gasses. The adsorption behaviors 349 in these materials are often explained by a micropore filling mechanism or a capillary model. In sharp 350 contrast, polymeric sorbents do not have intrinsic pores and the volume increases with increasing sorption 351 amount. In this case, the sorption of gasses follows Henry’s law when the sorption temperature is greater 352 than the 𝑇  of the polymer. The effect of the elastic modulus on the polymer is then substantial. Therefore, 353 the physicochemical sorption behaviors of polymers have not been quantitatively evaluated compared 354 with those of inorganic adsorbents. 355 Numerous studies on metal–organic frameworks (MOFs) have been reported over the past two decades 356 [59]. These materials possess a soft crystalline structure and exhibit unusual adsorption behaviors. 357 Kitagawa et al. reported that porous materials synthesized from copper (II) nitrate, 2,5-dihydroxybenzoic 358 acid, and 4,4′-bipyridine showed a gate opening/closing effect in response to the partial pressure of N2 359 [60]. Flexible metal–organic microporous materials can strongly adsorb acetylene over CO2 if the 360 chemical interaction inside the pores is optimized [61]. The swelling of rubbers was actively studied in 361 the 1940s. Flory and Rehner reported that the degrees of swelling can be predicted from the crosslinking 362 density of rubber and the χ parameters of the rubber and the solvent [62]. In the 1960s, gas sorption and 363 diffusion behaviors of polymer membranes were analyzed using the dual-sorption theory [63]. However, 364 substantially high sorption capacity of solvent vapors has not been well studied in either the polymer 365 science or the physicochemistry field. We combined low-temperature gas sorption measurements with 366 static and dynamic mechanical analyses of soft rubbers and evaluated the sorption behavior using the 367 Clausius–Clapeyron equation, estimations of the energy loss due to volume expansion, and the Flory–368 Huggins’ mixing enthalpy. To the best of our knowledge, at low temperatures, this work is the first report 369 of a quantitative evaluation of the sorption behavior of good solvents in soft rubbers. Based on the results 370 reported here, we have examined the separation potential of hexane in our scaled-up bench plant. Woven 371 nylon fabric packages containing PDMS powders were used for separation column to make the handling 372 of PDMS powders much easier and to provide the pathway of hexane gas. The details were explained in 373 Supporting Information (Figure S7). 374  375 4. Conclusions 376 PDMS rubbers are expected to be valuable sorbents for hexane recovery. We have explained that PDMS 377 macromers diluted with a large amount of hexane (or decane) can be loosely crosslinked, giving soft 378 rubbers with a reduced degree of entanglement. The elastic modulus exponentially decreased with 379 increasing volume ratio of the solvent. The degree of swelling of soft PDMS rubbers in hexane increased 380 substantially because the 3D configuration of the PDMS chains was memorized during the crosslinking. 381 Even after drying, the rubber returned to its original structure when immersed in hexane; that is, the 382 swelling memory effect was observed. Enhanced sorption ability was also observed when the PDMS 383 rubbers were exposed to hexane gas. In particular, more than 600 mg g−1 of hexane could be captured at 384 low temperatures, even though the partial pressure was low (10–20 Pa). We found that the high sorption 385 amount could be well explained by the substantial decrease in endothermic energy due to the small bulk 386 modulus of soft PDMS rubbers. 387 If regeneration is not desired, ACs will be the best choice to remove trace amounts of VOCs. However, 388 for the recovery of large amounts of hydrocarbons, the weakly sorbing PDMS rubber becomes important. 389 In fact, the recovery of VOCs at large scale is demanded in various fields, including the chemical industry 390 and the printing and paint industries. The recovery of halogenated solvents in the semiconductor industry 391 is also strongly demanded, as is the removal of greenhouse gasses in the oil and gas industry. We should 392 also emphasize that most of the industrially available adsorbents such as AC and zeolite are not useful 393 under high humidity conditions because of the strong affinity to water. On the other hand, our PDMS 394 rubber sorbs no more than 1 wt% of water, making it absolutely advantageous in the separation process 395 of VOCs even under saturated humidity. 396  397 Declaration of Competing Interest 398 The authors declare that they have no known competing financial interests or personal relationships that 399 could have appeared to influence the work reported in this paper. 400  401 Acknowledgements 402 The authors are grateful for the financial support of Cabinet Office, Government of Japan, Cross-403 ministerial Moonshot Agriculture, Forestry and Fisheries Research and Development Program, 404 "Technologies for Smart Bio-industry and Agriculture" (No. JPJ009237) and for JST-START program 405 (ST221006UM). The authors also appreciate Mr. R. Ema of National Institute of Technology, Tokyo 406 College for his supports in literature arrangement.  407  408 Appendix A. Supporting Information 409 Supplementary data to this article can be found online at https://⁓. Viscosities, elastic moduli, refractive 410 indexes, DSC thermograms, sorption isotherms, and Poisson ratios are shown in Figure S1–S6. 411 Calculation or estimation methods of 𝑟 , Δ𝐻 , Δ𝑆 , Δ𝐺 , Δ𝐻 , Δ𝐻 , 𝐸 , and comparisons of 412 sorption heats are described. 413  414  415 References 416 [1] F.I. Khan, A. K. Ghoshal, Removal of volatile organic compounds from polluted air, J. Loss Prev. 417 Process Ind. 13(6) (2000) 527–545. https://doi.org/10.1016/S0950-4230(00)00007-3. 418 [2] G.R. Parmar, N.N. Rao, Emerging control technologies for volatile organic compounds, Crit. 419 Rev. Environ. Sci. Technol. 39(1) (2008) 41–78. https://doi.org/10.1080/10643380701413658. 420 [3] M.S. Kamal, S.A. Razzak, M.M. Hossain, Catalytic oxidation of volatile organic compounds 421 (VOCs) – A review, Atmos. Environ. 140 (2016) 117–134. 422 https://doi.org/10.1016/j.atmosenv.2016.05.031. 423 [4] B.C. McDonald, J.A. de Gouw, J.B. Gilman, S.H. Jathar, A. Akherati, C.D. Cappa, J.L. Jimenez, 424 J. Lee-Taylor, P.L. Hayes, S.A. McKeen, Y.Y. Cui, S.W. Kim, D.R. Gentner, G. Isaacman-425 VanWertz, A.H. Goldstein, R.A. Harley, G.J. Frost, J.M. Roberts, T.B. Ryerson, M. Trainer, 426 Volatile chemical products emerging as largest petrochemical source of urban organic emissions, 427 Science. 359(6377) (2018) 760–764. https://doi.org/10.1126/science.aaq0524. 428 [5] P. Dwivedi, V. Gaur, A. Sharma, N. Verma, Comparative study of removal of volatile organic 429 compounds by cryogenic condensation and adsorption by activated carbon fiber, Sep. Purif. 430 Technol. 39(1-2) (2004) 23–37. https://doi.org/10.1016/j.seppur.2003.12.016. 431 [6] T.C. Bowen, R.D. Noble, J.L. Falconer, Fundamentals and applications of pervaporation through 432 zeolite membranes, J. Memb. Sci. 245(1-2) (2004) 1–33. 433 https://doi.org/10.1016/j.memsci.2004.06.059. 434 [7] L. Zhu, D. Shen, K.H. Luo, A critical review on VOCs adsorption by different porous materials: 435 Species, mechanisms and modification methods, J. Hazard. Mater. 389 (2020) 122102. 436 https://doi.org/10.1016/j.jhazmat.2020.122102. 437 [8] Y. Ding, Volatile organic compound liquid recovery by the dead end gas separation membrane 438 process: Theory and process simulation, Ind. Eng. Chem. Res. 58(12) (2019) 5008–5017. 439 https://doi.org/10.1021/acs.iecr.9b00586. 440 [9] J.G. Wijmans, A.L. Athayde, R. Daniels, J.H. Ly, H.D. Kamaruddin, I. Pinnau, The role of 441 boundary layers in the removal of volatile organic compounds from water by pervaporation, J. 442 Memb. Sci. 109(1) (1996) 135–146. https://doi.org/10.1016/0376-7388(95)00194-8. 443 [10] E.J. Simpson, W.J. Koros, R.S. Schechter, An emerging class of volatile organic compound 444 sorbents: Friedel−Crafts modified polystyrenes. 2. Performance comparison with commercially-445 available sorbents and isotherm analysis, Ind. Eng. Chem. Res. 35(12) (1996) 4635–4645. 446 https://doi.org/10.1021/ie950731l. 447 [11] R. Koyama, F. Tsunoda, I. Ichinose, H. Kanoh, Adsorption properties of methane, ethane, and 448 hexane on mesoporous organic polymers prepared by the flash freezing method, Langmuir. 36(9) 449 (2020) 2184–2190. https://doi.org/10.1021/acs.langmuir.9b03159. 450 [12] Y. Yang, L. Li, R.B. Lin, Y. Ye, Z. Yao, L. Yang, F. Xiang, S. Chen, Z. Zhang, S. Xiang, B. 451 Chen, Ethylene/ethane separation in a stable hydrogen-bonded organic framework through a 452 gating mechanism, Nat. Chem. 13 (2021) 933–939. https://doi.org/10.1038/s41557-021-00740-z. 453 [13] David. S. Sholl, Ryan P. Lively., Seven chemical separations to change the world, Nature. 532 454 (2016) 435–437. 455 [14] S. Brunauer, P.H. Emmett, E. Teller, Adsorption of gases in multimolecular layers, J. Am. Chem. 456 Soc. 60(2) (1938) 309–319. https://doi.org/10.1021/ja01269a023. 457 [15] W. Thomson, LX. On the equilibrium of vapour at a curved surface of liquid, London, Edinburgh, 458 Dublin Philos. Mag. J. Sci. 42(282) (1871) 448–452. 459 https://doi.org/10.1080/14786447108640606. 460 [16] P.J. Flory, Thermodynamics of high polymer solutions, J. Chem. Phys. 10(1) (1942) 51–61. 461 https://doi.org/10.1063/1.1723621. 462 [17] D.F. Sanders, Z.P. Smith, R. Guo, L.M. Robeson, J.E. McGrath, D.R. Paul, B.D. Freeman, 463 Energy-efficient polymeric gas separation membranes for a sustainable future: A review, 464 Polymer. 54(18) (2013) 4729–4761. https://doi.org/10.1016/j.polymer.2013.05.075 465 [18] Z. Si, J. Li, L. Ma, D. Cai, S. Li, J. Baeyens, J. Degreve, J. Nie, T. Tan, P. Qin, The Ultrafast and 466 Continuous Fabrication of a Polydimethylsiloxane Membrane by Ultraviolet-Induced 467 Polymerization, Angew. Chem. Int. Ed. Engl. 58(48) (2019) 17175-17179. 468 https://doi.org/10.1002/anie.201908386. 469 [19] W. Kong, J. Baeyens, P. Qin, H. Zhang, T. Tan, Towards an energy-friendly and cleaner solvent-470 extraction of vegetable oil, J. Environ. Manage. 217 (2018) 196-206. 471 https://doi.org/10.1016/j.jenvman.2018.03.061. 472 [20] W. Kong, Q. Miao, P. Qin, J. Baeyens, T. Tan, Environmental and economic assessment of 473 vegetable oil production using membrane separation and vapor recompression, Frontiers of 474 Chemical Science and Engineering 11(2) (2017) 166-176. https://doi.org/10.1007/s11705-017-475 1616-4. 476 [21] Z. Si, G. Li, Z. Wang, D. Cai, S. Li, J. Baeyens, P. Qin, A Particle-Driven, Ultrafast-Cured 477 Strategy for Tuning the Network Cavity Size of Membranes with Outstanding Pervaporation 478 Performance, ACS Appl Mater Interfaces 12(28) (2020) 31887-31895. 479 https://doi.org/10.1021/acsami.0c05859. 480 [22] E. Ismail, N.H. Lazim, A. Nakata, A. Iwasawa, R. Yamanaka, I. Ichinose, Plasma-induced 481 Interfacial Crosslinking of Liquid Polydimethylsiloxane Films and Their Organic Solvent 482 Permeation Performance, Chem. Lett. 49(11) (2020) 1286-1290. 483 https://doi.org/10.1246/cl.200504. 484 [23] S.M. Jordan, W.J. Koros, Permeability of pure and mixed gases in silicone rubber at elevated 485 pressures, J. Polym. Sci. Part B Polym. Phys. 28(6) (1990) 795–809. 486 https://doi.org/10.1002/polb.1990.090280602. 487 [24] R.D. Raharjo, B.D. Freeman, D.R. Paul, G.C. Sarti, E.S. Sanders, Pure and mixed gas CH4 and n-488 C4H10 permeability and diffusivity in poly(dimethylsiloxane), J. Memb. Sci. 306(1-2) (2007) 75–489 92. https://doi.org/10.1016/j.memsci.2007.08.014. 490 [25] L. Liu, E.S. Sanders, J.R. Johnson, O. Karvan, S. Kulkarni, D.J. Hasse, W.J. Koros, Influence of 491 membrane skin morphology on CO2/N2 separation at sub-ambient temperatures, J. Memb. Sci. 492 446 (2013) 433–439. https://doi.org/10.1016/j.memsci.2013.06.001. 493 [26] G. Wulff, Fourty years of molecular imprinting in synthetic polymers: origin, features and 494 perspectives, Microchim. Acta. 180 (2013) 1359–1370. https://doi.org/10.1007/s00604-013-0992-495 9. 496 [27] L. Chen, X. Wang, W. Lu, X. Wu, J. Li, Molecular imprinting: perspectives and applications, 497 Chem. Soc. Rev. 45 (2016) 2137–2211. https://doi.org/10.1039/C6CS00061D. 498 [28] V. Arrighi, S. Gagliardi, A.C. Dagger, J.A. Semlyen, J.S. Higgins, M.J. Shenton, Conformation of 499 cyclics and linear chain polymers in bulk by SANS, Macromolecules. 37(21) (2004) 8057–8065. 500 https://doi.org/10.1021/ma049565w. 501 [29] J.S. Higgins, K. Dodgson, J.A. Semlyen, Studies of cyclic and linear poly(dimethyl siloxanes): 3. 502 Neutron scattering measurements of the dimensions of ring and chain polymers, Polymer. 20(5) 503 (1979) 553–558. https://doi.org/10.1016/0032-3861(79)90164-2. 504 [30] W.H. Briscoe, Polymers and Nanoscience, in: Colloid. Found. Nanosci., Elsevier, 2014: pp. 107–505 133. https://doi.org/10.1016/B978-0-444-59541-6.00005-9. 506 [31] P.G. de Gennes, Scaling Concepts in Polymer Physics, Cornell University Press, 1979. 507 [32] N. Stafie, D.F. Stamatialis, M. Wessling, Effect of PDMS cross-linking degree on the permeation 508 performance of PAN/PDMS composite nanofiltration membranes, Sep. Purif. Technol. 45 (2005) 509 220–231. https://doi.org/10.1016/j.seppur.2005.04.001. 510 [33] B. Xu, J. Wu, G.B. McKenna, Mechanical and swelling behaviors of end-linked PDMS Rubber 511 and randomly cross-linked polyisoprene, Macromolecules. 46 (2013) 2015–2022. 512 https://doi.org/10.1021/ma302335u. 513 [34] A.M. Kansara, V.K. Aswal, P.S. Singh, Preparation and characterization of new 514 poly(dimethylsiloxane) membrane series via a ‘cross-linking’ reaction using monomolecular 515 trichloro(alkyl)silane of different alkyl chain and type, RSC Adv. 5(16) (2015) 51608–51620. 516 https://doi.org/10.1039/C5RA06433C. 517 [35] G.L. Jadav, V.K. Aswal, P.S. Singh, In-situ preparation of polydimethylsiloxane membrane with 518 long hydrophobic alkyl chain for application in separation of dissolved volatile organics from 519 wastewater, J. Memb. Sci. 492 (2015) 95–106. https://doi.org/10.1016/j.memsci.2015.05.050. 520 [36] W. Ogieglo, H. van der Werf, K. Tempelman, H. Wormeester, M. Wessling, A. Nijmeijer, N.E. 521 Benes, n-Hexane induced swelling of thin PDMS films under non-equilibrium nanofiltration 522 permeation conditions, resolved by spectroscopic ellipsometry, J. Memb. Sci. 431 (2013) 233–523 243. https://doi.org/10.1016/j.memsci.2012.12.045. 524 [37] S.K. Patel, S. Malone, C. Cohen, J.R. Gillmor, R.H. Colby, Elastic modulus and equilibrium 525 swelling of poly(dimethylsiloxane) networks, Macromolecules. 25(29) (1992) 5241–5251. 526 https://doi.org/10.1021/ma00046a021. 527 [38] E.J. Kappert, M.J.T. Raaijmakers, K. Tempelman, F.P. Cuperus, W. Ogieglo, N.E. Benes, 528 Swelling of 9 polymers commonly employed for solvent-resistant nanofiltration membranes: A 529 comprehensive dataset, J. Memb. Sci. 569 (2019) 177–199. 530 https://doi.org/10.1016/j.memsci.2018.09.059. 531 [39] S. Postel, C. Schneider, M. Wessling, Solvent dependent solute solubility governs retention in 532 silicone based organic solvent nanofiltration, J. Memb. Sci. 497 (2016) 47–54. 533 https://doi.org/10.1016/j.memsci.2015.09.014. 534 [40] F. Horkay, M. Zrinyi, E. Geissler, A.M. Hecht, P. Pruvost, Effect of neutral silica particles on the 535 macroscopic swelling and elastic properties of polydimethyl siloxane networks, Polymer. 32(5) 536 (1991) 835–839. https://doi.org/10.1016/0032-3861(91)90507-F. 537 [41] S.J. Lue, J.S. Ou, C.H. Kuo, H.Y. Chen, T.H. Yang, Pervaporative separation of azeotropic 538 methanol/toluene mixtures in polyurethane–poly(dimethylsiloxane) (PU–PDMS) blend 539 membranes: Correlation with sorption and diffusion behaviors in a binary solution system, J. 540 Memb. Sci. 347(1-2) (2010) 108–115. https://doi.org/10.1016/j.memsci.2009.10.012. 541 [42] E. Favre, Swelling of crosslinked polydimethylsiloxane networks by pure solvents: Influence of 542 temperature, Eur. Polym. J. 32(10) (1996) 1183–1188. https://doi.org/10.1016/S0014-543 3057(96)00062-6. 544 [43] E. Favre, Q.T. Nguyen, D. Sacco, A. Moncuy, R. Clement, Multicomponent polymer/solvents 545 equilibria: an evaluation of Flory-Huggins theory for crosslinked PDMS networks swelled by 546 binary mixtures, Chem. Eng. Commun. 140(1) (1995) 193–205. 547 https://doi.org/10.1080/00986449608936463. 548 [44] N. Schuld, B.A. Wolf, Solvent quality as reflected in concentration- and temperature-dependent 549 Flory-Huggins interaction parameters, J. Polym. Sci. Part B Polym. Phys. 39(6) (2001) 651–662. 550 https://doi.org/10.1002/1099-0488(20010315)39:6<651::AID-POLB1039>3.0.CO;2-1. 551 [45] A. Muramoto, Studies on the Interaction Parameter in Polysiloxane Solutions, Polym. J. 1 (1970) 552 450–456. https://doi.org/10.1295/polymj.1.450. 553 [46] Y. Zhao, B.E. Eichinger, Study of solvent effects on the dilation modulus of 554 poly(dimethylsiloxane), Macromolecules. 25(25) (1992) 6988–6995. 555 https://doi.org/10.1021/ma00051a041. 556 [47] Y. Hu, X. Chen, G.M. Whitesides, J.J. Vlassak, Z. Suo, Indentation of polydimethylsiloxane 557 submerged in organic solvents, J. Mater. Res. 26(6) (2011) 785–795. 558 https://doi.org/10.1557/jmr.2010.35. 559 [48] N.A. Neuburger, B.E. Eichinger, Critical experimental test of the Flory-Rehner theory of 560 swelling, Macromolecules. 21(10) (1988) 3060–3070. https://doi.org/10.1021/ma00188a026. 561 [49] G. Cocchi, M.G. De Angelis, F. Doghieri, Solubility and diffusivity of liquids for food and 562 pharmaceutical applications in crosslinked polydimethylsiloxane (PDMS) films: I. Experimental 563 data on pure organic components and vegetable oil, J. Memb. Sci. 492 (2015) 600–611. 564 https://doi.org/10.1016/j.memsci.2015.04.063. 565 [50] S.P. Malone, C. Vosburgh, C. Cohen, Validity of the swelling method for the determination of the 566 interaction parameter, Polymer. 34(24) (1993) 5149–5153. https://doi.org/10.1016/0032-567 3861(93)90261-8. 568 [51] W. Chassé, M. Lang, J.U. Sommer, K. Saalwächter, Cross-link density estimation of PDMS 569 networks with precise consideration of networks defects, Macromolecules. 45(2) (2012) 899–912. 570 https://doi.org/10.1021/ma202030z. 571 [52] T. Shiomi, K. Kuroki, A. Kobayashi, H. Nikaido, M. Yokoyama, Y. Tezuka, K. Imai, 572 Dependence of swelling degree on solvent composition of two-component copolymer networks in 573 mixed solvents, Polymer. 36(12) (1995) 2443–2449. https://doi.org/10.1016/0032-574 3861(95)97346-H. 575 [53] H. Yang, Q.T. Nguyen, Y. Ding, Y. Long, Z. Ping, Investigation of poly(dimethyl siloxane) 576 (PDMS)–solvent interactions by DSC, J. Memb. Sci. 164(1-2) (2000) 37–43. 577 https://doi.org/10.1016/S0376-7388(99)00187-8. 578 [54] J.B. Ott, J.R. Goates, (Solid + liquid) phase equilibria in binary mixtures containing benzene, a 579 cycloalkane, an n-alkane, or tetrachloromethane An equation for representing (solid + liquid) 580 phase equilibria, J. Chem. Thermodyn. 15(3) (1983) 267–278. https://doi.org/10.1016/0021-581 9614(83)90119-2. 582 [55] E.S. Domalski, E.D. Hearing, Heat capacities and entropies of organic compounds in the 583 condensed phase. Volume III, J. Phys. Chem. Ref. Data. 25(1) (1996) 1–525. 584 https://doi.org/10.1063/1.555985. 585 [56] H. Pan, J.A. Ritter, P.B. Balbuena, Examination of the approximations used in determining the 586 isosteric heat of adsorption from the Clausius−Clapeyron equation, Langmuir. 14(21) (1998) 587 6323–6327. https://doi.org/10.1021/la9803373. 588 [57] K.M. Watson, Thermodynamics of the liquid state, Ind. Eng. Chem. 35(4) (1943) 398–406. 589 https://doi.org/10.1021/ie50400a004. 590 [58] Y. Lai, Y. Hu, Probing the swelling-dependent mechanical and transport properties of 591 polyacrylamide hydrogels through AFM-based dynamic nanoindentation, Soft Matter. 14 (2018) 592 2619–2627. https://doi.org/10.1039/C7SM02351K. 593 [59] S. Kitagawa, R. Kitaura, S. Noro, Functional porous coordination polymers, Angew. Chemie Int. 594 Ed. 43(18) (2004) 2334–2375. https://doi.org/10.1002/anie.200300610. 595 [60] R. Kitaura, K. Seki, G. Akiyama, S. Kitagawa, Porous coordination-polymer crystals with gated 596 channels specific for supercritical gases, Angew. Chemie Int. Ed. 42(4) (2003) 428–431. 597 https://doi.org/10.1002/anie.200390130. 598 [61] R. Matsuda, R. Kitaura, S. Kitagawa, Y. Kubota, R.V. Belosludov, T.C. Kobayashi, H. Sakamoto, 599 T. Chiba, M. Takata, Y. Kawazoe, Y. Mita, Highly controlled acetylene accommodation in a 600 metal-organic microporous material, Nature. 436 (2005) 238–241. 601 https://doi.org/10.1038/nature03852. 602 [62] P.J. Flory, J. Rehner, Statistical mechanics of cross‐linked polymer networks II. Swelling, J. 603 Chem. Phys. 11 (1943) 521–526. https://doi.org/10.1063/1.1723792. 604 [63] W.R. Vieth, P.M. Tam, A.S. Michaels, Dual sorption mechanisms in glassy polystyrene, J. 605 Colloid Interface Sci. 22(4) (1966) 360–370. https://doi.org/10.1016/0021-9797(66)90016-6. 606  607  608 FIGURES 609 [COLOR] Figure 1. Preparation of PDMS rubbers and a schematic of the PDMS dispersions with 610 different concentrations. The sample names are shown in the yellow box. 611  612 [COLOR] Figure 2. Presumed crosslinking structures of PDMS_0 (left) and PDMS_h5 with solvent 613 molecules (right). 614  615 [COLOR] Figure 3. (a) Stress–strain curves for PDMS_h1, PDMS_h2, PDMS_h3, PDMS_h4, and 616 PDMS_h5. The inset is the stress–strain curves for PDMS_0 and PDMS_h1. (b) Pictures of PDMS_0 and 617 PDMS_h5 immediately before breaking and illustrations of their crosslinked polymer chains. 618  619 [COLOR] Figure 4. (a) Degree of swelling of PDMS_0, PDMS_h1, PDMS_h2, PDMS_h3, PDMS_h4 620 and PDMS_h5 rubbers in different hexane–ethanol mixtures. Swelling weights were measured using 621 sealed containers to reduce the experimental error. (b) Volume increase of PDMS_0 and PDMS_h5 by 622 hexane. 623  624 [COLOR] Figure 5. (a) Flory–Huggins’ Δ𝐻  (■) and 𝑇Δ𝑆  (♦) values of hexane/PDMS mixtures, 625 and Δ𝐻  (■) and 𝑇Δ𝑆  (♦) values of ethanol–PDMS mixtures at 298 K. (b) Δ𝐺  of hexane–626 PDMS (●) and ethanol–PDMS (●) systems. (c) Illustration of the fluctuation of PDMS chains and their 627 retention of hexane. 628  629 [COLOR] Figure 6. (a) Optical microscope images of PDMS_0 and PDMS_de4 powders prepared by 630 cryogenic grinding. (b) Sorption isotherms of PDMS_0 (●), PDMS_de0.5 (●), and PDMS_de4 (●) at 298 631 K. (c) Sorption isotherms of PDMS_0 and PDMS_de4 at 195 K. (d) Comparison of the sorption isotherms 632 of PDMS_de4 at 298 and 195 K. The x-axis is shown as the absolute pressure. The inset shows magnified 633 plots in the low-pressure range. 634  635 [COLOR] Figure 7. (a) Comparison of the condensation heat (Δ𝐻 ) and the sorption heat (Δ𝐻 ) of 636 PDMS_0 (●) and PDMS_de4 (●) calculated by the Clausius–Clapeyron equation at 298 and 195 K. (b) 637 The energy level diagrams for hexane sorption in PDMS_de4 at 298 and 195 K. Δ𝐻 : mixing enthalpy 638 change of hexane and PDMS; Expansion: energy loss by volume expansion of PDMS rubber; 639 Stabilization: decrease in total energy mainly due to entropy change. 640  641 [COLOR] Figure 8. The changes in (a) elastic modulus, (b) Poisson’s ratio, (c) bulk modulus, and (d) 642 energy loss due to volume expansion with increasing sorption amount of decane at 298 K. PDMS rubbers: 643 PDMS_0 (●) and PDMS_de4 (●). (e) The relationship between storage modulus and temperature for 644 PDMS_0 and PDMS_de4 in the dried state. (f) Energy loss at 195 K for PDMS_0 and PDMS_de4. The 645 insets in (a), (c), (d), and (f) are the magnified figures for PDMS_de4. 646  647 [COLOR] Figure 9. Comparison of hexane sorption isotherms of activated carbon [11] at 283 K and 648 PDMS_de4 at 195 K, and the advantages of soft PDMS rubbers that exhibit the swelling memory effect. 649  650 TABLES 651 Table 1. Composition of reaction mixture, degree of overlapping concentration, and material properties 652 of the obtained PDMS rubbers.  653  654 Table 2. Preparation conditions for PDMS rubbers, along with their dodecane content and DSC data. 655  656