Introduction

In nature, gases trigger vital stimuli in living organisms. Animals smell gases to find and distinguish foods, associates, and dangers. Not only animals but also insects and plants sense gases, such as CO2. For example, some plants control stomatal conductance in their leaves depending on the CO2 concentration1. Mosquitos detect CO2 gas in animals’ breaths to identify targets2. Similarly, synthetic materials that respond to specific gases and switch their properties have been developed3,4,5,6. However, the applications of such gas-responsive materials have remained limited because they are mostly produced in the form of solutions7,8,9,10,11,12,13, liquids14,15,16,17, nanoparticles18, hydrogels19,20, and organogels21,22. From the viewpoint of applications, gas-responsive bulk materials that can be molded into sheets, membranes, plates, and objects of various shapes are strongly desired. However, the field of gas-responsive bulk materials is almost unexplored and considerably few polymeric23,24,25,26, metallic27, and ceramic28 materials that can enhance their functions with the aid of specific gases have been reported. In particular, materials that respond to CO2 gas are important because it is one of the major greenhouse gases and its reduction, capture, and effective utilization are imperative to prevent global warming. To date, various chemical materials such as membranes29,30, nanoporous materials31, catalysts32,33, and reactive monomers34,35 have been developed to capture, separate, and utilize CO2 gas. In this report, CO2-responsive bulk materials that switch their mechanical properties with CO2 gas have been reported. Moreover, we propose gas-induced mechanical reinforcement to enhance the mechanical properties (i.e., modulus, strength, stretchability, and toughness) of bulk materials using specific gases. The materials exhibit remarkable enhancements in mechanical properties upon exposure to even low concentrations of CO2 and some other gases. Moreover, these materials offer broad application prospects, including but not limited to sensors, tunable adhesives, shape-memory materials, actuators, and strength-tunable materials. These materials will provide a new path toward the effective utilization of CO2 gas.

Results and discussion

Material design

To achieve gas-induced mechanical reinforcement, we programmed components that generate reinforcement structures and/or morphologies through chemical and/or physical interactions with specific gases. In particular, the formation of physical or chemical crosslinks, the induction of microphase separations, and the transitions in morphologies were considered. The formation of crosslinks is the simplest method to enhance the modulus and strength of materials. Moreover, the induction of microphase separation may be effective for toughening materials through the formation of nanodomains. If nanodomains are rubbery at room temperature, they behave similarly to viscoelastic nanofillers that prevent the fracture of the material. In this case, the material’s properties can be tuned by varying the size, number density, and viscoelasticity of nanodomains.

To demonstrate the aforementioned concept, we selected chemically crosslinked polydimethylsiloxane (PDMS) elastomers modified with amines (Fig. 1a). The high gas permeability of PDMS36 facilitates a rapid reaction between gas molecules and amines even in the bulk material. Amines are converted to ammonium carbamates that function as ionic crosslinks upon a reaction with CO2 gas16. Moreover, ammonium carbamates aggregate into nanodomains in the hydrophobic PDMS matrix (Fig. 1c). In this polymer, the size and viscoelasticity of the nanodomains can be tuned by varying the size of the amine moiety (Fig. 1a). The generated viscoelastic nanodomains effectively dissipate a large amount of energy and prevent crack propagation upon material deformation; thus, the fracture of the material is efficiently prevented (Fig. 1c).

Fig. 1: Molecular structures and CO2 toughening behavior of elastomers.
figure 1

a Chemical structures of PDMS elastomers modified with amines. b Photograph of a CL–M5 sheet that is colorless and transparent. c Compression behavior of a block of CL–M5 in N2 and CO2 gases. CL–M5 is extremely toughened with CO2 gas while it is brittle in a N2 environment. See Supplementary Movie 3 and Supplementary Movie 4 in detail.

Full size image

Materials synthesis and characterization

PDMS samples having monoamine and diamine groups, denoted as PDMS–Mx and PDMS–Dx (x indicates the content of amine in mol%), respectively, were synthesized via ring-opening polymerization (Supplementary Fig. 1). PDMS–Mx and PDMS–Dx were characterized via 1H-neuclear magnetic resonance (NMR) (Supplementary Fig. 2) and gel permeation chromatography (GPC). The weight-average molecular weight (Mw) and the polydispersity index (PDI) were determined via GPC using PDMS standards (Scientific Polymer Products Inc.); the amine concentrations determined via neutralization titration are listed in Table 1. PDMS–Mx and PDMS–Dx were chemically crosslinked using 1,4-butanediol diglycidyl ether (BDE; concentration = 8.5 × 10−5 mol g−1). Crosslinked PDMS–Mx and PDMS–Dx were denoted as CL–Mx and CL–Dx, respectively (Fig. 1a). Additionally, CL–T5 wherein PDMS was modified using tris(2-aminoethyl)amine (TAEA) was obtained via the addition of TAEA to epoxy-modified PDMS (Fig. 1a). The Mw and PDI values of epoxy-modified PDMS (KF-1001, Shin-Etsu Chemical Co., Ltd.) were determined to be 66700 and 1.89, respectively (Table 1). The concentration of the epoxy group was 2.9 × 10−4 mol g−1. CL–T5 contained 3 wt% TAEA, i.e., the amine concentration in CL–T5 was 6.3 × 10−4 mol g−1 (~4.7 mol%) without tertiary amine. The sheets of these elastomers (~ 1 mm thick) were colorless and transparent (Fig. 1b).

Table 1 Molecular characteristics of samples and parameters used for the simulation of SAXS patterns for samples with CO2 gas
Full size table

Microphase separation with CO2 gas

CL–M5, CL–D9, and CL–T5 sheets rapidly trap CO2 gas molecules even though they are relatively thick (~ 1 mm). The increase in the sheet weight under CO2 gas flow was monitored at 25 °C (Fig. 2a). The trapped CO2 molecules were slowly released under N2 gas flow. The amounts of CO2 molecules trapped in CL–M5, CL–D9, and CL–T5 were calculated to be 3.3 × 10−4, 5.1 × 10−4, and 2.1 × 10−4 mol g−1, respectively. These values were nearly half of the values of the concentration of the amine groups in each sample, indicating the formation of ammonium carbamates (Fig. 2a)16. When our elastomer materials were heated, the trapped CO2 was rapidly released (Supplementary Fig. 3). This result implies that the effect of CO2 on our materials is easily canceled by heating the materials. The formation of ammonium carbamates was further verified via Fourier-transform infrared (FT–IR) spectroscopy bands, N–H bending, C = O stretching, and C–N stretching bands at 1580, 1640, and 1470 cm−1, respectively16 (Fig. 2b). The formation of ammonium carbamates occurs even at low concentrations of CO2 gas (Fig. 2c). For example, 72% of amines in CL–D9 were converted to ammonium carbamates even with the N2/CO2 (95/5) mixed gas. Hydrophilic ammonium carbamates aggregate into nanodomains in the hydrophobic PDMS matrix as illustrated in Fig. 1c. In the small-angle X-ray scattering (SAXS) patterns, a broad and relatively strong scattering peak is exhibited for the samples reacted with CO2 gas owing to the formation of nanodomains, whereas marginal scattering from heterogeneities in the crosslinked PDMS matrix is observed in air (Fig. 2d). In case of noncrosslinked PDMS–M5 and PDMS–D9, similar scattering from the nanodomains with CO2 gas is observed; however, no scattering peak is detected in air (Supplementary Fig. 4). The size and number density of the nanodomains with ammonium carbamates were determined by simulating the experimental SAXS patterns. We found that the Yarusso–Cooper model37 satisfactorily reproduces the experimental SAXS patterns (Fig. 2d). This model assumes that spherical nanodomains with a radius (R1) are randomly distributed with the closest approach limitation (2Rca) in the matrix. The sample volume occupied by one nanodomain is Vp. The parameters used in the simulation of each sample are listed in Table 1. The volume (V = \(\frac{{{{{{\boldsymbol{4}}}}}}}{{{{{{\boldsymbol{3}}}}}}}{{{{{\boldsymbol{\pi }}}}}}{{{{{{\boldsymbol{R}}}}}}}_{{{{{{\boldsymbol{1}}}}}}}^{{{{{{\boldsymbol{3}}}}}}}\), nm3) and number density (ND ((10 nm)−3)) of nanodomains calculated using the simulation parameters are indicated in Fig. 2d. V increases with the increasing size of the amine group. Even though the V value of CL–M5 is relatively small, CL–M5 exhibits a considerably larger ND value than CL–D9, wherein the concentration of diamine side groups (~4.6 mol%) is comparable with the monoamine concentration in CL–M5. As shown in Fig. 2d, the intensity of the scattering peak of CL–M5 is small compared with that of CL–D9 despite the larger ND value of CL–M5, indicating that the electron density of the nanodomains in CL–M5 decreases upon the inclusion of polymer segments into the nanodomains. The inclusion of polymer segments into ionic aggregates is frequently observed in ionomers38. In our study, nanodomains are efficiently formed even at low concentrations of CO2 gas (N2/CO2 (90/10)) (Supplementary Fig. 5). Interestingly, V is almost insensitive to the CO2 concentration, whereas ND increases with the increasing CO2 concentration. This result confirms that the size of the amine group dominates the size and number density of the nanodomains upon CO2 gas exposure.

Fig. 2: Effects of CO2 on elastomer structure.
figure 2

a Weight changes of samples with different gas flows. The weights of samples increase with CO2 gas. PDMS is a homopolymer without amines (KF-96H-50000CS, Shin-Etsu Chemical). The chemical reaction between primary amines and CO2 is also shown. b FT–IR spectra of elastomers in N2 (black) and CO2 (colored) gases. Ammonium carbamates are generated with CO2. c Weight changes of CL–D9 in N2/CO2 mixed gases with different CO2 concentrations. d Experimental and simulated SAXS patterns of elastomers. The SAXS patterns with and without CO2 are indicated by the colored and black plots, respectively. The Yarusso–Cooper model is used for the simulation37. V (nm3) and ND ((10 nm)−3) correspond to the volume and number density of nanodomains, respectively.

Full size image

Enhancement of viscoelastic properties using CO2 gas

Upon the formation of nanodomains using CO2 gas, additional relaxation occurs owing to molecular motions in the nanodomains in linear viscoelastic measurements (Fig. 3a–e). For CL–M5 (Fig. 3a) and CL–D9 (Fig. 3d), additional relaxation occurred even in the N2/CO2 (90/10) mixed gas. Because of this additional relaxation, the shear storage modulus (Gʹ), shear loss modulus (Gʺ), and loss tangent (tanδ, Gʺ Gʹ−1) of the elastomers increase considerably at room temperature. In particular, this effect is pronounced for CL–D5 (Fig. 3c) and CL–D9 (Fig. 3d) in which additional relaxations occur near room temperature. The relaxation temperatures determined from the Gʺ peaks of CL–M5, CL–M10, CL–D5, CL–D9, and CL–T5 are less than −30 °C, approximately −30 °C, 21 °C, 23 °C, and 39 °C, respectively, while the glass transition temperature (Tg) of PDMS measured using differential scanning calorimetry is ~−122 °C (Supplementary Fig. 6). The relaxation temperature of the nanodomains increases with the increasing size of amine group, signifying that the viscoelastic properties of our materials can be tuned by modifying the size and content of the amine group upon CO2 gas exposure. Among our materials, CL–D9 toughened with CO2 gas exhibits the largest Gʺ value at room temperature (25 °C) (Fig. 3d). Mullins reported that Gʺ and the fracture energy values of elastomers are correlated39. This study demonstrates that effective energy dissipation during deformation is crucial for toughening materials. In fact, CL–D9 with CO2 gas exhibited high toughness as discussed below. Furthermore, CO2 gas rapidly changed the viscoelasticity of our materials owing to the high gas permeability of PDMS. The tensile storage modulus (Eʹ), tensile loss modulus (Eʺ), and loss tangent (tanδ, Eʺ Eʹ−1) of CL–D9 instantaneously increase with CO2 gas flow, whereas these parameters gradually return to their initial values when exposed to N2 gas flow at 25 °C (Fig. 3f). This behavior is consistent with the results of the CO2 uptake shown in Fig. 2a. The increases in Gʹ and Gʺ with the increasing CO2 concentration were similarly observed for PDMS oils containing amine groups16. However, the values of Gʹ and Gʺ were less than one kPa because of their small molecular wight and the absence of the chemical crosslinks.

Fig. 3: Effects of CO2 on the viscoelastic properties of elastomers.
figure 3

Effects of each gas on the shear storage modulus (Gʹ), shear loss modulus (Gʺ) and loss tangent (tanδ) of 1 Hz for CL–M5 a, CL–M10 b, CL–D5 c, CL–D9 d, and CL–T5 e at 1 Hz. For CL–M5 and CL–D9, N2/CO2 mixed gases were also used. The vertical broken line indicates 25 °C. f Changes in the tensile storage modulus (Eʹ), tensile loss modulus (Eʺ), and tanδ of CL–D9 in N2 and CO2 gases measured at 1 Hz and 25 °C.

Full size image

Enhancement of mechanical properties via reinforced with gases

Our elastomers are considerably toughened with CO2 gas owing to the formation of viscoelastic nanodomains. For example, a CL–D9 sheet could be very easily stabbed in air (Supplementary Movie 1). However, this sheet became remarkably toughened with CO2 gas (Supplementary Movie 2). To quantitatively estimate the toughness of the materials, fracture energies were measured via tearing tests using trouser-shaped test pieces40 (Fig. 4a). The fracture energies of the elastomers are dramatically enhanced when exposed to CO2 gas. Our elastomers exhibit considerable brittleness in N2 gas and in air, where their fracture energy values remained less than 60 J m−2. However, the fracture energies of elastomers considerably increase with CO2 gas (Fig. 4a). In particular, CL–D9 in CO2 gas (100%) exhibits a maximum value of ~1420 J m−2. This value is ~35 times as large as that measured in N2 gas. Furthermore, our materials are effectively toughened even at low concentrations of CO2 (Fig. 4a). For example, the fracture energy value of CL–D9 in the case of the N2/CO2 (90/10) mixed gas is ~970 J m−2, which is ~25 times that of CL–D9 in N2 gas. Although the effect of CO2 gas concentration is examined only for CL–M5 and CL–D9 in this study, a similar trend is expected for the other samples. Not only CO2 gas, but also other gases such as hydrogen chloride (HCl) and acetic anhydride (AcAh) gases are available for toughening our elastomers (Fig. 4a). Such HCl and AcAh gases form ammonium chloride and amide groups in our elastomers, respectively (Supplementary Fig. 7a). Such groups aggregate into nanodomains, functioning as viscoelastic nanofillers (Supplementary Fig. 7b and Supplementary Fig. 7c). Unlike ammonium carbamates, these groups are stable at room temperature in air.

Fig. 4: Effects of CO2 on the mechanical properties of elastomers.
figure 4

a Effects of different gases on the fracture energies of elastomers. For CL–M5 and CL–D9, N2/CO2 mixed gases were also used. Fracture energies were determined by tearing tests for trouser-shaped test pieces40. b Relationship between the fracture energy and shear loss modulus (Gʺ) of elastomers at 25 °C. c Fracture energy against GʺAint (Aint = 4πR12×ND). d Tensile testing of elastomers measured at 26 °C ± 1 °C in N2, air, and CO2 gases. The strain rate was 0.09 s−1. e Compression tests for CL–M5 measured at 26 °C ± 1 °C in N2 and CO2 gases. The strain rate was 0.03 s−1.

Full size image

The toughening behavior of the presented elastomers is affected by the changes in their viscoelasticity with the increasing concentrations of CO2 gas. Mullins demonstrated that the fracture energy value of elastomers exhibits a correlation with their Gʺ39. Similarly, the fracture energy values of the elastomers in our study exhibits a correlation with Gʺ (Fig. 4b), whereas they do not exhibit a correlation with Gʹ and tanδ (Supplementary Fig. 8). This result clearly indicates that the generation of viscoelastic nanodomains41 is the main cause for the toughening of our materials. We speculate that the fracture of our materials is prevented owing to energy dissipation via the viscoelastic behavior of the nanodomains. Furthermore, each elastomer exhibits different fracture-energy-versus-Gʺ relations. CL–Mx (green) and CL–Dx (pink) exhibit different slopes, while CL–Dx and CL–T5 (orange) show similar fracture-energy-versus-Gʺ relations. Moreover, the relation of CL–D9 with HCl deviates from the relation of CL–Dx with CO2 gas. Although the viscoelasticity of nanodomains mainly dominates the elastomer toughness, the morphology (size and ND) of nanodomains is also expected to affect the elastomer toughness. Figure 4c shows that the fracture energy values of all elastomers exhibit a single relation in terms of a value of Gʺ × Aint where Aint represents the total area of the interface between the nanodomains and matrix for a constant sample volume of 1000 nm3, which is calculated using 4πR12ND (nm2(10 nm)−3). This result demonstrates that the nanodomain interface plays an important role in the toughening of our elastomers in addition to the viscoelasticity of nanodomains. At the present stage, we speculate that the interface of nanodomains prevents the propagation of microcracks in the material, thereby functioning as a toughening mechanism.

Contrary to the poor stretchability and fracture stress in N2 gas, the modulus, stretchability, and fracture stress of our elastomers drastically increases with the increasing CO2 concentration owing to the formation of nanodomains that function as physical crosslinks (Fig. 4d). In case of CL–Mx samples, nanodomains are soft and do not function as effective physical crosslinks at room temperature because the relaxation temperature of nanodomains is considerably lower than room temperature (Fig. 3a, b). Therefore, the modulus of CL–M10 is not remarkably enhanced with CO2 gas while the stretchability is largely improved owing to the prevention of fracture. The toughening of our elastomers with CO2 gas is displayed during the compression tests. Upon compression, a block of CL–M5 is completely fractured in N2 gas (Fig. 4e, Supplementary Movie 3). However, a CL–M5 block reacted with CO2 gas does not fracture even at high compression (~20 MPa) despite the generation of some cracks (Fig. 4e, Supplementary Movie 4).

Role of detachment of ammonium carbamates from nanodomains in elastomer toughening

To understand the effects of nanodomains on material toughening, in situ SAXS measurements were conducted for CL–D9 under stretching. In this experiment, the specimen was stretched in the horizontal direction at a strain rate of 0.09 s−1, and two-dimensional (2D) SAXS patterns were recorded at an exposure time of 1 s. One-dimensional (1D) SAXS patterns were extracted from the stretching and transverse directions with a narrow angle range (±5°) in the 2D SAXS pattern (Fig. 5a). The structural information of the nanodomains for each direction was separately obtained from the simulations of the 1D SAXS patterns based on the Yarusso–Cooper model37 (Fig. 5b). In Fig. 5c, the R1, V, and ND values of the nanodomains were plotted as a function of strain for each direction. Interestingly, the thickness of nanodomains along the stretching direction slightly decreases upon stretching, as illustrated in Fig. 5d. This is attributed to the detachment of strongly stretched ammonium carbamates from the nanodomains. Moreover, the V value of nanodomains gradually decreases with stretching (Fig. 5c). Herein, oblate spheroid-shaped nanodomains with a short axis along the stretching direction are assumed to calculate V. The size and distribution of the oblate spheroid-shaped nanodomains along the transverse direction do not substantially change with stretching because the stacking of ammonium carbamates is rearranged along the transverse direction with stretching to reduce the steric constraint of stretched polymer chains attached to ammonium carbamates (Fig. 5d); this indicates that the stress is concentrated at the nanodomains, which promotes mechanical reinforcement. Furthermore, the detachment of strongly stretched ammonium carbamates from nanodomains plays an important role in dissipating the excess stress locally concentrated in the stretched material, thereby preventing the fracture of our material under high stretching42. A similar toughening mechanism for covalently crosslinked networks via detachment of coexistent weak crosslinks such as the ionic one has been frequently observed43.

Fig. 5: In situ SAXS analysis of CL–D9 with stretching.
figure 5

a 2D SAXS patterns of CL–D9 in CO2 gas at various nominal strains (ε). The specimen was stretched at 0.09 s−1 and 25 °C. b 1D SAXS profiles extracted from the stretching (green) and transverse (orange) directions (±5°). The experimental patterns were simulated using Yarusso–Cooper model37. c Nominal stress, R1, V, and ND of CL–D9 in CO2 gas against ε. Data for the unstretched state are indicated using the magenta symbols. For R1 and ND, the data for the stretching and transverse directions are indicated using solid and open symbols, respectively. d Schematic illustration for the detachment of ammonium carbamate attached to highly stretched polymer chains from nanodomains.

Full size image

Toughening of polyethylene with CO2 gas

To demonstrate the applicability of our method to other polymers, the mechanical reinforcement of polyethylene (PE) with CO2 gas was examined. PE is one of the most important polymers in our modern society. Vast amounts of PE are used daily as films, sheets, plates, tubes, and moldings. Therefore, the mechanical reinforcement of PE with specific gases can exemplify the remarkable impact our findings on potential applications and technologies. In this study, we prepared a TAEA-modified PE film (E–T), as shown in Fig. 6a, with a thickness of ~200 μm. E–T rapidly traps CO2 gas molecules via the formation of ammonium carbamates (Supplementary Fig. 9a and Supplementary Fig. 9b). The trapped CO2 is quickly released when the specimen is heated (Supplementary Fig. 9c). The generated ammonium carbamates microphase-separate into nanodomains in the PE matrix (Fig. 6b). The V and ND values of nanodomains determined via the Yarusso–Cooper model simulation35 of the SAXS pattern are 3.4 (nm3) and 22 ((10 nm)−3), respectively. The formation of nanodomains with CO2 gas mechanically reinforces and toughens E–T. The Eʹ and Eʺ of E–T increase more than twofold with CO2 gas (Fig. 6c). Owing to the mechanical reinforcement with CO2 gas, the fracture stress of E–T in CO2 gas (~14 MPa) is much higher than that in N2 ( ~ 10 MPa) during the tensile test (Fig. 6d). The fracture energy of E–T increases from 3500 J m−2 in N2 to 4800 J m−2 in CO2 gas (Supplementary Fig. 9d). E–T is also reactive with HCl, AcAh, and acetic acid (AcOH) gases (Supplementary Fig. 10). Although AcOH molecules are released from E–T upon heating, E–T/HCl and E–T/AcAh are stable even at 110 °C.

Fig. 6: Mechanical reinforcement of polyethylene with CO2 gas.
figure 6

a Chemical structure of E–T. b Experimental and simulated SAXS patterns of E–T. SAXS patterns with and without CO2 are shown as magenta and black plots, respectively. The Yarusso–Cooper model was used for the simulations37. c Effects of CO2 gas on Eʹ, Eʺ, and tanδ of E–T measured at 1 Hz and 25 °C. d Tensile behaviors of E–T measured at 26 °C ± 1 °C in N2 and CO2 gases. The strain rate was 0.09 s−1.

Full size image

Conclusion

Herein, we developed PDMS elastomers that are mechanically reinforced and toughened with CO2, HCl, and AcAh gases. The key aspect of the material design involves programming chemical groups to form physical crosslinks with specific gases. Amine groups were examined as the simplest example of CO2-responsive chemical groups. The amines form ammonium carbamates with CO2 gas that physically crosslink polymer chains and mechanically reinforce the PDMS elastomers. Further, nanodomains with ammonium carbamates that function as viscoelastic nanofillers dissipate a large amount of energy upon material deformation and thus lead to mechanical toughening. Furthermore, this design can be applied to various polymers, such as PE. These CO2-responsive materials exhibit considerable potential for practical applications, including but not limited to sensors, tunable adhesives, shape-memory materials, actuators, and strength-tunable materials. These materials provide a new path toward the effective utilization of CO2 gas.

Methods

Details of sample preparation are described in the Supplementary Information.

Preparation of CL–Mx, CL–Dx, and CL–T5

CL–Mx and CL–Dx were prepared via the crosslinking of PDMS–Mx and PDMS–Dx with 1,4-butanediol diglycidyl ether (BDE; concentration = 8.5 × 10−5 mol g−1). The Mw and PDI values of PDMS–Mx and PDMS–Dx are presented in Table 1. The amine concentrations of PDMS–M5, PDMS–M10, PDMS–D5, and PDMS–D9 were 5.2 mol% (6.8 × 10−4 mol g−1), 10 mol% (1.3 × 10−3 mol g−1), 4.7 mol% (6.2 × 10−4 mol g−1), and 9.2 mol% (1.1 × 10−3 mol g−1), respectively. Each polymer (7.94 g) was dissolved in dry THF (10 mL) with BDE (0.138 g) and poured into a Teflon petri dish. The mixture with PDMS–Mx was dried at 50 °C for 17 h. The obtained CL–Mx sheet was further dried at 90 °C for 48 h in a vacuum. In the case of the mixture with PDMS–Dx, the mixture was dried at 35 °C for 39 h and then at 60 °C for 7 h. The mixture was further heated at 90 °C for 14 h in a vacuum to obtain a CL–Dx sheet. CL–T5 was obtained via the addition of tris(2-aminoethyl)amine (TAEA) to epoxy-modified PDMS (KF-1001, Shin-Etsu Chemical Co., Ltd.). The Mw and PDI values of KF-1001 were 66700 and 1.89, respectively. KF-1001 (9.70 g) and TAEA (0.300 g) were dissolved in dry chloroform (80 mL), and the solution was stirred at 90 °C for 4 h and poured into a Teflon petri dish. The solution was dried at 60 °C for 48 h, and the obtained CL–T5 sheet was further dried at 90 °C for 24 h in a vacuum. The thicknesses of the CL–Mx, CL–Dx, and CL–T5 sheets were ~1 mm.

Reaction of CL–D9 with HCl and AcAh gases

CL–D9 sheets and a glass petri dish containing HCl aqueous solution (35%) or AcAh were placed in a glass container. The glass container was kept at 40 °C for 24 h. The reacted sheets were vacuum dried at 35 °C for 24 h.

Preparations of E–T

Poly(ethylene-co-glycidyl methacrylate) (PEGMA, Sigma-Aldrich Co. LLC, 0.790 g) and TAEA (0.0605 g) were dissolved in dry toluene (30 mL) at 91 °C, and the solution was stirred for 5 h. The solution was then poured into a Teflon petri dish and slowly dried at 103 °C for 24 h. The obtained E–T film with a thickness of ~0.2 mm was further dried at 100 °C for 24 h in a vacuum. The obtained E–T film was transparent.

Measurements

Before the measurements, the sample sheets were further vacuum-dried at room temperature for at least 24 h and carefully placed into the instruments to avoid moisture absorption.

A 400 MHz JEOL-ECS400 spectrometer was used for 1H-NMR measurements. CDCl3 containing tetramethylsilane was used as a solvent. GPC measurements were conducted using an EXTREMA HPLC system (JASCO Co.) equipped with a polystyrene gel column (Shodex GPC LF-804). THF was used as the eluent at 40 °C. PDMS standards (Scientific Polymer Products Inc.) were used for calibration.

DSC performed using a DSC7020 instrument (SII). The samples were heated from −150 °C to 100 °C at a rate of 10 °C min−1. Thermogravity measurements were performed using STA200 (HITACHI) at a gas flow of 100 mL min−1. N2, CO2, and N2/CO2 mixed gases were used. The concentration of the N2/CO2 mixed gas was controlled using a gas mixer CUBE GM3 (FCON Co., Ltd.).

An FT/IR-6600 spectrometer (JASCO Co.) equipped with a triglycine sulfate detector was used for FT–IR measurements at a resolution of 4 cm−1 under N2 or CO2 gas flow aof 1.5 L min−1. For the transmittance mode in case of E–T, a thin sample film was used to ensure that the band absorbance were in the Lambert–Beer law range. The attenuated total reflection mode was performed for CL–Mx, CL–Dx, and CL–T5 using an ATR PRO ONE (JASCO) with an ATR correction program.

Synchrotron SAXS measurements were performed using the BL-15A2 beamline at the Photon Factory of the High Energy Accelerator Research Organization (KEK) in Tsukuba, Japan. The X-ray wavelength was 0.092 nm, and PILATUS 2 M was used as a detector. Stearic acid and silver behenate were used as calibration standards. Samples fully exposed to CO2 gas were quickly installed into the beamline and measured in air. For in situ SAXS measurements with stretching, a custom-made tensile device (AcroEdge Co., Ltd.) with a sample chamber was installed in the beamline and the specimen was stretched at 0.09 s−1. CO2 gas was flown into the sample chamber at 3.0 L min−1. Prior to the measurement, the specimen was exposed to CO2 gas for 1.5 h in the sample chamber.

A rheometer (MCR302, Anton Paar) with 4-mm-diameter plates was used for linear viscoelasticity measurements. A shear strain of 0.1% was applied at 1 Hz. A temperature sweep test was conducted from –30 °C to 150 °C using a Peltier cooling system at a heating rate of 3 °C min−1. A flow rate of 3.0 L min−1 was used for each dry gas. DMA measurements were performed using an MCR702e instrument (Anton Paar). Measurements were performed in the tensile mode at 1 Hz with a tensile strain of 0.1% and at constant temperature (25 °C). The sample dimensions were 20 mm × 5.0 mm × 1.0 mm and 20 mm × 5.0 mm × 0.2 mm for CL–D9 and E–T, respectively. Dry N2 and CO2 gases were flown into the sample chamber at a rate of 3.0 L min−1.

Tensile and compression measurements were conducted using an AND force tester MCT-2150 with a custom-made sample chamber at 26 °C ± 1 °C. N2 or CO2 gases were flown into the chamber at 3.0 L min−1. Sample pieces for tensile tests were shaped into JIS 7 dumbbell-shaped tensile bars. Prior to the measurement, the specimen was exposed to dry N2 or CO2 gas for more than 2 h in the sample chamber. The initial gauge length was set to ~20 mm. The sample pieces were stretched at 100 mm min−1. For compression tests, a sample block (5 mm × 5 mm × 5 mm) was compressed at 100 mm min−1. Tearing tests were performed using an AND force tester MCT-2150 at 26 °C ± 1 °C. Trouser-shaped test pieces were torn at 100 mm min−1. Prior to the measurement, the specimen was exposed to dry N2 or CO2 gas for more than 2 h in the sample chamber. The fracture energy was calculated following the literature40.