Emulsion Gels as Precursors for Porous Silicones and All-Polymer Composites—A Proof of Concept Based on Siloxane Stabilizers

In spite of its versatility, the emulsion templating method is rather uncommon for the preparation of porous silicones. In this contribution, two siloxane-containing stabilizers, designed to be soluble in polar (water) and non-polar (toluene) solvents, respectively, were used in low concentrations to produce stable emulsions, wherein polysiloxane gels were obtained by UV-photoinitiated thiol-ene click cross-linking. The stabilizers exhibited negative interfacial tension, as measured by Wilhelmy plate tensiometry. The emulsion gels evolved into porous silicones (xerogels), with tunable morphology and properties. According to TEM and SEM investigations, the emulsion template was preserved in the final materials. Several parameters (e.g., the structure of the polysiloxane precursors, composition of the emulsion gels, nature of the continuous phase, cross-linking conditions, or additives) can be varied in order to obtain porous elastic materials with desired properties, such as Janus membranes, absorbent monoliths, all-polymer porous composites, or silicone-swollen gels. The feasibility of these types of materials was tested, and exemplary porous silicones were briefly characterized by contact angle measurements, mechanical testing, and absorption tests. The proposed method is simple, fast, and economic, uses very little amounts of stabilizers, and can be adjusted as a green technique. In this contribution, all the silicon-based materials with a convenient design were prepared in house.


Introduction
Emulsion templating is one of the most important methods for producing porous materials. It generally involves reactions that take place in the external (continuous) phase, and lead to monoliths. A lot of techniques and experimental setups, as well as various chemical reactions, have been demonstrated for production of a wide range of porous materials, including polymers, organic-inorganic hybrids or composites, inorganic structures, or porous carbon materials [1]. Several types of emulsions may form by proper choice of the system components and concentration, and these can be polymerized to form porous materials with tunable morphology. For example, complex emulsion structures have been obtained [2] by using osmotically driven water-oil phase separation.
In particular, emulsion gels are two-phase systems in which the continuous phase is a gel, and serve as templates for highly porous materials. In most cases, the disperse phase consists in oil droplets dispersed in a hydrogel (or a biopolymer gel), and these are widely used in food products [3]. The reverse combinations of organic gels with high water content as the internal phase are precursors for low density, porous materials [4]. As in the general case of emulsions, the preparation of emulsion gels involves the use of a stabilizer, which is We have reported several new types of siloxane-based amphiphilic compounds, and among them, pyridyl-modified siloxanes showed remarkable potential. They proved to be surface-active in water or in organic solvents, depending on the length of the dimethylsiloxane segment, and could manifest phase transfer ability, besides other properties [27].
Herein, we explore the potential of two pyridyl-modified siloxane stabilizers, specially designed to be soluble in water or toluene, for emulsion templating of silicone gels, aiming to produce porous monoliths or films, with a minimum amount of additives. These surfactants showed negative interfacial tension and thus can promote quasi-spontaneous emulsification. The photo-induced cross-linking of vinyl-polysiloxanes offers a fast, clean, and simple alternative for the preparation of stable, porous silicone materials, and is also a tool for better understanding the complexity of the water/toluene emulsion system containing the pair of in-house prepared surfactants. The main purpose of this research was to test the feasibility of the emulsion templating method, adapted for dilute solutions of these surfactants and using thiol-ene click photo-addition cross-linking, in preparation of porous silicones. Several variations, aiming to tune the characteristics of the final materials, were identified. For example, the transient gel phase can be preserved for a long time, especially when the toluene solvent is replaced by less volatile octamethylcyclotetrasiloxane (D4). The addition of water-soluble polymers allows one-step preparation of composite porous materials. Moreover, many parameters can be tuned to obtain optimized materials, according to envisaged applications.
The stabilizers (D and T) were synthesized as previously described [27] by the addition of 4-aminopyridine to epoxy-terminated siloxane oligomers.

Methods
Surface, interfacial, and continuous surface tension measurements were done with a Sigma 700 automated tensiometer from KSV, using the Wilhelmy plate method. For interfacial tension measurements, special care was taken for accurate detection of the interface, including staining of the water layer with methylene blue.
The TEM analysis of the samples was performed using the Hitachi HT-7700 equipment operated at 100 kV accelerating voltage. This instrument is dedicated to the study of soft materials and allows observation of details in high magnification and good contrast conditions. Samples of the studied emulsions were drop cast onto TEM grids (300 mesh, copper grids, Ted Pella). Excess of the liquid was blotted away and the samples were let to dry overnight in a vacuum oven at 70 • C. Inspection of the samples was conducted after cooling the samples chamber at −170 • C.
For SEM investigations, depending on the sample, two microscopes were used: Quanta 200 (FEI Company, Brno, Czech Republic), using a LFD (Large Field Detector), working in low vacuum mode at 20 kV without any coating, and Verios G4 UC (Thermo Scientific, Waltham, MA, USA), using a CBS (Concentric Backscatter Detector. Brno, Czech Republic) Gels 2022, 8,377 4 of 17 at 5 kV, in which case the samples were coated with 6 nm of platinum with a Leica EM ACE200 Sputter coater to provide electrical conductivity and to prevent charge buildup during exposure to the electron beam. All the materials were analyzed in cross-section obtained by breaking or cutting with scissors.
Fourier transform infrared (FT-IR) spectra were registered on a Bruker Vertex 70 Spectrometer in transmittance mode, in the 400-4000 cm −1 spectral range, with a resolution of 4 cm −1 and accumulation of 32 scans.
The contact angle was measured with a DSA25 Drop Shape Analyzer (Kruss), in the Sessile Drop mode, using water droplets of 1 µL. The values of the contact angle are the mean of a minimum of 3 determinations.
Mechanical measurements were made using an Instron 3365, two-column universal mechanical testing machine. The stress-strain measurements were made at an extension rate of 50 mm/min, at room temperature (laboratory conditions). All samples were measured two times and the average of the obtained values was reported. The compression tests were done for monolith samples at 10 mm/min compression rate and different compressive strains (10%, 30%, and 50%).
Swelling experiments were done by immersing pieces of porous silicone materials in the solvents for 30 min and draining and weighing the swollen samples. The absorption capacity (g/g) was calculated with the formula: (m − m 0 )/m 0 , where m 0 and m represent the weight of the sample before and after absorption of the solvent.

Preparation of Emulsion Gels and Xerogels
In preliminary experiments (PE1-3), in a dilute toluene solution of T, a dimethylmethylvinyl siloxane copolymer with 8.5% vinyl groups (copolymer concentration 20% w/v), the cross-linker (15% w/w reported to the copolymer) and photoinitiator DMPA (2.4%) were dissolved. Distilled water was added (10% or 20% reported to toluene) with or without PEG dissolved, and the mixture was processed with a SilentCrusher Homogenizer (Heidolph) at 14,000 rpm. Then, the mixture was either poured in a Teflon vessel or kept in an open vial, and irradiated with a laboratory lamp (Herolab) at 365 nm for 10-15 min. The materials were allowed to dry freely in an ambient atmosphere.
In the optimized experiments (EG1-10), 1% toluene solution of T (typically 2.4 mL) was mixed with 1% water solution of D (according to Table 1), by manual shaking and sonication for 30 s in an ultrasonic cleaning bath (Elmasonic P, Elma, Singen, Germany, at 37 kHz and 90% power). Then, the silicone precursor (0.5 g) and the cross-linker SHCL (in equimolar ratio reported to vinyl groups) were added to the formed emulsion and dissolved by manual shaking and sonication for 30 s in the cleaning bath. The photoinitiator (DMPA, 12 mg) was added, and after gentle agitation, the mixture was submitted to UV irradiation with a 400 W Phillips UV-A Lamp, for 3-5 min. The materials were rinsed with water several times to remove excess SHCL and other impurities, and allowed to dry freely in ambient atmosphere, or, alternatively, were supplementary dried in a vacuum oven at 70 • C for 24 h. The composition of all the samples is described in Table 1. In all cases, the ratio between the vinyl-silicone, the cross-linker, and the catalyst was kept constant. For sample EG9, the stabilizer T was first dissolved in D4 (instead of toluene), with the aid of a drop of chloroform and then the entire protocol described above was followed.

Surface Properties of the Stabilizers
Two siloxane compounds bearing pyridyl terminal groups and having one (disiloxane, D) or 13 (telechelic, T) siloxane units were prepared according to our previous report [27] (Scheme 1). These compounds, with similar structure, are soluble in water and nonpolar organic solvents (toluene), respectively. Surface activity in aqueous solution was demonstrated for D. A very close homologue of T showed remarkable phase transfer ability of organic molecules from water into chloroform by the formation of water-in-chloroform (micro)emulsions [27].

Surface Properties of the Stabilizers
Two siloxane compounds bearing pyridyl terminal groups and having one (disiloxane, D) or 13 (telechelic, T) siloxane units were prepared according to our previous report [27] (Scheme 1). These compounds, with similar structure, are soluble in water and nonpolar organic solvents (toluene), respectively. Surface activity in aqueous solution was demonstrated for D. A very close homologue of T showed remarkable phase transfer ability of organic molecules from water into chloroform by the formation of water-in-chloroform (micro)emulsions [27]. In this contribution, we further explored the surface properties of these modified siloxanes (Table 2). They have very different HLB values [27] that are 12 for D and 5 for T, which indicate applications in direct (o/w) or reverse (w/o) emulsions, respectively. First, their interfacial tension against complementary liquids was measured by the Wilhelmy plate method. By this method, we found small negative values, i.e.,: In this contribution, we further explored the surface properties of these modified siloxanes ( Table 2). They have very different HLB values [27] that are 12 for D and 5 for T, which indicate applications in direct (o/w) or reverse (w/o) emulsions, respectively. First, their interfacial tension against complementary liquids was measured by the Wilhelmy plate method. By this method, we found small negative values, i.e.,: −4.2 mN/m for an aqueous solution of D versus toluene, and −2.95 mN/m for a toluene solution of T versus water. The plausible explanation of the observed phenomenon is the strong amphiphilic character of the siloxane surfactants, which promotes spontaneous distribution of these molecules within the interfacial region. Repeated measurements with care for accurate detection of the interface, including staining of the water layer with methylene blue were done. In addition, interfacial tension was also measured for D with a continuous surface tension module, providing the appropriate parameters setup. It was observed that the values obtained for different concentrations of the D solutions were close to those measured in the usual surface/interfacial tension mode ( Figure S1). The surface tension of the emulsions formed with each surfactant was also measured ( Table 2). The mean values for the toluene-in-water (o/w) emulsions in both cases, as well as for the water-in-toluene (w/o) phase in the case of T surfactant (Figure 1), were very close, around 26 mN/m. However, the standard deviation was rather high for the emulsion formed by D, which indicates poorer stability.
character of the siloxane surfactants, which promotes spontaneous distribution of these molecules within the interfacial region. Repeated measurements with care for accurate detection of the interface, including staining of the water layer with methylene blue were done. In addition, interfacial tension was also measured for D with a continuous surface tension module, providing the appropriate parameters setup. It was observed that the values obtained for different concentrations of the D solutions were close to those measured in the usual surface/interfacial tension mode ( Figure S1). The surface tension of the emulsions formed with each surfactant was also measured ( Table 2). The mean values for the toluene-in-water (o/w) emulsions in both cases, as well as for the water-in-toluene (w/o) phase in the case of T surfactant (Figure 1), were very close, around 26 mN/m. However, the standard deviation was rather high for the emulsion formed by D, which indicates poorer stability.

Morphology of the Emulsions and Templated Xerogels
Negative interfacial tension values are correlated with the spontaneous formation of microemulsions. In our case, formation of an emulsion at gentle manual shaking was indeed observed. Depending on the water content and on the nature and level of the external energy, the emulsification process occurred differently. First we focused our attention on the organic-soluble T surfactant in contact with water. Under weak agitation, two types of emulsions formed (Figure 1), which is in line with the previous observation of phase transfer ability.
The TEM investigations revealed completely different morphology in the two emulsions (Figure 1b,c). The w/o macroemulsion has large dimensions of the internal phase domains, within 400 nm-1 mm range, while the o/w emulsion presents particles of 100-200 nm assembled in a composite morphology. Interestingly, a completely different situation occurred when a high-energy ultrasonication was applied, using an ultrasonic processor (VC 505) ( Figure S2). This time both emulsions formed, but the o/w emulsion was much more stable. Moreover, its milky aspect indicates a macroemulsion.
To prepare the porous silicones, a UV-photoinitiated cross-linking reaction was used, which is a thiol-ene addition between a vinyl-containing polysiloxane and a trifunctional thiol (SHCL), as depicted in Scheme 2. This reaction is well known in organic chemistry

Morphology of the Emulsions and Templated Xerogels
Negative interfacial tension values are correlated with the spontaneous formation of microemulsions. In our case, formation of an emulsion at gentle manual shaking was indeed observed. Depending on the water content and on the nature and level of the external energy, the emulsification process occurred differently. First we focused our attention on the organic-soluble T surfactant in contact with water. Under weak agitation, two types of emulsions formed (Figure 1), which is in line with the previous observation of phase transfer ability.
The TEM investigations revealed completely different morphology in the two emulsions (Figure 1b,c). The w/o macroemulsion has large dimensions of the internal phase domains, within 400 nm-1 mm range, while the o/w emulsion presents particles of 100-200 nm assembled in a composite morphology. Interestingly, a completely different situation occurred when a high-energy ultrasonication was applied, using an ultrasonic processor (VC 505) ( Figure S2). This time both emulsions formed, but the o/w emulsion was much more stable. Moreover, its milky aspect indicates a macroemulsion.
To prepare the porous silicones, a UV-photoinitiated cross-linking reaction was used, which is a thiol-ene addition between a vinyl-containing polysiloxane and a trifunctional thiol (SHCL), as depicted in Scheme 2. This reaction is well known in organic chemistry [29,30], and also employed in silicones chemistry [23,[31][32][33], as a quick (click), clean process, allowing functionalization or cross-linking of polysiloxanes.  [29,30], and also employed in silicones chemistry [23,[31][32][33], as a quick (click), clean cess, allowing functionalization or cross-linking of polysiloxanes. The completion of the cross-linking reaction can be observed in the FT-IR spect of material EG6, compared with the starting vinyl-silicone Co-8.5 and the SHCL (Fi 2). In the spectrum of the initial dimethylmethylvinylpolysiloxane (Co-8.5), the absorp bands corresponding to vinyl groups are located at 1600 and 3055 cm −1 , while the ba at 1263 cm −1 , 1097 cm −1 , 1024 cm −1 , and 800 cm −1 are characteristic of polysiloxanes. thiol absorption band in SHCL appears at 2569 cm −1 , and the band at 1732 cm −1 is assig to the ester groups. In the cross-linked material, neither the vinyl nor the thiol bands present anymore, and additional bands coming from the surfactants are visible at cm −1 and 3433 cm −1 (secondary amine and adsorbed water) [27]. The polysiloxane si ture remained unchanged, while the ester band suffered a slight shift due to change H-bonding. The completion of the cross-linking reaction can be observed in the FT-IR spectrum of material EG6, compared with the starting vinyl-silicone Co-8.5 and the SHCL (Figure 2). In the spectrum of the initial dimethylmethylvinylpolysiloxane (Co-8.5), the absorption bands corresponding to vinyl groups are located at 1600 and 3055 cm −1 , while the bands at 1263 cm −1 , 1097 cm −1 , 1024 cm −1 , and 800 cm −1 are characteristic of polysiloxanes. The thiol absorption band in SHCL appears at 2569 cm −1 , and the band at 1732 cm −1 is assigned to the ester groups. In the cross-linked material, neither the vinyl nor the thiol bands are present anymore, and additional bands coming from the surfactants are visible at 1657 cm −1 and 3433 cm −1 (secondary amine and adsorbed water) [27]. The polysiloxane signature remained unchanged, while the ester band suffered a slight shift due to changes in H-bonding.
When the dilute toluene solution of T was used as a continuous phase, having a vinylcontaining siloxane copolymer dissolved, together with an SH-containing cross-linker and the photoinitiator, a low amount of water was introduced as the internal phase (10-20%). The mixture was submitted to UV irradiation and the obtained gels ( Figure S3) were dried in ambient conditions before SEM analysis. As can be observed in Figure 3, porous materials were obtained, presenting a "double emulsion structure" [2] with both large (micron size) pores distributed on a "sea island" pattern and small (submicron) ones, as well as combinations of these (Figure 3c). A similar composition, wherein PEG was dissolved in the water phase, gave a polymer-polymer composite. Its SEM images (Figure 3d,e) clearly show micron size PEG particles (instead of the large pores in the previous example) and small pores (Figure 3d). The observed morphology of this composite material confirms the fact that the initial emulsion served as a template, since PEG particles are indeed formed in the w/o domains. When the dilute toluene solution of T was used as a continuous phase, having a vinylcontaining siloxane copolymer dissolved, together with an SH-containing cross-linker and the photoinitiator, a low amount of water was introduced as the internal phase (10-20%). The mixture was submitted to UV irradiation and the obtained gels ( Figure S3) were dried in ambient conditions before SEM analysis. As can be observed in Figure 3, porous materials were obtained, presenting a "double emulsion structure" [2] with both large (micron size) pores distributed on a "sea island" pattern and small (submicron) ones, as well as combinations of these ( Figure 3c). A similar composition, wherein PEG was dissolved in the water phase, gave a polymer-polymer composite. Its SEM images ( Figure  3d,e) clearly show micron size PEG particles (instead of the large pores in the previous example) and small pores (Figure 3d). The observed morphology of this composite material confirms the fact that the initial emulsion served as a template, since PEG particles are indeed formed in the w/o domains.  In the next step, the formulation was changed, by addition of D (1%) in the aqueous phase as co-surfactant. This allowed us to obtain stable emulsions just by gentle manual mixing, while short time sonication in an ultrasound bath additionally increased the stability and homogeneous aspect. The complex composition of the as-obtained emulsion can be visualized in Figure 4.  In the next step, the formulation was changed, by addition of D (1%) in the aqueous phase as co-surfactant. This allowed us to obtain stable emulsions just by gentle manual mixing, while short time sonication in an ultrasound bath additionally increased the stability and homogeneous aspect. The complex composition of the as-obtained emulsion can be visualized in Figure 4. Several types of disperse objects were observed by TEM, corresponding to double emulsions (w/o and o/w), as well as to triple emulsions (o/w/o and w/o/w), and overall a large polydispersity of size and shape. Nevertheless, the emulsion was stable enough, as to preserve the morphology much longer than the timeframe of the UV cross-linking of the polysiloxane mixture. Since both toluene-based and water-based emulsion patterns were detected, a logical assumption is that a bicontinuous phase forms during sonication, allowing coexistence of all these aggregates.
With the surfactants pair, silicone gels with 33-50% water content could be easily obtained ( Figure 5), which were processed as monoliths or thick films after drying in ambient conditions. Several types of disperse objects were observed by TEM, corresponding to double emulsions (w/o and o/w), as well as to triple emulsions (o/w/o and w/o/w), and overall a large polydispersity of size and shape. Nevertheless, the emulsion was stable enough, as to preserve the morphology much longer than the timeframe of the UV cross-linking of the polysiloxane mixture. Since both toluene-based and water-based emulsion patterns were detected, a logical assumption is that a bicontinuous phase forms during sonication, allowing coexistence of all these aggregates.
With the surfactants pair, silicone gels with 33-50% water content could be easily obtained ( Figure 5), which were processed as monoliths or thick films after drying in ambient conditions. For f = 0.25, the multiple emulsion structure [2] was observed by SEM in the xerogel (Figure 6a,b). Additional elements were a thin smooth layer on the surface, which is due to evaporation of toluene during heat-generating UV irradiation, and the wrinkled walls which result from drying of the silicone gel phase (Figure 6b). For f = 0.25, the multiple emulsion structure [2] was observed by SEM in the xerogel (Figure 6a,b). Additional elements were a thin smooth layer on the surface, which is due to evaporation of toluene during heat-generating UV irradiation, and the wrinkled walls which result from drying of the silicone gel phase (Figure 6b). For compositions with f = 0.33 and f = 0.5, gradient emulsion structure [2] was frozen in the xerogels, with smaller pores on the surface and much larger pores on the bottom of the films or monoliths, which still preserve the multiple emulsion aspect (Figure 6c). In such a case, Janus membranes can be obtained in a one-step click reaction, since they present two surfaces with different morphology and properties. All the described materials were obtained based on a high molecular mass siloxane copolymer, which plays the role  For compositions with f = 0.33 and f = 0.5, gradient emulsion structure [2] was frozen in the xerogels, with smaller pores on the surface and much larger pores on the bottom of the films or monoliths, which still preserve the multiple emulsion aspect (Figure 6c). In such a case, Janus membranes can be obtained in a one-step click reaction, since they present two surfaces with different morphology and properties. All the described materials were obtained based on a high molecular mass siloxane copolymer, which plays the role of a gelator, since it dissolves in toluene and ensures a relatively high viscosity, thus facilitating the stability of water droplets [5]. When this copolymer was replaced with a similar one with lower content of vinyl groups, and its proportion in the emulsion gel was diminished to half (sample EG10M), higher porosity was observed in the resulted monolith (Figure 6d). In another experiment, a long telechelic oligomer of M n~1 0,000 g/mol and vinyl end-groups was used instead of the copolymer. Thus, a less viscous mixture was obtained, and a looser cross-linking by chain ends. In order to increase the viscosity, a lower amount of toluene was used, and thus the proportion of water internal phase increased to f = 0.8. However, in this case (sample EG7), a phase separation occurred during UV cross-linking, leading to a thin film and a latex. The film had a wrinkled appearance in SEM, with some micron-size pores and small pores, preserving to a certain extent the morphology discussed above (Figure 6e,f). The latex evolved in a fragile material formed by cross-linked polymer particles, bearing nanopores (Figure 7a,b). The same complex morphology could be observed by TEM of a latex sample placed on a grid from the initial dispersion (Figure 7c,d). Water-in-oil droplets coexist with the latex particles and collapse on their surface by Ostwald ripening, creating small holes after water evaporation. When the water phase contained PEG in 20% proportion reported to the siloxane matrix, submicron PEG particles were observed within the entire cross-section of the film, including in the pores (Figure 8). This is an easy and convenient way of obtaining allpolymer porous composites, from two incompatible polymers. The morphology of the composites (although not extensively studied yet) was different in the two examples shown in Figure 3e (prepared without co-surfactant) and in Figure 8. In the former case, the PEG particles were roughly around 1 micrometer in diameter, while in the latter these were around 200 nm. The difference comes from the particularities of each emulsion in the presence of polysiloxane. Thus, PEG was "filling" the large pores in the first example, and the small ones in the second, while the complementary set of pores remained open. Therefore, using or not using the co-surfactant is one way of controlling the composites' morphology. In both cases, the PEG particles were relatively uniformly dispersed within the film thickness.  When the water phase contained PEG in 20% proportion reported to the siloxane matrix, submicron PEG particles were observed within the entire cross-section of the film, including in the pores (Figure 8). This is an easy and convenient way of obtaining all-polymer porous composites, from two incompatible polymers. The morphology of the composites (although not extensively studied yet) was different in the two examples shown in Figure 3e (prepared without co-surfactant) and in Figure 8. In the former case, the PEG particles were roughly around 1 micrometer in diameter, while in the latter these were around 200 nm. The difference comes from the particularities of each emulsion in the presence of polysiloxane. Thus, PEG was "filling" the large pores in the first example, and the small ones in the second, while the complementary set of pores remained open. Therefore, using or not using the co-surfactant is one way of controlling the composites' morphology. In both cases, the PEG particles were relatively uniformly dispersed within the film thickness.   Taking into account the information from electron microscopy observations, we proposed a schematic representation of the entire process of pore formation (Figure 9). The initial system contains surfactant T dissolved in toluene (a). When water containing cosurfactant D is added, w/o emulsion starts to form spontaneously until an excess of aqueous phase accumulates and the micelles of D entrap toluene, forming o/w emulsion. According to TEM (Figure 4), several types of emulsion droplets coexist, so it is supposed that a bicontinuous phase is present, allowing formation of triple emulsions as well (b). At this point, the silicone mixture is added, i.e., the vinyl copolymer, the SH cross-linker and the photoinitiator (Scheme 2). Due to increased hydrophobicity of the silicone (which is basically a slightly modified PDMS, known for its very low surface tension), the water from the bicontinuous phase is forced into w/o droplets, and the proportion of o/w/o increases, wherein we denoted o as the "toluene + silicone" phase. These changes are suggested by the absence of interconnected large pores or cracks in the final materials (which would indicate a bicontinuous emulsion gel template) according to SEM. Next, the UV-initiated cross-linking occurs within a few minutes (c). During gelation in steady conditions, the larger droplets, containing more water, tend to accumulate at the bottom of the vessel, templating larger pores in the final materials with medium or high internal phase (d).
It is worth mentioning that the pore diameters were largely polydisperse in all the samples, as was estimated in Figure 6, with "large" pores of tens of microns and "smaller" pores, partly distributed inside the larger ones, due to particularities of these complex emulsions. We suppose that adjustments for more uniform pore size are possible by the level of the energy input, which would control the size of the droplets in the emulsion template. In the meantime, the polydispersity may be of interest for certain applications, as reported in the case of materials with gradient pore size for scaffolds and extracellular matrix [34,35]. Similar size and morphology of the pores and some adjustments in pore distribution by nature and concentration of the surfactants were reported by Abshirini et al. [36], for porous silicones obtained by hydrosilylation in emulsion templating. Even larger pores were reported for porous silicones (sponges or foams) obtained with solid NaCl [37] or supercritical CO 2 [38] as templates.
gested by the absence of interconnected large pores or cracks in the final materials (which would indicate a bicontinuous emulsion gel template) according to SEM. Next, the UVinitiated cross-linking occurs within a few minutes (c). During gelation in steady conditions, the larger droplets, containing more water, tend to accumulate at the bottom of the vessel, templating larger pores in the final materials with medium or high internal phase (d). It is worth mentioning that the pore diameters were largely polydisperse in all the samples, as was estimated in Figure 6, with "large" pores of tens of microns and "smaller" pores, partly distributed inside the larger ones, due to particularities of these complex emulsions. We suppose that adjustments for more uniform pore size are possible by the level of the energy input, which would control the size of the droplets in the emulsion template. In the meantime, the polydispersity may be of interest for certain applications, as reported in the case of materials with gradient pore size for scaffolds and extracellular matrix [34,35]. Similar size and morphology of the pores and some adjustments in pore distribution by nature and concentration of the surfactants were reported by Abshirini et al. [36], for porous silicones obtained by hydrosilylation in emulsion templating. Even larger pores were reported for porous silicones (sponges or foams) obtained with solid NaCl [37] or supercritical CO2 [38] as templates.

Brief Characterization of the Obtained Xerogels
Porous silicone materials are very interesting for a number of applications, including sensors, filtration membranes, delivery of active principles, or cell growth, to mention only a few. By using the concept developed in this work, very small amounts of siliconecontaining surfactants are used in a simple method that allows various adjustments in order to tune the final properties according to the needs. For example, the cross-linking density can be adjusted by modifying the content of vinyl groups in the polysiloxane precursor, while the viscosity of the system depends (for a fixed phase ratio) on the molecular mass of the polysiloxane precursor. The amount of internal phase (water) influences the emulsion and monolith morphology, as well as the porosity of the materials. We expect that the rate of the photoinduced reaction, which can be adjusted by the structure and proportion of the thiol cross-linker and by the irradiation conditions, might also help optimize the high porosity products. The addition of complementary polymers (such as the hydrophilic PEG) opens a wide window for composite materials bearing porous structures at the same time. Another modification could be the replacement of toluene as the continuous phase with octamethylcyclotetrasiloxane (D4) or silicone oil, allowing the long-term preservation of the monolith in the swollen state, as tested here for sample EG9M.
Selected monoliths or films obtained herein were briefly characterized in order to verify the hypothesis of tunable properties. We have to mention that the materials were not optimized yet, since different properties are required for different applications, which will be developed in the future.
Water contact angle (CA) measurements showed a dual character on some of the porous membranes, with very high contact angle on the upper side (where toluene rapid evaporation ensures an almost continuous silicone layer) and lower values on the bottom side (where in many cases larger pores form due to gravimetrical separation of water-reach droplets). Such a case is sample EG5, which can be regarded as a Janus membrane. Its surface behavior is in line with the observed morphology (Figure 6c). When a highly porous structure resulted, a pseudo-hydrophilic behavior was obtained on both sides, due to the rapid absorption of water in the pores (sample EG6). The values presented in Table 3 were read at the contact of the sessile drop with the surface. When observed in dynamic regime, the contact angle decreased very quickly, due to the porous structure, and air bubbles were visible once water penetrated the membrane, as can be seen in Figure S4 (gif file). If we compare a sample with PEG particles encapsulated with the corresponding one without PEG (EG4 vs. EG5), the difference in CA on both sides was not as pronounced in the former case as in the latter, due to a more compact texture, with pores "blocked" by PEG particles, according to SEM images. For a film that rapidly separated from the water-reach phase (EG8), the increased hydrophobic character of the polysiloxane manifested due to a lower level of porosity. Mechanical tests of the films were conducted in stress-strain setup and the Young's modulus was calculated in each case, showing values lower than 1 MPa, with a general decreasing trend with an increase in porosity (Table 4), similar as reported by other authors [10]. Low maximum strain was a characteristic of these porous materials, since large pores act as mechanical defects in this kind of test. Nevertheless, flexible films were obtained, and the mechanical properties are largely tunable by proper choice of the starting vinyl-silicone. In fact, when the same silicone precursor was cross-linked in the same conditions, without the emulsion gel transient state, its mechanical properties were worse; the film was brittle and could not withstand typical stress-strain measurements. When a silicone precursor with low content of vinyl groups was used (EG8), its Y modulus was lower than that of the similar sample with high cross-linking density (EG6), but a more pronounced separation of the film after UV irradiation was observed for the same composition. On the other hand, the monolith obtained with Co-1.8 (EG8M) had poor mechanical resistance, so it was not further tested. The other three monoliths, prepared from f = 0.5 compositions with both siloxane copolymers in different proportion, were tested in compression mode, either in dry or swollen state (Video S1-video). Satisfactory results were obtained in this stage, for non-optimized soft materials. The samples withstood important compression force and resisted several cycles without damage ( Figure 10). The Young's modulus of the monoliths, measured in compression mode, was one order of magnitude lower than those of the films, measured by tensile tests. When the two monoliths analyzed in swollen state are compared (EG9M and EG10M), lower Y was obtained for the more porous one, but the nature of the swelling liquid is also important. It was found that the monolith swollen in D4 required less effort for the same compression degree, which would mean that D4 has a lubricating effect, due to its high affinity with the matrix. Such materials will be optimized for pressure sensors.  Swelling experiments were done in three different solvents and the results are presented in Figure 11. The absorption capacity was relatively high, compared with the silicone reference (without pores) and generally increased with the internal phase fraction (porosity). Values close to 500% could be obtained in certain cases. Higher absorption capacity for toluene was measured for a monolith (EG8M) compared with the Swelling experiments were done in three different solvents and the results are presented in Figure 11. The absorption capacity was relatively high, compared with the silicone reference (without pores) and generally increased with the internal phase fraction (porosity). Values close to 500% could be obtained in certain cases. Higher absorption capacity for toluene was measured for a monolith (EG8M) compared with the corresponding film (EG8). On the other hand, this monolith did not preserve its integrity after the absorption experiment. corresponding film (EG8). On the other hand, this monolith did not preserve its integrity after the absorption experiment. Figure 11. Absorption capacity (g/g) of xerogels after 30 min of immersion in different solvents.

Conclusions
Two compounds containing short (D) or long (T) dimethylsiloxane segments, and terminal pyridyl groups, designed to be soluble in water or toluene, respectively, were investigated for interfacial properties and emulsions stabilization. Negative interfacial tension values were measured by Wilhelmy plate tensiometry, and quasi-spontaneous emulsification was observed with 0.1% and 1% toluene solution of T. By adding vinylcontaining polysiloxanes to the formed emulsion, followed by photo-initiated cross-linking, emulsion gels were obtained, which were the templates for silicone xerogels. When only T was used as a stabilizer, a low water internal phase could be introduced. By using an aqueous solution of D as a co-surfactant, a stable emulsion with complex structure and a medium or high internal phase was formed by gentle manual mixing and brief sonication in a cleaning bath. The resulted porous materials preserved the morphology of the emulsion template. The presence of a second polymer (PEG), dissolved in the aqueous phase, allowed in situ formation of submicron particles embedded into the silicone matrix, thus opening a novel perspective for easy, one-pot preparation of all-polymer composites from incompatible polymer pairs.

Conclusions
Two compounds containing short (D) or long (T) dimethylsiloxane segments, and terminal pyridyl groups, designed to be soluble in water or toluene, respectively, were investigated for interfacial properties and emulsions stabilization. Negative interfacial tension values were measured by Wilhelmy plate tensiometry, and quasi-spontaneous emulsification was observed with 0.1% and 1% toluene solution of T. By adding vinylcontaining polysiloxanes to the formed emulsion, followed by photo-initiated cross-linking, emulsion gels were obtained, which were the templates for silicone xerogels. When only T was used as a stabilizer, a low water internal phase could be introduced. By using an aqueous solution of D as a co-surfactant, a stable emulsion with complex structure and a medium or high internal phase was formed by gentle manual mixing and brief sonication in a cleaning bath. The resulted porous materials preserved the morphology of