Hierarchically Porous Ceramic and Metal‐Ceramic Hybrid Materials Structured by Vat Photopolymerization‐Induced Phase Separation

Additive manufacturing techniques such as vat photopolymerization have laid the foundation for impressive advances in the 3D structuring of ceramic materials. However, simultaneous structuring of these complex‐shaped ceramic objects on the sub‐µm scale, an essential feature for a wide range of applications in separation, energy conversion and storage, adsorption or sensing, has remained a tremendous challenge. This study demonstrates how complex‐shaped polymer‐derived SiOC ceramics exhibiting hierarchical porosity ranging from the sub‐µm‐ to the millimeter‐range can be generated by combining vat photopolymerization with photopolymerization‐induced phase separation using preceramic polymer‐based phase‐separating resins. In addition to allowing single‐step, multi‐level structural control, this new processing concept allows for the chemical modification of the 3D‐printable, phase‐separating preceramic polymer resins using organometallic compounds, including the possibility to generate functional metal nanoparticles in situ during the polymer‐to‐ceramic conversion. In this manner, a chemical toolbox is provided, facilitating the introduction of Ni, Co, Mo, or La into the hierarchically structured SiOC matrix. The versatile applicability of this new materials design approach is demonstrated by employing complex‐shaped, hierarchically porous monoliths containing in situ generated Ni nanoparticles as heterogeneous catalysts for CO2 methanation, with a profound increase in catalyst performance attained by oxidative post‐treatment of the metal‐ceramic hybrid material.


Introduction
Ceramic materials combine excellent mechanical, thermal, and chemical stability, and have thus emerged as material class of choice for a wide range of technological processes taking place in harsh conditions.Owing to their thermal stability, consolidation of ceramics typically takes place via sintering of particulate starting materials, which strongly reduces the flexibility in both nano-and macro-scale materials design.The preceramic polymer (PCP) route provides an alternative approach toward the design and implementation of ceramics through a bottom-up approach.By converting (mostly Si-based) polymeric starting materials into amorphous or nanocrystalline ceramics through a controlled thermal conversion treatment, the full range of polymer chemistry and polymer technology can be utilized for compositional control, chemical functionalization, and processing options. [1]Ceramics such as SiOC, SiC, or Si 3 N 4 can, in this manner, be derived from polysiloxane, polycarbosilane, or polysilazane compounds at temperatures far below temperatures required for conventional sintering.While primarily utilized for the fabrication of ceramic fibers upon its inception, [2] the field of polymer-derived ceramics (PDCs) has evolved into a thriving and prolific field of research and scientific activity, with the overall aim to create novel materials suited for the needs of future applications, owing to excellent mechanical and thermal stability. [3,4]A field of steadily rising interest is the utilization of PCPs for the design of porous ceramics, due to the wide range of porosification approaches accessible, but also due to the circumvention of inherent limitations of PDC technology stemming from the evolution of gaseous decomposition products. [5]As porous ceramics are often used in harsh environments, PDCs can play a particularly important role in fields such as separation, energy conversion, or as catalyst supports. [6,7]n addition to the possibility to introduce and control pore architectures, the PCP approach facilitates a straightforward way to introduce functional features by in situ formation, such as metal nanoparticles [8] or percolating carbon networks, [9] or by integration with 2D materials. [10]The introduction of mesoporosity into PDCs has been of particular recent interest for prospective applications in clean energy technologies. [11]14][15] In the recent past, light has been successfully employed as a structuring agent for the shaping of matter, spanning the full dimensional range from the macro-to the microscale.A major advantage of using light is its unmatched performance in temporal and spatial control, i.e., the possibility to initiate a photoreaction at a specific place and a specific time. [16]Light-based additive manufacturing (AM) has become one of the primary tools in 3D structuring of materials, the most prominent techniques being powder bed fusion and vat photopolymerization.Vat photopolymerization (VPP), which encompasses various laser-based or digital light processing (DLP)-based techniques, has become of particular importance due to the high resolution and precision not attainable by other AM techniques. [17]VPP generally relies on the presence of photocurable groups facilitating polymerization either via chain-growth or step-growth mechanisms, [18] thus being effectively limited to polymer systems.Nonetheless, it has been successfully applied for the structuring of ceramics primarily through a colloidal approach, by introducing ceramic powders into a photopolymerizable matrix, with subsequent debinding of the organic matrix and conventional sintering of the ceramic particulates. [19]Alternatively, photocurable PCPs can be used.In seminal works by Zanchetta et al. [20] and Eckel et al., [21] VPPbased structuring of polysiloxane-based systems was achieved either via acrylate moieties or via thiol-ene chemistry, respectively, followed by a pyrolytic conversion to obtain ceramics.In recent years, VPP has been implemented for many other PCP systems, including polycarbosilanes and polysilazanes, [22][23][24][25] albeit with more limited success in obtaining competitive materials due to the reduced number of options in reaction mechanisms in combination with the chemical sensitivity of typical non-oxide precursors.Recent developments in the field of light-based AM of PCPs involve the functionalization of polysiloxane-derived AMstructured ceramic parts, including the in situ formation of metal nanoparticles [26][27][28] or the fabrication of composites to improve mechanical performance. [29]ight can not only be used to facilitate structuring of materials on the macro scale, but can also be utilized as a tool to control a material's structure in the sub-μm domain.In contrast to strategies aiming at improving the spatial resolution of light exposure, an example being two-photon polymerization, [30] nanostructural control can also be gained by using light to affect physicochemical interactions of the starting material on the molecular level, e.g. by achieving phase separation between thermodynamically incompatible components leading to spinodal decomposition and resulting in discrete nano-scaled domains. [31]One possibility to stimulate this process by light is photopolymerizationinduced phase separation (PIPS), [32] with a high flexibility in controlling the resulting morphologies by composition of the starting mixture, polymerization temperature, and light intensity. [33]n contrast to thermally induced polymerization, photopolymerization excels in the high spatio-temporal control of the stimulus initiating polymerization.By selection of adequate phases, which should be miscible at the onset of the process but immiscible after polymerization, distinct properties such as mechanical stability can be improved. [34]In contrast, if one of the separated phases is selectively removed after polymerization by extraction or thermal decomposition, nanoporous materials can be obtained, with the resulting pore morphologies replicating the initial phase morphology. [35,36]Consequently, by controlling the phase separation process itself, the resulting pores can be tailored, with morphologies ranging from discrete globular appearance to interpenetrating networks, the latter of which are often preferred in applications involving the efficient transport of media, which includes filtration or catalysis. [37]Very recently, PIPS has been combined with VPP to achieve simultaneous nanoand macrostructuring of polymer materials, with the aim to either spatially control functional material phases in 3D printed objects, [38] to increase mechanical properties of printed parts, [39] or to obtain nanoporous 3D objects. [40]However, up to this point, the concept of combining PIPS with VPP has been nearly exclusively limited to organic materials.To our knowledge, the sole translation of this concept from the organic into the inorganic material domain was reported by Moore et al., [41] who developed nanoporous, 3D printed glasses.After combining acrylate monomers with poly(diethoxysiloxane) as inorganic precursor, phase separation was initiated during VPP carried out using a DLP-based desktop system.Furthermore, it was shown that the resulting mesopore size could be locally tuned in the sub-μm range by modifying the respective light intensity.
With respect to ceramic materials, various AM approaches were utilized for the generation of hierarchically porous structures, examples including the use of capillary suspension-based inks, [42][43][44] direct foam writing, [45] or stable emulsions. [46]However, these approaches generally rely on the use of ceramic particles, with the respective limitations already mentioned.In terms of molecular precursors, geopolymers [47] and preceramic polymers [48] have been successfully used to obtain hierarchically porous ceramics via direct ink writing, while VPP has been used in combination with sol-gel-based systems to generate hierarchically structured silica materials. [49]he objective of this work is to, for the first time, demonstrate the successful implementation of vat photopolymerization-induced phase separation for the generation of hierarchically porous, complex-shaped ceramics using light as structure-directing stimulus.Furthermore, we fully exploit the potential of preceramic polymer technology by modifying the phase-separating preceramic resins with metal-organic moieties while retaining their printability, thus providing a straightforward material toolbox for the introduction of nanoscaled metal and metal-based centers into the hierarchically structured ceramic matrices.The applicability of this concept for prospective green technologies is demonstrated by employing our 3D printed, metal-modified ceramics with multiscale porosity in a CO 2 methanation model setup, highlighting the versatile nature of this combinatory processing approach.

Design and Evaluation of Phase-Separating Resin Formulations
The interactions between the photosensitive acrylate (TMPTMA) and the preceramic polymer solution (PSO-TPM) were investigated to produce a photopolymerizable preceramic resin with distinct photopolymerization-induced phase separation behavior, suitable for structuring via vat photopolymerization (Figure 1A).As the acrylate is largely immiscible with the polymer solution, tetraethyl orthosilicate (TEOS) was employed as solubilizing agent.The miscibility behavior of the resin, depending on the concentration of the three main components, was investigated by cloud-point titration.Figure 1B depicts the localized miscibility gap as green area where phase-separation occurs within the unpolymerized resin.The counteracting behavior of TEOS, which is highly miscible with PSO, and TMPTMA, which has low miscibility with PSO, was used to design resin compositions with different degrees of phase separation.Four resin compositions were selected to study the effect of phase separation during vat photopolymerization and the final pore formation within the ceramic material (Figure 1B).To ensure sufficient mechanical stability of the printed components and a bicontinuous network between acrylate and preceramic compounds, an acrylate content of at least 35 wt.% was used.Resins 1 and 2 are located at the edge of the miscibility gap; therefore, non-separated resins are initially present, with phase separation induced by photopolymerization.Resin 3 is located well within the miscible region of the system, and therefore a different degree of separation during polymer- ization is to be expected.Resin 4, on the other hand, is located well within the miscibility gap, and therefore a phase-separated resin is present even before photopolymerization. Figure 1C visualizes morphological changes taking place during processing, including photopolymerization-induced phase-separation during printing as well as porosification during pyrolytic conversion.As the homogeneous preceramic resin undergoes photopolymerization, the acrylate monomers polymerize and form a rigid network.Within the acrylate network, a bicontinuous network of preceramic constituents is entrapped.During the subsequent thermal decomposition process, PSO is converted into SiOC, while the acrylate decomposes, thus creating a hierarchically porous ceramic network.Likewise, the solvent (TPM) and TEOS are removed by evaporation during the thermal treatment.
Prior to printing of the resin compositions, photorheology measurements coupled with in situ near-infrared (NIR) spectroscopy were conducted to study the photopolymerization behavior of the resins.The most decisive factor in these investigations is the change in storage modulus (G′) (Figure 2A), which is an indicator for the speed of the reaction and the final stiffness of the cured polymer.The second indicator for the kinetic characteristics of the photoreaction is the course of the double-bond IR signal, shown in Figure 2B, which is calculated as doublebond conversion (DBC) with respect to the starting intensity of the methacrylate signal (6100-6200 cm −1 ).In addition, the normal force between the two plates of the rheometer is recorded, which is affected by photopolymerization-induced shrinkage.At the start of the photoreaction (t = 0), a steep rise in G´and DBC is evident for all resin compositions, indicating a rapid photoreaction within the first seconds of light exposure.Likewise, a steep decline in normal force, also attributed to a rapid progress of photopolymerization, is observed.The starting point of the curing reaction is typically given as the gel point (t gel ), characterized by the overlap between the storage modulus and the loss modulus (Table S1, Supporting Information).Here, the first difference between the polymerization kinetics of the resins becomes apparent, as resin 3 shows a higher t gel value.The decreased reaction rate is caused by the lower concentration of acrylate within resin 3, which indicates a general kinetic hindrance of the polymerization process by the nominally non-reactive components within the resin (PSO, TPM, TEOS).In addition, this composition is furthest away from the miscibility gap, which in turn causes a stronger kinetic hindrance during photopolymerization-induced phase separation.In line with these findings, resin 3 exhibits a lower shrinkage (indicated by the normal force) resulting from the lower acrylate content.Furthermore, a decreased final modulus is observed due to the decreased acrylate concentration.When comparing resin composition 4, located well within the miscibility gap, with the two resin compositions close to the edge of the miscibility gap (resin 1 and 2), no difference in the initial reaction rate is found.However, resin 4 undergoes a considerably higher shrinkage (Figure 2C), which can be attributed to mobility restrictions due to resin heterogeneity. [50]The differences between the two resin compositions located at the boundary of the immiscible region (resins 1 and 2) are particularly apparent in the photorheology measurements.Although resin 2 contains a higher acrylate content, the reaction rates of the two formulations are comparable.Resin 2 exhibits a lower t gel (Table S1, Supporting Information), while resin 1 achieves a significantly increased final modulus, again highlighting the essential role of phase separation kinetics on the formation of the resulting photopolymer network.A similar conclusion can be derived when observing the DBC characteristics (Figure 2B): all four resins exhibit a comparable DBC behavior, but the network formation differences become apparent via the storage modulus. [50]Based on its increased modulus, a tighter network with a higher degree of cross-linking can be anticipated in resin 1 after photopolymerization, suggesting the presence of a well-developed bicontinuous network.
Cylindrical samples were successfully printed using all four resin compositions (printing parameters are listed in Table S1, Supporting Information), and were converted into distinct porous structures by pyrolysis at 700 °C.A pronounced openporous network can be observed in all four derived ceramic materials (Figure 3).A gradient in pore size is present within each printing layer, generally exhibiting a larger pore size in intralayer regions closer to the light source and a smaller pore size at the intralayer regions further away from the light source.As the phase separation is attributed to a spinodal decomposition, the formation of the bicontinuous network depends on the polymerization rate. [40,41]Therefore, a higher polymerization rate is present at the onset of the layer formation, which decreases with the ongoing layer formation.The reduced polymerization rate can mainly be attributed to the light attenuation during the layer build-up, leaving lower light intensities for the end of the layer formation. [41]hen comparing the pore morphologies of samples derived from different resin compositions, resin 3 deviates from the other resins.Increased pore sizes and strut sizes can be observed, also reflected in a shift in pore size as determined by mercury intrusion (Figure S1, Supporting Information).The difference in morphology is in accordance with the photoreaction characteristics, as slower reaction rates generally result in larger phase domains within the bicontinuous network.Furthermore, the decreased acrylate content is responsible for the increase in strut size.Resin 4, in contrast, exhibits the lowest median pore diameter (Table 1), which can be correlated with the increased acrylate concentration, the increased polymerization rate, and the high shrinking stress.Resin 2, located at the boundary of the immiscible region, exhibits a pore morphology comparable to resin 4. In this case, the similarity in acrylate content appears to have a larger effect on the resulting pore morphology than the pre-separation present in resin 4. Resin 1, in spite of also being located at the boundary of the miscibility gap, exhibits a pore morphology distinct from the one observed for resin 2. Resin 1 exhibits a narrower pore size distribution (Figure S1, Supporting Information) with a well-defined pore morphology.These morphological descriptors indicate that the well-balanced polymerization behavior of resin 1 results in the most uniform continuous network within the tested formulations.
In addition to the macroporous structures obtained through phase separation-based templating, an additional class of pores in the micro-and mesopore range can be found in pyrolyzed samples via N 2 physisorption.All structures exhibit Type II isotherms, albeit indicating only a minor amount of mesoporosity within the pyrolyzed samples (Figures S3 and S4, Supporting Information).The presence of micropores can be mainly attributed to the polymer-to-ceramic conversion, with the outgassing or decomposition of volatile substances during the thermal treatment causing the formation of transient pore structures. [51]The retention of micropores within the ceramic structure largely depends on the pyrolysis temperature. [52]To investigate this relationship for the produced structures, samples of printed resin 1 were pyrolyzed at 600, 700 and 800 °C (Table 1, Table 1.Specific surface area (BET, N 2 physisorption), average pore size (BJH, N 2 physisorption), and median pore opening diameter (mercury intrusion) of printed specimens from different resin compositions after pyrolysis (resin 1: at 600, 700, and 800 °C; resins 2-4: at 700 °C).
Average pore size  Figures S2 and S4, Supporting Information).The collapse of microporosity becomes evident at 800 °C, with a drastic decrease in specific surface area (SSA).While the highest SSA is -as expected -obtained after pyrolysis at 600 °C, an SSA of 249 m 2 g −2 is still present after treatment at 700 °C, which is comparably high for a PDC system, especially for a monolithic structure. [5,52]This behavior is most likely caused by the use of the acrylate compound and solvents (TPM, TEOS).In analogy to the change in SSA, an increase in pore size with increasing temperature can be detected via the BJH method, related to the closing of small micropores within the material.Based on these findings illustrating the general temperature dependence of meso-and microporosity within the material, other resin compositions apart from resin 1 were only evaluated after pyrolysis at 700 °C.In these systems, the acrylate concentration within the resins appears to correlate with the resulting SSA: resin 3, containing the lowest amount of acrylate, exhibits the lowest SSA value, which may further be attributed to the decreased phase separation tendency observed for this system.Resins 1 and 2, located at the boundary of the immiscible region, yield comparable SSA values and mesopore sizes.In contrast, resin 4 exhibits an increased SSA value, with the irregular distribution of the acrylate within this system potentially leading to a more pronounced degradation of the acrylate on a molecular level.
With respect to the chemical composition of the products, comparable carbon contents were found in all resin formulations pyrolyzed at 700 °C (Figure S5A, Supporting Information).In case of oxygen content, resin 3 exhibited an increased oxygen content compared to the rest of the formulations, the high amount of TEOS and a low acrylate content likely being responsible.When comparing carbon and oxygen contents of resin 1 pyrolyzed at 600, 700 and 800 °C, a decrease in carbon and oxygen content with increasing temperature was observed (Figure S5B, Supporting Information).As a mass loss was also observed in TGA experiments between 600 and 800 °C, the decrease of oxygen and carbon content can be attributed to the elimination of residual organic structures.
Based on photorheological investigations mentioned earlier, resin 1 demonstrates the most favorable properties for use in vat photopolymerization of porous SiOC within these material combinations.The significantly higher storage modulus, fast polymerization time, and moderate shrinkage indicate a beneficial behavior with respect to phase separation and photopolymerization, in addition to its suitable handling properties after printing.With respect to the pore structure after pyrolytic conversion, resin 1 exhibits the most uniform pore size distribution, a direct result of placing the resin formulation at the boundary of the miscibility gap and balancing the acrylate-to-preceramic ratio toward a bicontinuous structure.Consequently, this resin formulation was used for the generation of complex-shaped structures, as well as for a subsequent modification with metal compounds.
Two designs were selected to showcase the capability of structuring complex macroporous geometries, thus adding an additional level of porosity via vat photopolymerization.Feature sizes of 300 and 400 μm were achieved for gyroid (Figure 4A) and lattice structures (Figure 4B) in the printed state, respectively.Phase separation within the printed components is indicated by the material's turbidity.After cleaning with TPM solvent, the printed structures were thermally cured at 130 °C in N 2 to induce condensation reactions within the preceramic phase.After this stage, the microstructure within the polymer samples (Figure S6, Supporting Information) is controlled by phase separation, resulting in the formation of globular domains consisting of either acrylate-containing phase or preceramic phase.Already at this stage, the first formation of porosity is evident, resulting from the evaporation of the solvent (TPM).Subsequently, the printed complex-shaped structures were successfully converted into hierarchically porous SiOC structures by pyrolysis at 700 °C in Ar atmosphere (Figure 4C,D).During this process, a linear shrinkage of 24.1 ± 0.1% within the plane perpendicular to the printing direction (XY) and 26.8 ± 0.5% in the printing direction (Z) was found, with an overall ceramic yield of 20.4 ± 0.1%.This anisotropy in linear shrinkage can be attributed to a compaction of the printing layers during pyrolysis. [29]he possibility to generate fully accessible macro-, meso-and micropores within a complex-shaped monolithic ceramic object provides a powerful means to fabricate tailored structures for specific applications in diverse technological fields including heterogeneous catalysis, separation, energy conversion, or bioengineering.

Metal Modification of the Preceramic Polymer
In a next step, the preceramic base polymer (PSO) was modified with metal precursors of different elements, targeting a prospective application of the produced hierarchical porous ceramic material in heterogeneous catalysis.[55] The modification was achieved by a reaction of the polysiloxane's hydroxy groups and the metal center of the acetylacetonate, resulting in the formation of oxygen-metal bonds (Si-O-M).As the metal centers provide valences for two reactions, partial crosslinking through oxygen-metal bridges can be expected, alongside the addition of metal acetylacetonate side chains.The anticipated schematic reaction is depicted in Figure 5A.To elucidate the in-teraction between the polysiloxane and various metal acetylacetonates (Ni, Co, Mo, and La), FTIR measurements of unmodified and modified polymers were conducted (Figure 5B).The formation of Si-O-M bonds could be confirmed after the reaction with the Ni, Co, and Mo acetylacetonate by the detection of a newly emerging absorption band at 930 cm −1 . [53,56,57]Furthermore, an increase in the M-O absorption band at 665 cm −1 was found, further underlining the incorporation of the metal into the polymeric structure.During the reaction, the acetylacetonate ligand leaves via acetylacetone, which is removed by the stripping of the solvent after the reaction.However, since monosubstituted hydroxy groups also remain present, the presence of acetylacetone side groups can still be detected in the modified polysiloxanes via C═O, C═C, and C─H vibrations at 1572, 1542, and 1423 cm −1 , respectively.In contrast to the Ni-, Co-, and Mo-modified polymers, no Si-O-M formation is evident after modification with La.This deviation results from an acid-base reaction of the acetylacetonate with the propionic acid used in this case.La is therefore introduced into the precursor via a sol-gel reaction involving propionic acid. [58]

Characterization of Metal-Modified Resins
Photorheology measurements coupled with NIR spectroscopy were conducted to elucidate the influence of metal modification on the photoreactivity of the printing resins, using the same conditions as for the unmodified resin.The same ratios of (modified) polymer (metal-PSO), solvent (TPM, propionic acid), acrylate (TMPTMA) and TEOS were used as in resin composition 1 (see Section 2.1).Again, homogeneous and transparent resins could be prepared, which undergo a well-defined phase separation during photopolymerization. Figure 6 shows the photopolymerization characteristics of the metal-modified resins in comparison to the unmodified resin 1.The modification reduces the maximum storage modulus for all metal-modified systems, correlating with a lower stiffness of the printed parts.Especially in case of the Co-modified resin, a significant reduction in storage modulus is observed compared to the unmodified as well as Ni, Mo, La-modified resins (Figure 6A).When the reactivity, i.e., the gel point of the systems, is compared (Table S2, Supporting Information), a slight increase is found with the addition of La, Mo, and Ni.Here, La exhibits the strongest increase.The increase in reactivity is attributed to the catalytic effect of the metals on the photoreaction. [28]However, the Co-modified system exhibits a significantly lowered reaction rate.The main reason for the decrease can be seen in the dark coloration caused by the Co complex, which sharply reduces the light intensity within the forming layer and thus decreases the polymerization rate.When comparing the DBC of the unmodified resin and the metal-modified resins, La, Mo, and Ni again exhibit a higher reaction rate (Figure 6B).The lower reaction rate and lower final curing rate of the Co is likewise reflected in the DBC measurement.As a consequence of the higher achieved modulus within the unmodified resin, a higher shrinkage stress of the unmodified system is observed (Figure 6C).In addition to photorheological investigations, curing tests were conducted on the 3D printer to derive the printing parameters for the actual production of parts (Figure 6D).In these measurements illustrating the curing depth as a function of exposure time, the same trends observed via photorheology are found: the unmodified resin shows the highest curing depth at a given time, followed by Ni, La, and Mo, respectively.Here, the Co resin again exhibits the lowest curing rate, with curing times > 30 s required for achieving a curing depth suitable for the vat photopolymerization process.

Additive Manufacturing and Pyrolytic Conversion of Metal-Modified Phase-Separating Resins
Based on the photo-rheology results and curing depth determinations, complex gyroid structures were fabricated for all metalmodified resins.When comparing the printed sample parts (Figure 7A-D), a striking difference in coloration is evident, caused by the different used metal precursors.For all variations, the characteristic turbidity derived from phase separation is apparent.As all metal modifications introduce a color change into the printing resin, no contrast agent had to be used, instead utilizing the precursor coloration itself as contrast agent.In case of Ni, Mo, and La, the exposure time during printing could thus be kept on a similar level as in the unmodified resin (Table S5, Supporting Information).The Co-modified resin exhibited a dark coloration and light attenuation, resulting in significantly longer exposure times for each layer (Table S5, Supporting Information).Significant differences were observed when comparing the quality of the printed structures in terms of surface characteristics, printability, and handleability.The modification of PSO with the Ni precursor resulted in a positive effect on printability, in particular in terms of resolution.Together with a good handleability of the parts and high resin stability, highly resolved parts with favorable surface characteristics could be produced.The Comodified parts showed significantly lower stiffness, thus particular care had to be taken when handling the samples for postprocessing.Nonetheless, high-quality components with a suitable surface finish could be produced.The Mo-modified variant showed a decent stiffness resulting in good handleability during post-processing, which included washing with TPM solvent for several cycles, as pronounced gelation of the resin within the structures was evident.La-modified parts exhibited the highest stiffness and, therefore, favorable handling characteristics compared to the other metal-modified materials.Due to partial gelation of the resin within the printed structure, cleaning of printed parts was conducted with a more aggressive solvent (isopropanol) and conducted for several cleaning cycles.The result of this post-processing is visible in the higher surface roughness of the La-modified sample.Based on the results of the systematic printing tests involving the distinct metal-modified resins, the Ni-modified system exhibited the most promising behavior in terms of processability of the resin and post-processing of printed parts.
When comparing the microstructure of the different metalmodified parts before pyrolytic conversion (Figure S7, Supporting Information), the Ni-, Co-and La-modified samples exhibit microstructural features comparable to the unmodified sample.Phase-separated polymer clusters are visible in all cases, with porosity caused by already evaporated solvent (TPM, propionic acid).In contrast to the unmodified sample, a secondary phase is visible between the polymer domains, indicating the separation of a metal-rich phase.The Mo-modified system, in contrast, contains a different microstructural situation with a considerably less phase-separated structure.This difference in structure indicates either an interference of the Mo precursor with the phase separation process or a reaction with one of the other preceramic compounds.This assumption is likewise supported by the difference in mass loss during pyrolytic conversion, observed by thermogravimetric measurements (Figure S8, Supporting Information).Here, the Mo-modified specimen exhibits a significantly lower mass loss compared to the other metal-modified materi-als or to the unmodified sample.The difference in mass loss is most likely caused by a formation of bonds between TEOS and PSO caused by the presence of Mo.In contrast, Ni-, Co-, and Lamodified specimens show a higher mass loss than the unmodified system due to the additional introduction of decomposable constituents.
The printed metal-modified parts were subsequently pyrolyzed at 700 °C in Ar (Figure 7E-H).All structures could be converted into metal-modified ceramic structures without severe crack formation.The previously discussed effect of the different metals on the quality of the components also becomes apparent in the converted state.The uneven surface on the La components caused by the differences in post-processing is equally visible after pyrolysis.In contrast, the Mo-modified specimen exhibits a much more uniform, glossy surface.In case of the Ni-modified material, the printed shape is retained most precisely, the surface appearance even surpassing the unmodified sample.
When examining the microstructure of the surface of metalmodified ceramic samples by SEM (Figure 7I-L), the Ni-, Co-, and La-modified materials exhibit an open porous surface, as originally targeted through the utilization of polymerization-induced phase separation.The materials contain struts connected by spherical structures, similar to the surface of the unmodified parts, with an apparent decrease in strut thickness in the case of the metal-modified variants.In contrast to the other materials, the Mo-modified part exhibits an almost closed surface, resembling a wavy microstructure.The difference in surface morphology again reflects the difference in the polymeric state of the Mo-modified variant.
The cross-sections of the pyrolyzed parts, perpendicular to their printing direction, show the microstructure within the metal-ceramic material (Figure 7M-P).A porous structure comparable to the unmodified material is found for all metalmodified materials.The effect of differences in porosity throughout each printing layer is more pronounced in the Ni-, Co-, and La-modified samples compared to the unmodified material, which is likely caused by the difference in coloration of the modified printing resins.In contrast, the Mo-modified sample shows a uniform distribution of porosity within the material.Mercury intrusion porosimetry similarly illustrates the macropore size distribution (Table 2 and Figure S9, Supporting Information).The bimodal distribution of the pore opening diameter is particularly pronounced for the Ni-modified specimen and likewise detectable for the Co sample, fitting the observed situation in the microstructure.A more uniform, yet still bimodal distribution is present in the La-modified sample.The Mo-modified material, however, shows a continuous pore size distribution, with an apparently high amount of relatively small pore open diameters, most likely caused by the relatively closed surface structure.
The presence of meso-and microporosity is reflected in N 2 physisorption measurements (Table 2 and Figure S10, Supporting Information).The Ni-modified sample displays a comparable specific surface area as the unmodified sample, while a minor surface area reduction is evident within the Co-, Mo-, and La-modified samples.[61] As is the case for the unmodified samples pyrolyzed at the same conditions, only a minor amount of mesoporosity could be detected by the BJH method, the isotherms depicting a Type II shape.
The bulk elemental composition of Ni-, Co-, Mo-, and Lamodified materials pyrolyzed at 700°C was determined by combining combustion, hot gas fusion, and EDX methods (Figure S11, Supporting Information).The variation of the metal precursor does not significantly affect the carbon content of the material, which varies between 28 wt.% and 32 wt.%, but a deviation in oxygen content was observed for the La-modified sample.
Here, the difference in printing solvent (changed from TPM to propionic acid) or the different number of ligands in the precursor may be the cause.For Ni, Co, and Mo, an underrepresentation of the metal content compared to the anticipated 5 wt.% was found.Most likely, the assumed mass loss for the calculation was not meeting the actual mass loss of the requisite sample.
In comparison to the unmodified material (Resin 1_700: 22.4 wt.% C and 24.2 wt.% O, see Figure S5, Supporting Information), a significant increase in C and O content was found for metal-modified materials, which may be attributed to the incorporation of organic material derived from the acetylacetonate ligands in the metal precursors.
As the formation of nano-sized metal clusters within the hierarchic porous ceramic material is anticipated by the modification of the polysiloxane with several metal precursors, the formation of separated secondary metal phases was studied with transmission electron microscopy (TEM) and X-ray diffraction (XRD).Through the material contrast given by scanning transmission electron microscopy in high angle annular-dark field (STEM-HAADF), the presence of Ni and Co particles within the SiOC matrix is further underlined (Figure 8E,F) and proven by the use of energy-dispersive X-ray spectroscopy (EDX) mappings (Figure 8I,J).The presence of a bimodal particle size distribution is again reflected in STEM-HAADF and EDX images.In contrast, the La-modified ceramic shows the formation of nanosized particles under 2 nm within the SiOC matrix (Figure 8D).In addition, broader La-rich domains could be identified by STEM-HAADF with a size in the range of 100 nm, verified by EDX mappings (Figure 8H,L).A different situation was found for the Mo-modified ceramic, as no distinctive Mo particles or domains could be detected (Figure 8C).Furthermore, no distinct Mo-rich domains were identified by EDX (Figure 8K), indicating a relatively homogeneous distribution of Mo within the SiOC matrix.Variations in contrast within the STEM-HAADF image (Figure 8G) can thus be attributed to morphology effects within the prepared samples.
The difference in segregation of the studied metals can mainly be attributed to their reducibility by means of carbothermal reduction. [53]As Ni-and Co-precursors are typically reduced at temperatures beneath 400 and 500 °C, respectively, the presence of elemental metal is expected. [60,62,63]This reducibility is well reflected in the Gibbs free energy change of the different metal oxidation reactions versus the oxidation of carbon. [64]Therefore, the expected presence of metallic face-centered (fcc) Ni and Co phases could be determined by XRD measurements (Figure S12, Supporting Information) and selected area electron diffraction (SAED) (Figure S13, Supporting Information).A well-crystallized fcc-Ni phase could be determined within the Ni-modified ceramic by well-defined reflexes in XRD and corresponding diffraction spots in SAED measurements of selected larger Ni particles.The Co-modified exhibits a lower degree of crystallinity within the Co phase, as broader reflexes and fewer diffraction spots could be measured in this case, following the reducibility of the given metal.Following the principle of the Scherrer equation, the crystallite size of the Ni and Co metal phases was calculated (JCPDS: fcc-Co: 04-014-0167, fcc-Ni: 00-004-0850). [62]Crystallite sizes of 23.9 ± 0.2 and 12.3 ± 0.9 nm for Ni and Co could be determined, respectively, in accordance with the given larger particle hierarchy in each system.As no segregation of a Mo phase could be determined by TEM measurements, the amorphous nature of the Mo-modified ceramic is also reflected in the XRD measurement.When considering the Gibbs free energy of the carbothermic Mo reduction, a reaction is likely at temperatures above 750 °C.
Therefore, no formation of elemental Mo was expected following this modification.As Mo shows a high tendency for carbide and silicide formation, an integration of Mo within the SiOC matrix in the form of an amorphous MoSiOC structure can be underlined by thermodynamic considerations. [65]In previous reports, the segregation of crystalline molybdenum carbide or molybdenum silicide phases within PDC matrices could only be detected at pyrolysis temperatures above 1100 °C. [65,66]In the case of the La-modified ceramic, an isolated consideration of the metal's reducibility is insufficient, as no carbothermal reduction can be expected in this case.La exhibits a high tendency to form oxides, similar to Zr, Hf, or Lu. [53]As La has not been incorporated into a polymer-derived SiOC matrix in prior studies, the mechanistic behaviour can only be derived from similar elements: in these cases, formations of amorphous oxide nano-particles and larger metal oxide enrichments within the SiOC matrix have been reported at lower pyrolysis temperatures, with crystallization only occurring at elevated temperatures (above 1400 °C). [53,67,68]herefore, the presence of a segregated amorphous lanthanum oxide phase within the SiOC matrix is suggested within this study.
Table 3. Specific surface area, average pore size, and median pore opening diameter of printed specimens derived from 20 wt.% Ni-modified resin 1 after pyrolysis at 600 °C in Ar and after oxidative post-treatment, respectively.

Catalytic Properties
Based on the results of the comparative study of Ni-, Co-, Mo-, and La-modified SiOC ceramics, the Ni-modified system stands out with respect to its processability for vat photopolymerization as well as the formation of individual nano-scaled metal particles within the hierarchically porous SiOC matrix.Ni, which is a highly suitable catalyst in CO 2 utilization reactions such as CO 2 methanation, [69] has been increasingly considered as a model material with regard to the immobilization of catalytically active centers by in situ formation of metal nanoparticles within PDCs. [55,60,70,71]Consequently, the Ni-modified system developed in this study was evaluated toward its prospective use as monolithic catalyst material, providing both catalyst and carrier functionality.The first process modification toward this prospective application was an increase of the Ni content within the final material from 5 wt.% to 20 wt.%, which was successfully accomplished without a need for changing the polysiloxane functionalization approach or the printing parameters.In analogy to the material variant containing 5 wt.%Ni, a monolithic gyroid design was implemented (Figure S14A, Supporting Information).The design was only slightly modified to provide cylindrical gyroid parts suitable for testing of the catalytic performance in a tubular reactor.The second modification to improve the material's use as catalyst structure included a reduction in pyrolysis temperature to 600 °C, resulting in an increased specific surface area (Table 3).
Here, an additional benefit can be anticipated from changes in surface characteristics of SiOC pyrolyzed at lower temperatures, which have been found beneficial for CO 2 methanation. [28,71,72]In this manner, printed structures could successfully be converted into porous Ni/SiOC monoliths (Figure S14B, Supporting Information).
During an initial evaluation of the catalytic performance of Ni/SiOC monoliths pyrolyzed at 600 and 700 °C in a CO 2 methanation model reaction setup, no significant activity was found, indicating a hindrance in the accessibility of the active Ni sites for the gaseous reaction educts (Figure 9A).To verify this assumption, the monolithic samples were crushed into powders, resulting in a significant increase in CO 2 conversion of up to 26% for the material pyrolyzed at 600 °C.Based on these results, a blockage of Ni sites can indeed be assumed.By selectively removing the free carbon within the Ni/SiOC structure by an oxidative post-treatment in air at 440 °C, the loss of free carbon also being reflected by a color change from black to grey (Figure S14C, Supporting Information), carbon domains potentially blocking access to active sites were successfully removed from the porous monoliths.While the specific surface area is largely unaffected by this treatment (Table 3), a change in the shape of the N 2 adsorption isotherm was found (Figure S15A, Supporting Information).The change from a Type II to a Type IV isotherm in the air-treated sample suggests the evolution of mesoporous structures within the material, which is verified by the BJH plot (Figure S15B, Supporting Information), depicting an increased pore volume at a pore width between 5 and 25 nm.A similar trend is visible in the mercury intrusion measurements, which also depict pores in a size range below 20 nm (Figure S16B, Supporting Information).Therefore, the free carbon within the SiOC system, incorporated by using the acrylate and acetylacetonate precursor, appears to act as a template for mesoporosity: by removal of this free carbon during oxidative post-treatment, the resulting mesoporous channels act as pathways for the reactants (CO 2 and H 2 ) toward the catalytically active sites.This last step of template removal consequently yields another class of pores in the mesoporous range.
Consequently, by oxidative post-treatment, the monolithic structures yielded CO 2 conversion values up to 48 % at a reaction temperature of 400 °C (Figure 9A).The catalytic efficacy, which was examined at various reaction temperature steps between 200 and 400 °C, showed a distinct temperature dependence, with an increased conversion rate at elevated temperatures.By retaining the reaction conditions at 400 °C for a duration > 50 h, the stability of the catalyst was evaluated (Figure 9B), showing only a minor drop in CO 2 conversion within the first 3 h before remaining at a relatively high level of > 42 % over the remainder of the experiment.Throughout all catalytic performance tests involving the material after oxidative post-treatment, a CH 4 selectivity of 100% was observed.
When the catalytic performance is put into perspective by comparison with previous works on metal-modified PDCs for catalysis applications in the literature, a clear overcoming of current limitations can be demonstrated.In the past, several reports on the use of metal-modified PDCs for catalysis, especially CO 2 methanation, have been shown, albeit mainly limited on powdered materials. [59,60,73,74]To overcome these limited shaping possibilities, Schumacher et al. [71] reported on the generation of PDC-based, Ni-modified monolithic catalysts via freeze-casting; however, the determination of the catalytic activity of the materials was also only carried out in powdered form.Consequently, a distinct evaluation of the suitability of monolithic PDC-based components for this purpose has been lacking so far.Very recently, we successfully generated Ni-modified PDC monoliths exhibiting catalytic activity by vat photopolymerization, [28] showing a clear difference in the catalytic performance between powdered and monolithic material: the accessibility of active centers appeared to be significantly hindered in the monolithic Ni-SiOC components, thus requiring crushing of the material for substantial catalytic conversion.The same observation was made for the hierarchically porous materials generated in this present work.The oxidative post-treatment of the monolithic catalyst and the resulting generation of the first monolithic PDC catalyst material with substantial catalytic activity thus presents an enormous step toward the applicability of these materials.When comparing the activity to the performance of the previously mentioned powdered materials, the CO 2 conversion of 48% is at the same level or even higher than for powdered Ni-SiOC materials (a comparison given in Table S3, Supporting Information).Furthermore, a vast improvement in selectivity of the material toward CH 4 can be achieved in comparison to previously reported SiOC-based materials, thus bringing in situ formed PDC catalyst materials on eye level with conventionally produced catalysts in terms of selectivity. [69]The last decisive factor for the applicability of a catalyst is its long-term stability: here, the monolithic samples generated within this work show improved performance in selectivity during long-term use compared to SiOC materials decorated with active catalyst centers by conventional impregnation techniques. [72]The high stability of the in situ formed metal nanoparticles is further highlighted by the limited change in crystallite size, with only a minor increase from 7.6 to 10.8 nm over the 50 h reaction period at 400 °C, indicating the anticipated limitation in sintering phenomena in these environmental conditions.
Methods: Cloud-Point Titration: An analytical scale and pipettes were used to carry out the cloud-point titrations.The solutions were homogenized by magnetic stirring.As solubility and phase separation are temperature-dependent, a constant temperature of 25 °C was ensured using a water bath between component additions.All substances were stored in this water bath to temper them prior to the experiment.Data points within the miscibility diagram were generated by alternating the drop-wise addition of TEOS and TMPTMA to the compositions located on the phase boundary obtained from the fixed compositions approach.This way, the solution became clear upon the addition of TEOS and turbid again when enough TMPTMA was introduced.
The miscibility gap was located by a perpendicular movement away from the MK-TPM axis toward the bottom right corner of the pseudoternary miscibility diagram (Figure 1B).Once the appearance of turbidity was recorded, the addition of TMPTMA was stopped.
Resin Compounding: The resin for vat photopolymerization of green bodies was prepared in batches of 10 g each.The light absorber Sudan Orange G (0.055 wt.%) was dissolved in TMPTMA by magnetic stirring.Simultaneously, the PSO powder was dissolved in TPM in a 1:1 weight ratio by magnetic stirring at 40 °C for 1 h.Subsequently, TEOS was stirred into the TMPTMA solution.In a third step, the PSO-TPM solution was added drop-wise, and the mixture was stirred for another 5 min, yielding an orange-colored liquid of low viscosity.Finally, under the exclusion of blue light, 1 wt.% of the photo-initiator BAPO was added and dissolved by ultrasonication for 10 min.For the production of complex parts on the CeraFab printer, 0.6 wt.% proprietary Type I-photoinitiator with an absorption maximum in the near UV (Lithoz GmbH, Austria) was employed.The ratios between the three main components for each respective resin composition are given in Table S7 (Supporting Information).
Metal Modification: To modify the preceramic base polymer (PSO) with the corresponding metal compounds, the required amount of metal acetylacetonate was weighed into a dry round-necked flask under inert nitrogen atmosphere in a glove box.The amount of metal acetylacetonate was selected to yield a concentration of 5 wt.% and 20 wt.% elemental metal in the final ceramic, respectively, expecting a ceramic yield of 80 wt.%.After removing the diaphragm-covered flask from the glove box, the solvent (THF for Ni and Mo; DCM for Co; PA for La) was added with a syringe.The weight ratios of precursors to solvent were 1:5 for Ni, Co and Mo.A 1:1 ratio of propionic acid to PSO was used for the modification with La.The metal acetylacetonate solutions were stirred for 10 min to dissolve the metal precursor.Subsequently, PSO was added to the solution under air to start the reaction.After a reaction time of 2 h at room temperature, the solvent was removed by rotary evaporation (40 °C, 25 mbar) for Ni, Co, and Mo.The La-modified PSO was used in solution, and the propionic acid was likewise employed as the printing solvent, as opposed to TPM used in the other systems.The Ni-, Co-and Mo-modified PSO were dissolved in TPM (using a 1:1 mass ratio) and stirred for 30 min at 40 °C.
The modified printing resins were compounded using the same compositional ratios as used for resin 1 of the unmodified system (Table S7, Supporting Information): 44 wt.% of modified PSO-TPM solution (or, alternatively, PSO-propionic acid for the La-modified system) was mixed with 12 wt.%TEOS and 44 wt.% TMPTMA. 1 wt.% each of Genorad 21 and Genorad 16 were added to stabilize the printing resin against premature gelation.Prior to printing, 1 wt.% of BAPO was added under the exclusion of blue light and dissolved at 30 °C in an ultrasonic bath for 10 min.
Additive manufacturing: A 3D desktop printer (Anycubic Photon Mono 4K) was used for vat photopolymerization of simple geometries from unmodified resins and for the generation of both simple and complex structures from metal-modified resins (printing parameter are given in Tables S4 and S5, Supporting Information).The light exposure was realized with a monochrome LCD screen with LEDs at a wavelength of 405 nm.The resin vat was equipped with a FEP foil (ELEGOO, FEP 2.0 Release Film Liner, 0.127 mm thickness).In contrast, a CeraFab 7500 printer (Lithoz GmbH) was used to generate parts of highly complex geometry from the unmodified resin 1 (printing parameters are given in Table S6, Supporting Information).In this case, a DLP projector with a wavelength of 460 nm was utilized.For both printers, the defoamed resin was poured into the vat before the printing process.The exposure conditions were adjusted based on individual resin characteristics, determined by light exposure tests.After printing, the cured structures were gently detached from the printing platform using a razor blade.The components were then ultrasonically cleaned for 10 min in TPM for the unmodified as well as for Ni-, Co-, and Mo-modified systems.Isopropyl alcohol (IPA) was employed for cleaning the La-modified parts.Subsequently, residual uncured resin and the remaining solvent were removed by compressed air.The cleaned structures were then carefully dried and stored in a desiccator until pyrolysis was carried out.
The structures were pyrolyzed at maximum temperatures ranging from 600 to 800 °C for 2 h.Prior to pyrolysis, samples were thermally cured at 130 °C (heating rate: 0.5 K min −1 ) for 12 h in N 2 .The pyrolysis temperature profile consisted of several distinct heating segments to ensure complete conversion or decomposition of individual phases, respectively, as well as the formation of porous structures (general heating rate: 0.5 K min −1 , cooling rate 2 K min −1 ).Details can be found in the Supporting Information.
The oxidative post-treatment of the Ni-modified samples was likewise conducted in the tube furnace.The pyrolyzed specimens were treated at 440 °C (heating rate 1 K min −1 , 4 h dwell time) in flowing synthetic air (0.6 L min −1 ).Cooling was carried out at a rate of 2 K min −1 .
Characterization: Photopolymerization reactions were investigated using a real-time IR-photorheology setup, as described by Gorsche et.al., [50] to obtain information on curing properties and on time-resolved double bond conversion.The setup consisted of a Bruker Vertex 80 FTIR spectrometer equipped with a near-infrared (NIR) optic, coupled with a rapid scan module and an Anton Paar MCR302 WESP rheometer.The measurement parameters were selected to simulate the conditions during vat photopolymerization.A broadband Exfo OmniCure 2000 UV-light source, emitting between 270 and 550 nm with an intensity of 6 mW cm −2 , was chosen in combination with an external MCT IR detector (Burker Dig-iTech).The rheometer was equipped with a steel parallel-plate measurement system with a diameter of 25 mm.The sample thickness was set to a gap of 0.05 mm between the plate and the optical window.All measurements were carried out at room temperature.Before irradiation, the respective resin samples were sheared with a strain of 1% and a constant shear rate of 1 Hz.To adjust the exposure time for the printer, the curing depth was determined at varying exposure times on the desktop printer using a sliding gauge.For polymer characterization, attenuated total reflectance Fourier-transform infrared spectroscopy (ATR-FTIR) was conducted on a Bruker Tensor37 with a platinum diamond ATR unit.The spectra were recorded from 400 to 4000 cm −1 at a resolution of 4 cm −1 .Thermogravimetric analysis (TGA) and differential thermal analysis (DTA) were conducted using a Netzsch STA 449 C instrument.A heating ramp of 5 K min −1 from 30 to 1500 °C under flowing Ar (50 mL min −1 ) was selected for all measurements, the sample amount being 30-40 mg.In addition, ceramic yield and the linear shrinkage were determined with an analytical balance and a sliding gauge, respectively.Powder X-ray diffraction (XRD) was performed with a PANalytical XPert Pro MPD using CuK 1,2 -radiation at a diffraction angle 2Θ between 5°and 100°.Morphology and macrostructure of printed and pyrolyzed structures were observed with a digital microscope (Keyence VHX-5000).Scanning electron microscopy (SEM) images of surfaces and fracture surfaces were recorded using a FEI ESEM Quanta 200 FEG instrument.Transmission electron microscopy (TEM) measurements were recorded using a FEI TECNAI F20 instrument equipped with a field emission gun working at 200 kV acceleration voltage.Scanning transmission electron microscope (STEM) images were recorded using a Fishione HAADF detector controlled by the Gatan DigiSTEM II system.TEM images were recorded using a Gatan Rio 16 camera.Energy dispersive Xray spectroscopy (EDX) mappings were recorded using an EDAX-AMETEK Apollo XLTW SDD detector (energy resolution 132 eV).Specific surface areas of the materials were determined by N 2 adsorption at −196 °C according to BET theory using 5 measurement points (ASAP 2020, Micromeritics).BJH calculations were used for the determination of pore size in the mesopore range, assuming a cylindrical pore geometry.The samples were degassed in vacuum for 4 h at 300 °C prior to measuring.Pore opening diameters and pore size distributions were evaluated by mercury intrusion porosimetry (Pascal 140/440, ThermoScientific) with a maximum applied intrusion pressure of 400 MPa.Carbon and oxygen content of pyrolyzed materials were determined using combustion analysis (Leco C-230) or the hot fusion method (Leco TC-400), respectively.
Testing of the catalytic activity during CO 2 methanation was performed in a self-assembled fixed-bed steel reactor with an inner diameter of 6 mm at atmospheric pressure.0.5 g of the monolithic or powder material were pre-treated at 400 °C for 4 h in a 25 mL min −1 gas flow consisting of 10 vol% H 2 in He.The catalytic activity was evaluated using a feed of 25 mL min −1 (containing 10 vol% CO 2 , 40 vol% H 2 and 50 vol% He) at temperature steps of 50 °C between 200 and 400 °C, with a dwell time of 2 h at each temperature step.The composition of the outlet gases was analyzed by gas chromatography using an Inficon Micro GC Fusion instrument.

Figure 1 .
Figure 1.A) Schematic of the vat photopolymerization process.B) Pseudo-ternary miscibility diagram of preceramic resin constituents obtained by cloud-point titration.The phase-separated region (i.e., the miscibility gap) is indicated as filled green area.Points 1-4 indicate the composition of resins used for additive manufacturing of porous PDCs.C) Schematic of the molecular/microstructural situation in the homogeneous preceramic printing resins, after photopolymerization-induced phase-separation during additive manufacturing, and after pyrolytic conversion with simultaneous polymer burn-out.

Figure 2 .
Figure 2. Results of photorheology experiments for the different resin compositions, over the course of curing: A) Storage modulus.B) Double bond conversion calculated from NIR signals.C) Normal force.

Figure 3 .
Figure 3. A-H) Comparison of porous structures obtained from different resin compositions after pyrolysis at 700 °C in Ar (top: sample surfaces; bottom: fracture surfaces).

Figure 5 .
Figure 5. Metal-modification of preceramic PSO: A) Schematic of the metal modification reaction: Formation of Si─O─M bonds and partial cross-linking of the polymer.B) FTIR spectra of PSO before and after modification reaction with Ni, Co, Mo, and La, characteristic Si-O-M band at 930 cm −1 attributed to the modification reaction.

Figure 6 .
Figure 6.Results of photorheology and curing experiments for resin 1 modified with Ni, Co, La, and Mo: A) Storage modulus.B) Double bond conversion calculated with NIR signals.C) Normal force, over the course of curing.D) Curing depth as a function of exposure time on the printer.
Figure 9A-D depicts TEM images of the pyrolyzed Ni, Co, Mo, and La modified SiOC ceramic.Within the Ni-and Comodified ceramics, segregated metal nano-particles are visible.In both cases, bimodal particle size distributions are evident (Figure 8A,B), with Ni and Co particles being present in two size categories; in addition to the larger particles at a size of 10-50 nm (Ni) or 5-15 nm (Co), another category of particles sized in the range of 3 nm (Ni) or 2 nm (Co), respectively, is evident.

Figure 8 .
Figure 8. Transmission electron microscopy investigations of metal phases within printed metal-modified SiOC pyrolyzed at 700 °C in Ar.A-D) TEM images of Ni-, Co-, Mo-, and La-modified SiOC.E-H) STEM-HAADF images of Ni-, Co-, Mo-, and La-modified SiOC.I-L) Elemental distribution of Ni, Co, Mo, and La within metal-modified SiOC (EDX maps).

Figure 9 .
Figure 9. Catalytic activity of pyrolyzed Ni-SiOC materials for a CO 2 methanation model reaction.A) CO 2 conversion as a function of reaction temperature for monolithic and powder samples with 20 wt.% Ni loading, pyrolyzed at 600 or 700 °C.B) Long-term stability of 20 wt.% Ni-SiOC monolithic structures after oxidative post-treatment during CO 2 methanation testing (reaction temperature 400 °C).

Table 2 .
Specific surface area, average pore size, and median pore opening diameter of metal-modified SiOC materials derived from printed phaseseparating resins after pyrolysis at 700 °C in Ar.