Ceramic Open Cell Foams Featuring Plasmonic Hybrid Metal Nanoparticles for In Situ SERS Monitoring of Catalytic Reactions

This work presents porous zirconia‐toughened alumina ceramics functionalized with Au@Pd/Au@Pt core–shell nanoparticle (NP) for in situ monitoring of catalytic reactions via surface‐enhanced Raman scattering (SERS) which is augmented by the open cell foam structure of the ceramic support. In this respect, the porous ceramic enables efficient light trapping and propagation onto the coated surface, which provides good accessibility of the catalyst, while the core–shell particles are equipped with a catalytically active shell and a plasmonic core which enables SERS sensing. The metallic hybrid core–shell NPs are synthesized by the Au‐seed mediated method and colloidally deposited onto the open porous ceramic matrix prepared via the polymer replica method. The Au@Pt NP functionalized porous ceramic show a Raman enhancement factor up to 106, which is significantly higher than that of non‐porous samples. In situ reaction monitoring via SERS is demonstrated by the Pt‐catalyzed reduction of 4‐nitrothiophenol to 4‐aminothiophenol, showing high specificity for analysis of reactants and products. This multifunctional material concept featuring ceramics‐augmented SERS and catalytic activity could be extended beyond real‐time, sensitive reaction monitoring toward high temperature reactions, photothermal catalysis, bioprocessing and ‐sensing, green energy conversion, and related applications.


Introduction
Porous ceramics are well-established as supports for metal nanoparticle catalysts, where they are required to maintain spatial distribution and to facilitate product separation from the catalytic nanoparticles. [1][2][3][4] However, ceramic catalyst carriers usually complicate in situ reaction monitoring, because they block and/or scatter electromagnetic radiation at a wide range of wavelengths. This problem could be solved by specifically designing the porous support, [5,6] so that it can utilize scattering processes to enable or even enhance Raman scattering-based sensing of reactants.
Ideal substrates for metal nanoparticle catalysts should feature chemical and thermal stability, high surface area and large open porosity providing high accessibility of the substrates. In this context, ceramics have obvious advantages over metal-and polymer-based materials in terms of mechanical and thermal stability, resistance against corrosion and chemical erosion. Additionally, ceramics can be processed with a wide range of porous structures with low density, high mechanical stability and high specific surface area for 3D catalyst impregnation. In some instances, ceramics like alumina can act not only as a support, but also as a catalyst for reactions such as isomerization, alkylation, catalytic cracking and hydroforming. [7] Of different crystal structural phases (e.g., alpha, beta and gamma), alpha-alumina has the lowest catalytic activity and is mostly processed with a relatively low specific surface area as opposed to the "activated" alumina phases that exhibit meso-and microporosity along with a high abundance of active sites. [8] Here, we focus on passive alpha alumina supports structured as macroporous open cell foams.
In situ metal-catalyzed reaction monitoring has been realized via various techniques, such as UV-vis absorption spectroscopy, infrared (IR) spectroscopy and surface-enhanced Raman spectroscopy (SERS). [9][10][11][12][13][14] While UV-vis absorption spectroscopy provides only limited chemical information, IR spectroscopy functions inefficiently in aqueous systems due to strong IR absorption of water. On the other hand, Raman spectroscopy, especially SERS allows monitoring and investigation This work presents porous zirconia-toughened alumina ceramics functionalized with Au@Pd/Au@Pt core-shell nanoparticle (NP) for in situ monitoring of catalytic reactions via surface-enhanced Raman scattering (SERS) which is augmented by the open cell foam structure of the ceramic support. In this respect, the porous ceramic enables efficient light trapping and propagation onto the coated surface, which provides good accessibility of the catalyst, while the core-shell particles are equipped with a catalytically active shell and a plasmonic core which enables SERS sensing. The metallic hybrid core-shell NPs are synthesized by the Au-seed mediated method and colloidally deposited onto the open porous ceramic matrix prepared via the polymer replica method. The Au@Pt NP functionalized porous ceramic show a Raman enhancement factor up to 10 6 , which is significantly higher than that of non-porous samples. In situ reaction monitoring via SERS is demonstrated by the Pt-catalyzed reduction of 4-nitrothiophenol to 4-aminothiophenol, showing high specificity for analysis of reactants and products. This multifunctional material concept featuring ceramics-augmented SERS and catalytic activity could be extended beyond real-time, sensitive reaction monitoring toward high temperature reactions, photothermal catalysis, bioprocessing and -sensing, green energy conversion, and related applications.
of reaction processes by detecting specific molecular vibrational modes of reactants, intermediates and products at different reaction time spots, providing direct information of the chemistry of the analyzed reaction. [15][16][17][18] SERS uses plasmonic nanostructures to excite the localized plasmon resonance in metallic nanoparticles, which leads to significantly increased Raman intensity. [19] The most commonly used plasmonic particles are noble metal NPs, such as gold (Au) and silver (Ag), because of their oxidation resistance and pronounced localized plasmon resonance. [20] However, most of the reported SERS substrates are based on 2D assemblies, which have been demonstrated to provide very high Raman enhancement, [21,22] but are difficult to adapt to more complex setups.
Li et al. have presented 3D plasmonic nanostructures as high-quantum-efficiency UV photodetectors based on anodic aluminium oxide (AAO) matrix substrates, where the light confinement was spatially extended from 2D to 3D. It was reported that the incident light is concentrated by the Au nanostructures and the strong optical interference within the AAO matrix, leading to light trapping followed by a high Raman enhancement factor (EF). [23] Polymeric materials were also reported as 3D SERS substrates. For example, 3D pyramid polymethyl methacrylate (PMMA) functionalized with Ag NPs exhibited EFs up to 10 8 . [24] Ag NPs@MoS 2 /PMMA polymer composites showed a detection limit of 10 −10 m of the melamine standard solution. [25] Other examples include plasmonic 3D photonic crystals, 3D plasmonic nanoantenna based on multielectrode arrays, 3D plasmonic nanorod forests on paper or 3D SERS substrates based on graphene. [26][27][28][29] However, current research on 3D plasmonic SERS substrates focuses mostly on the nano-scale structure of the plasmonic particles and their supports, [30,31] not on macroscopic technical materials.
While being able to catalyze a wide range of chemical reactions, the group VIIIB transition metals palladium (Pd) and platinum (Pt) exhibit a high imaginary part of the complex permittivity and therefore only generate weak SERS EFs due to dampening of the plasmonic oscillation. [32] To make up for the weak EFs, Pd/Pt-based SERS substrates usually demand perfect ordering and uniform deposition with high surface coverage, that also consider the size, shape, geometry, and position of the metallic nanostructures. [33][34][35][36][37][38] Conversely, while conventionally used as excellent SERS-active metals, Au NPs have a size limit of ≈5 nm for efficient catalytic activity, which is too small for SERS-activity. [39,40] In this context, the design of metallic hybrid nanostructures which enable the combination of various properties of the different metals in one structural unit were reported. [12,13,[41][42][43][44][45][46] Particularly, metallic hybrid NPs consisting of Au in combination with Pd or Pt can lead to both plasmonic and catalytic activity in a single core-shell particle. For example, Hu et al. prepared Au core/Pd shell (Au@Pd) NPs with a size from 35 to 100 nm by the Au-seed mediated method. The Au@Pd NPs distributed on a Pt electrode surface showed Raman EF up to 2.7 × 10 3 . [41] Xie et al. presented the synthesis of bifunctional Au/Pt/Au nanoraspberries for in situ monitoring of platinum-catalyzed reactions. [13] The synthesized raspberrylike Au/Pt bimetallic NPs have two exposed metal surfaces, resulting in both SERS and catalytic activity in a single bifunctional unit. However, most of the current research on metallic hybrid NPs for molecule detection and reaction monitoring via SERS focused on synthesis, characterization, and dispersions of metallic hybrid NPs. [12,13,[41][42][43][44][45] In our own previous work, [5] we have introduced the concept of plasmonic porous ceramics-based zirconia-toughened alumina (ZTA) functionalized with Ag NPs that combine the optical properties of plasmonic nanostructures with advantages of ceramics, such as stability at extreme temperatures and pressures, chemical inertness, dielectric properties and biocompatibility. [47][48][49][50][51] In these experiments, silver (Ag) NPs were deposited onto the open porous ceramic matrix through chemical reduction. Processing of the ceramic open cell foam can be easily upscaled and tailored for various technical applications. [52] Using macroscopic ceramic open cell foams as substrate one could realize high accessibility by light, which lead to efficient measuring from all directions. Most strikingly, the porous ceramic structure showed significantly higher EFs than those of non-porous structures, which is most likely caused by light trapping and multiple scattering within the porous ceramic structure. [23] This additional enhancement effect resulted in efficient molecule detection via SERS despite the unordered deposition of the plasmonic particles on the heterogeneous ceramic matrix.
In this work, we fully realize this new material concept by preparing functionalized ceramic open cell foams with metallic hybrid core-shell NPs (Au@Pd, Au@Pt). The core-shell hybrid NPs are designed to exhibit a highly SERS-active Au core and a catalytically active Pd/Pt shell with a large surface area. The particles are prepared with various shell thicknesses via the Au-seed mediated method [42] and colloidally deposited on the ceramic matrix after particle synthesis. The zirconiastabilized alumina matrix of the porous ceramic acts as the catalyst support, which is designed to augment the SERS-based reaction monitoring through light trapping and scattering which is characterized by the diffuse reflectivity of the sample. The general material concept was tested by analyzing the wellknown reaction of 4-nitrothiophenol (4-NTP) to 4-aminothiophenol (4-ATP).

Results and Discussion
In this work, metallic hybrid core-shell NPs consisting of an Au core and a Pd or Pt shell have been synthesized. The shell thickness can be manipulated by varying the amount of the corresponding reactant. The Au core was synthesized by chemical reduction of HAuCl 4 with TSC. In a second step, the Pd/Pt shell was reduced on the surface of the Au core by the remaining surface-adsorbed TSC (with the addition of L-ascorbic acid in the case of Pt). In this seed growth mechanism, the Au seed is likely involved in catalyzing the reduction of PdCl . [57] Scanning electron microscopy (SEM) images of the ceramic surface functionalized with Au@Pd hybrid NPs and transmission electron microscopy (TEM) images of the synthesized Au@Pd particles with different amounts of the palladate precursor are shown in Figure 1. The TEM images reveal increasing size of the core-shell structure of the NPs with increasing precursor concentration with diameters of 42.05 ± 3.29, 54.01 ± 4.58, 75.94 ± 3.28, and 93.18 ± 3.62 nm, corresponding to palladate reaction concentrations of 0.1, 0.2, 0.8, and 1 mm, respectively. At the highest precursor concentration, a rough shell becomes visible (Figure 1d inset). Since pure Au particles prepared with this method show sizes of 39.1 ± 2.13 nm, the change of the core-shell particle size indicates the change of the shell thickness. On the SEM images, the dark grey background represents alumina, the lighter grey areas the zirconia grains, which provide higher mechanical strength to the intricate porous ceramic structure. [58] The smallest bright dots are the metallic hybrid NPs. The resulting metallic hybrid NP coverage on the ceramic surface was slightly above 20% of the total surface area. Figure 2 shows SEM micrographs of ZTA surfaces functionalized with Au@Pt NPs and TEM images showing equivalent metallic hybrid NPs from the reaction medium. The deposited Au@Pt NPs have a diameter of 44.32 ± 0.57, 48.11 ± 1.93, 50.84 ± 4.7, and 70.78 ± 5.58 nm corresponding to chloroplatinic acid reaction concentration of 0.1, 0.2, 0.8, and 1 mm, respectively. Here, the rough shell structure is already visible at the second lowest concentration of the platinate precursor (Figure 2 insets). In general, Pd shells on the Au core surface reached higher thickness than Pt shells under identical reaction conditions. Based on the postulated seed growth mechanism, [57] the growth of the Pt shell is slower than the Pd shell due to the higher reactivity of the PdCl 4 2− precursor. UV-vis absorption spectra of suspensions of the metallic hybrid NPs show shifting intensity peaks toward higher wavelength with increasing shell thickness, as shown in Figure 3. The spectrum of the 39 nm Au core NPs is given as a reference. The Au NP suspension showed a sharp localized surface plasmon resonance (LSPR) peak at 526 nm, while the 42 and 54 nm sized Au@Pd NP suspensions showed LSPR peaks at 539 nm ( Figure 3a). For the 76 nm sized Au@Pd NPs, the LSPR peak shifted to 549 nm and became broader. In the case of the 93 nm sized Au@Pd NPs, the absorption peak was located at 561 nm and became much flatter. The flattening of the LSPR peak indicates that the damped plasmon oscillations in the Pd shell dominate the plasmon resonance of the NPs, as was also reported for similar hybrid particles. [41] Figure 3b presents the absorption spectra of the Au@Pt hybrid NPs. Due to the smaller Pt shells, the 44, 48, and 51 nm sized Au@Pt NPs only showed slightly redshifted LSPR peaks at 529 nm, along with pronounced broadening for 51 nm shells. When the size of the Au@Pt NPs increased to 71 nm, the absorption curve became much flatter and the resonance peak vanished. This again indicates that the damped plasmon resonance of the Pt shell dominates over the Au core with higher shell thickness. Meanwhile, the width of the Au@Pt NP size distribution increased from 0.21 to 0.58 with increasing size of the particle, indicating more pronounced polydispersity of the Au@Pt NP, which also contributed to flattening of the LSPR peak of the larger Au@Pt NPs. [56] The localized surface plasmon resonance (LSPR) peaks of Au@Pd core-shell NPs were more pronouncedly redshifted than that of Au@Pt NPs. From the literature, the main resonance frequencies of the localized surface plasmon of Pd, Pt, and Au are 0.10, 0.35, and 1.40 eV, [59] respectively, corresponding to 12 398, 3542, and 886 nm. The main resonance of Pd lies at a much higher wavelength than that of Pt, which leads to the more pronounced redshift of the Au@Pd LSPR peaks. In addition, the Pd shell was significantly thicker than the Pt shell, so the damped plasmon oscillations in the Pd shell should more pronounced.

Molecule Detection with Plasmonic Porous Ceramics
Raman spectra of the porous ZTA substrate before (a) and after 39 nm Au nanoparticle functionalization (b) are shown in Figure S1, Supporting Information. The Raman shifts at 1370 and 1400 cm −1 can be assigned to alpha-alumina. The deposition of the metal nanoparticle alone does not affect the Raman spectra of the ZTA substrate. Raman spectra of pure pyridine (a) on a bare ZTA substrate and of a 10 −2 m pyridine droplet on porous ZTA functionalized with 39 nm Au nanoparticles (b) are presented in Figure S2, Supporting Information. The porous sample functionalized with 39 nm Au NPs showed an EF ≈3.3 × 10 6 .
The ability of the porous ceramics functionalized with metallic hybrid NPs for sensitive molecule detection can be determined via SERS using pyridine as reporter molecule. Figure 4a shows the Raman spectra of the reporter molecule deposited on substrates functionalized with differently sized Au@Pd NPs. The Raman peaks at 1012 and 1038 cm −1 can be assigned to the pyridine ring breathing mode (υ 1 ) and symmetric ring deformation (υ 12 ), respectively, enhanced by the LPR of Au. [60] The intensity peak at 1026 cm −1 can be associated to the symmetric ring deformation (υ 12 ) of pyridine adsorbed on Pd, which is reported to occur at 1001-1005 cm −1 . [60] The frequency shift of the pyridine symmetric ring deformation (υ 12 ) can be interpreted based on the vibrational coupling effect. [61][62][63][64][65] The shift in vibration wavenumbers for totally symmetric modes of pyridine can be related to the strength of the adsorption bond between pyridine and the metal cluster. For group Adv. Mater. Interfaces 2023, 10, 2300207   [63] which corresponds to the measured Raman signal at 1026 cm −1 in this work (Figure 4a). Such a slight shift of 4 cm −1 can be explained in terms of quasiharmonic nature of the molecules at ambient temperature. Figure 4b shows the Raman spectra of diluted pyridine on plasmonic porous ceramic substrates functionalized with differently sized Au@Pt NPs. The Raman peak at 1015 cm −1 can be assigned to pyridine adsorbed on the Pt shell. [60] The Raman enhancement factor of the substrates was calculated as described in our previous work [5] using the Raman intensity at 1026 and 1015 cm −1 for Au@Pd and Au@Pt NPs functionalized substrates, respectively, and the intensity of pure pyridine adsorbed on a porous ZTA sample without metal nanoparticles. In accordance to the UV-vis spectra discussed above, the EFs of the substrates functionalized with the metallic core-shell NPs show a decreasing trend with increasing shell thicknesses. The EFs of the substrates functionalized with Au@Pd NPs reduced from 8.0 × 10 4 to 2.8 × 10 4 , when the size of the core-shell particle was enlarged from 42 to 93 nm. However, the EFs are still in the same order of magnitude. The EFs decreased rapidly from 1.5 × 10 6 to 3.0 × 10 5 , when the size of the Au@Pt NPs increased from 48 to 51 nm. At 71 nm, the Au@Pt NPs did no longer show any characteristic Raman peaks of pyridine, indicating that the substrate lost the Raman enhancement effect.
Using transition metals as SERS substrates, the SERS activity of pure transition metals was reported to be rather weak. [60] Through the thin layer of the transition metal on top of a SERS-active Au core, the high SERS enhancement can work from a certain distance away from the core via the longrange effect of the electromagnetic field. [66][67][68][69] According to the electromagnetic mechanism of the Raman enhancement effect, [20] the closer the probe molecule to the SERS-active nanostructure, the higher the Raman enhancement. Accordingly, the thinner the NP shell, the closer the reporter molecule is placed to the Au core and the higher the Raman EF. The ceramic substrates deposited with Au@Pd NP show EFs in the order of 10 4 . In comparison, the EFs are much higher, up to 10 6 , in the case of substrates functionalized with Au@Pt NPs. However, the effective permittivity of hybrid metal nanoparticles depends not only on the size of Au core but also the surrounding layer. In the case of metallic hybrid NPs, the metal with interband electron transitions at lower frequency dominates the plasmon response. [32] Since the maximum quality localized surface plasmon frequencies of Q LSP, Pd and Q LSP, Pt is significantly lower than that of Q LSP, Au , [59] the Pd and Pt shells dominate the plasmon response. Therefore, the Au@Pd and Au@Pt NPs showed generally weaker SERS enhancement than conventional Au and Ag NPs. Figure S3, Supporting Information, shows the Raman spectrum of pyridine on plasmonic porous ceramic functionalized with 48 nm Au@Pt nanoparticles, measured with a 532 nm laser. The Raman EF is ≈1.2 × 10 4 . In comparison, the sample measured with the 633 nm laser (Figure 4b, red spectrum) showed an EF ≈1.5 · 10 6 . Although the Au@Pt core-shell particles exhibit an absorption peak ≈529 nm, the 532 nm laser as excitation source provided weaker Raman EFs than the 633 nm laser. This phenomenon can be explained by the redshift of the local enhancement field caused by plasmonic hot spots, which has been experimentally demonstrated and theoretically calculated by several studies. [20,81,82]

Catalytic Function of the Plasmonic Porous Ceramics
The catalytic activity of the metallic core-shell NPs on the porous ceramic support was investigated with the reduction of 4-NTP by sodium borohydride (NaBH 4 ) to 4-ATP with trans-4,4-dimercaptoazobenzene (trans-DMAB) as intermediate at room temperature, which can be catalyzed both by Pd and Pt ( Figure 5). [40] The reaction was first analyzed via UV-vis spectroscopy. Here, the reactant 4-NTP shows an absorbance peak at ≈412 nm, while the product 4-ATP is colorless and shows no absorbance peak. By observing the intensity change of the absorbance peak at 412 nm, one can easily interpret the change of the 4-NTP concentration and with that the reduction process of 4-NTP to 4-ATP. In this context, we assume that the ZTA substrate is completely inert and does not take part in the chemical reaction. [7,8] Adv. Mater. Interfaces 2023, 10, 2300207  Figure 6 shows the absorption spectra of the reduction of 4-NTP using functionalized substrates with differently sized Au@Pd NPs as catalyst. With 42 and 54 nm Au@Pd NPs, the 4-NTP absorbance peak at 412 nm could still be observed after 30 min reaction time, indicating that the reduction was not complete. With the 76 or 93 nm sized Au@Pd as catalyst, the absorbance peak at 412 nm disappeared after 2 min, suggesting 4-NTP has been reduced to 4-ATP.
It is worth noticing, that the substrate functionalized with 54 nm Au@Pd NPs showed a faster reaction rate than the substrate functionalized with 42 nm NPs within the first 20 min reaction time. Although the absorbance peak of the reactant 4-NTP seems to be weaker on the 54 nm functionalized sample after 30 min reaction time, the difference was not significant and within the standard deviation of the data (listed in Table S1, Supporting Information).   more surface area of the catalytically active metal for more efficient catalysis. Figure S4, Supporting Information, shows the absorption spectra of the reduction of 4-NTP using Au nanoparticle functionalized sample as substrate at the initial point and after 30 min reaction time. After 30 min reaction time, the spectrum of the reactant 4-NTP was not significantly changed, which demonstrates that the bare Au-NPs do not show catalytic activity.

Comparing Porous and Dense Ceramic Substrates
Analogous to our previous work with Ag NPs, [5] we detected high EFs of the porous substrates functionalized with the herein described core-shell particles. To highlight the additional enhancement from the porous structure, Figure 8a shows Raman spectra of pyridine on porous (red) and dense (grey) substrates, each functionalized with 48 nm Au@Pt. The dense substrate had the same particle coverage and ceramic matrix as the open cell foam but was processed into a compacted pellet instead. In this case, the porous sample delivered around seven times higher EF than that of the dense sample, which corresponds to our previous findings. [5] The EF was ≈10 times higher when using 4-NTP as the reporter molecule. Figure 8b shows the hemispherical reflectance of bare ZTA (dashed spectra) and ZTA functionalized with 48 nm Au@Pt NPs (line spectra) in both dense (grey) and porous (red) form. The bright white porous ZTA substrate showed an effective reflectance back into the integrating sphere between 20% and 30% at wavelengths above 500 nm. The remaining light is either transmitted through the structure, reflected toward the sides or adsorbed by the ceramic matrix. For comparison, an unfunctionalized dense ZTA pellet had a total reflectivity of about 85%. It is noticeable that unfunctionalized porous ZTA shows increasing reflectance with increasing wavelength, which might be related to the light propagation inside the semitransparent ceramic matrix. A similar behavior has been modeled by Li et al. via Monte Carlo ray tracing radiative transfer analysis of an alumina open cell foam. [70] After functionalization with 48 nm sized Au@Pt NPs, the reflectance of the open cell foam was reduced by roughly 80% to below 5% over the entire measured wavelength range (including the wavelength of the incident Laser of the Raman spectrometer at 633 nm), which also manifests in the dark grey color of the material. For comparison, the reflectivity of the dense pellet was only reduced by 40%. This indicates that the previously reflected light was efficiently adsorbed by the deposited NPs inside the porous structure. Relating to the adsorption spectrum of the 48 nm sized Au@Pt NPs (Figure 3b red spectrum), no obvious feature at the adsorption maximum of the core-shell particles at 529 nm was observed in the reflectivity spectra of the functionalized sample which might indicate a broadening of the absorption spectra of the adsorbed particles compared to the dispersions discussed above.
Likewise, the catalytic activity of dense ZTA pellets was compared to the porous substrates, which is shown in Figure S5  with hardly a change in reactant concentration. After 30 min, the absorption intensity of the reaction 4-NTP was decreased by 5.9%, while it decreased by 22.5% for the porous substrate. Accordingly, the functionalized porous substrate showed better catalytic activity, which agreed with our expectation, since the porous ceramic offered higher specific surface area for deposition of the core-shell particles.

In Situ Reaction Monitoring
After SERS enhancement and catalytic activity of the plasmonic porous ceramic have been investigated, these two aspects can be combined for in situ reaction monitoring via SERS. Figure 9 shows the in situ reaction monitoring via SERS at different time points using porous ceramics with 48 nm Au@Pt NPs as a representative case study.
At t = 0 min, the Raman intensity peaks at 1108 and 1572 cm −1 correspond to the CN stretching and phenylring CC stretching of 4-NTP, respectively (grey dashed line). At t = 2 min, the appearance of the peak at 1142 cm −1 indicates the formation of the reaction intermediate trans-DMAB (red dashed line). The relative peak intensity at 1108 cm −1 is also much weaker than that at t = 0 min, indicating the reduction of the reactant 4-NTP. At t = 20 min, we can observe the appearance of a new peak at 1595 cm −1 , which can be assigned to the phenyl-ring CC stretching of 4-ATP (blue dashed line). Thus, at t = 30 min, the concomitant decrease and increase of the peak intensity at 1572 and 1595 cm −1 can be understood as the reduction of 4-NTP to 4-ATP within the experimental setup. Should the differential Raman cross section of the respective vibrational modes be known, the associated kinetics can be studied, as well.

Conclusions
In this work, plasmonic porous ceramics were functionalized with metallic hybrid core-shell NPs through colloidal deposition during reduction of the respective precursor salts. The substrates deposited with Au@Pd NPs showed Raman EFs of ≈10 4 , while the substrates functionalized with Au@Pt NPs delivered EFs up to 10 6    The open porous ceramic structure allows possible application in complex 3D configurations and measuring in all directions. Moreover, the plasmonic porous ceramic showed significantly higher EFs than comparable non-porous samples, due to efficient absorption and scattering of light within the porous ceramic structure.
Accordingly, we were able to demonstrate that the bifunctional plasmonic porous ceramics can be applied for catalytic reaction monitoring via SERS. With Raman EFs up to 10 6 , the reaction monitoring of plasmonic porous ceramics enables high sensitivity through electromagnetic and chemical enhancement and a high chemical specificity by detecting Raman active vibrational modes of each single molecule during the reaction. [71][72][73] This material concept could be further adapted for bioprocessing and -sensing, green energy conversion and related applications. Moreover, the plasmonic nature of core-shell particles in combination with the inert and temperature-stable ceramic matrix could be utilized for high-temperature reactions up to 600 °C, [2,[74][75][76] while localized plasmonic heating via the photothermal effect could be applied for photothermal catalysis or similar methods. [77][78][79][80]
Processing of Plasmonic Porous Ceramics: The ZTA porous ceramics were fabricated by the polymer replica method with recoating according to previous work. [5,53] In short, the 30 ppi (pore per inch) PU foams were cut into 15 × 15 × 10 mm small blocks and fully submerged in 80 wt% ZTA ceramic slurry. Excess slurry in the foam was squeezed out manually. After drying, the coated PU foam was coated again with 60 wt% ZTA ceramic slurry. Blocked pores were cleared carefully with compressed air. After complete drying, the coated PU foam was sintered with a multi-step sintering program in a furnace (Nabertherm HT04/17, Nabertherm GmbH, Germany). The furnace was heated with a rate of 1 °C min −1 to 110, 250, and 600 °C, with dwelling times of 2, 3, and 3 h, respectively. After burning out of the PU templates, the porous ceramics were sintered at 1650 °C for 3 h with a heating rate of 3 °C min −1 and cooled to room temperature within 8 h.
The deposition of the core shell hybrid plasmonic NPs was realized by the established chemical reduction of noble metal ions. In a 100 mL 3-neck round flask, a porous ceramic sample was hung with a thin thread and submerged in 14.705 mg 45 mL TSC solution. The reaction solution was stirred and heated. When the temperature reached 100 °C, 5 mL 10 mm chloroauric acid was added dropwise to the reaction solution.
The reaction solution turned red rapidly. To functionalize these particles with a shell of Pd, yielding Au@Pd NPs, 1.632/3.264/13.057/16.322 mg potassium tetrachloropalladate were added to reaction solution (K 2 PdCl 4 reaction solution concentration 0.1, 0.2, 0.8, and 1 mm), respectively after 10 min stirring. The reaction solution was stirred and kept at 100 °C until the color changed from red to dark grey. [41,54,55] In order to functionalize the gold particles with a shell of Pt, yielding Au@Pt NPs, 24.39/48.79/195.12/243.93 µL 8 wt% chloroplatinic acid solution was added to the reaction solution (H 2 PtCl 6 reaction solution centration 0.1, 0.2, 0.8, and 1 mm), respectively after 10 min stirring when the Au core NPs were formed. Subsequently, 8.81/17.62/70.48/88.1 µg L-Ascorbic acid was added to the reaction mixture, respectively. The temperature was held at 100 °C for about 1 h until the color changed from red to dark brown. [42,43,56] After the color changed to dark grey or brown, respectively, the porous ceramic sample was taken out of the reaction flask, flushed with DI water and dried in an evacuated desiccator.
Characterization: For scanning electron microscopy (SEM), the ceramic samples functionalized with metallic hybrid NPs were cut into small fragments (≤2 mm) and glued on a clean sample holder. The samples were investigated with SEM (Supra 40, Carl Zeiss, Germany) at 1.00 kV with a working distance between 1.9-3.1 mm. The NP surface coverage was calculated via analyzing the SEM images using the software ImageJ. For transmission electron microscopy (TEM), the metallic hybrid NP suspensions were highly diluted and dried on a Formvar coated copper film. The dried NPs were analyzed by TEM (EM 900, Zeiss, Germany). The sizes of the metallic NPs were determined by analyzing the TEM images using ImageJ and the polydispersity was assessed by Gaussian fitting of the size distribution.
UV-vis spectra were taken with a Multiskan GO microplate spectrophotometer (Thermo Scientific, Germany). For the absorption spectra of the metallic hybrid core-shell NPs, the measured suspensions were taken from the reaction solution in section 2.2 after deposition of the core-shell NPs. For monitoring the reduction of 4-nitrothiophenol (4-NTP), plasmonic porous ceramics were submerged in 5 mL of 0.5 mm 4-NTP. With the addition of 5 mL 2.5 mm NaBH 4 to the 4-NTP solution, a timer was started. UV-vis spectra at different reaction time points were recorded.
Raman spectra were recorded on a LabRam ARAMIS (Horiba Jobin Yvon) micro Raman Spectrometer equipped with laser working at 633 nm and less than 20 mW. The use of a 10× objective (Olympus) with a numerical aperture of 0.25 provides a focus spot of about 2.6 µm diameter. Raman spectra were collected in the range of 800 to 1200 cm −1 for pyridine, 900 to 1300 cm −1 and 1450 to 1700 cm −1 for the reduction of 4-NTP respectively, with a spectral resolution of ≈2.1 cm −1 using a grating of 1800 grooves mm −1 and a thermoelectrically-cooled CCD detector (Synapse, 1024 × 256 pixels). The obtained spectra were analyzed with the LabSpec 5 suite and processed with the Baseline function (Type: lines, Degree: 4, Attached: yes, Autofitting).
The hemispherical reflectance was acquired with a VERTEX 80v (Bruker GmbH, Germany) FTIR spectrometer equipped with a Ø 7.5 cm white polytetrafluoroethylene sphere and a Si-Diode 2.4 mm detector for the visible spectrum. Three consecutive measurements from 450 to 1100 nm were performed at ambient temperature and 2.5 mbar and averaged. The reflectivity was calibrated against the Spectralon Diffuse Reflectance Standard 99% (Labsphere, Inc., USA).

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.