Integrated Photothermal Nanoreactors for Efficient Hydrogenation of CO2

To alleviate the energy crisis and global warming, photothermal catalysis is an attractive way to efficiently convert CO2 and renewable H2 into value-added fuels and chemicals. However, the catalytic performance is usually restricted by the trade-off between the dispersity and light absorption property of metal catalysts. Here we demonstrate a simple SiO2-protected metal–organic framework pyrolysis strategy to fabricate a new type of integrated photothermal nanoreactor with a comparatively high metal loading, dispersity, and stability. The core-satellite structured Co@SiO2 exhibits strong sunlight-absorptive ability and excellent catalytic activity in CO2 hydrogenation, which is ascribed to the functional separation of different sizes of Co nanoparticles. Large-sized plasmonic Co nanoparticles are mainly responsible for the light absorption and conversion to heat (nanoheaters), whereas small-sized Co nanoparticles with high intrinsic activities are responsible for the catalysis (nanoreactors). This study provides a new concept for designing efficient photothermal catalytic materials.


Introduction
Converting CO 2 and renewable H 2 into fuels and feedstock chemicals through photothermal catalysis offers a promising way to mitigate the energy crisis and global climate change [1][2][3][4][5][6][7]. This photothermal catalytic technique can use light and heat from the sun, which greatly broadens the utilization range of the solar spectrum [8,9]. Since the seminal research by Ye et al. [10] in 2014, numerous attempts have been made to develop efficient catalysts for photothermal CO 2 conversions [10][11][12][13][14][15]. However, it has been a long-standing challenge to develop high-performance photothermal catalytic systems with broadband optical absorption across the full solar spectrum range and enhanced intrinsic catalytic activities.
Most photothermal catalysts are composed of oxide-supported metallic nanoparticles, which serve as light absorbers and active sites [16,17]. To construct a desirable photothermal catalyst, light absorption capability, light-to-heat conversion efficiency, and intrinsic reactivity are crucial factors to be taken into consideration [18][19][20]. However, for conventional photothermal catalysts, these factors have encountered a trade-off [21]. Sunlight absorption and localized heating can be improved by increasing metal loadings, which often leads to decrease in metallic dispersity and intrinsic catalytic activity. By contrast, the decrease in metal particle loadings results in enhanced dispersity and high atomic utilization of metals, but at the expense of the sunlight-harvesting ability over the entire solar spectrum [22]. Recently, our previous work discovered that with the increase in cobalt loadings, cobalt-decorated silicon nanowires arrays (SNAs@Co) exhibited strong absorption in the near-infrared range [23]. Meanwhile, the further increase of cobalt loadings led to large particle size and low catalytic activity [24]. Therefore, it is highly desirable but challenging to develop new To this end, we demonstrate a confined pyrolysis strategy for the synthesis of an integrated photothermal nanoreactor (IPR) with higher metal loading and dispersity compared with that of traditional supported metal catalysts. The newly developed IPR, characteristic of a unique core-satellite structure, achieves a dramatic synergy effect of the photothermal effects of large-sized plasmonic Co nanoparticles (nanoheaters) and great intrinsic catalytic performance of small-sized Co nanoparticles (nanoreactors). In addition, the spatial confinement effect of a thin layer of SiO 2 is deployed to ensure the stability of satellite Co nanoparticles and provide heat insulation and nanoscale greenhouse effect [12]. The IPR not only inherits the excellent light absorption ability but also exhibits a superior activity in photothermal CO 2 catalysis. The well-designed construction of photothermal catalysis offers a novel strategy for developing high-performance photothermal catalysts via sized engineering for solar fuel production from the greenhouse gas CO 2 .

Preparation of Co@SiO 2
The preparation of zeolite imidazolium framework-67 (ZIF-67) was based on the reported work [25,26]. Specifically, 6.09 g of Co(NO 3 ) 2 ·6H 2 O was dispersed in 210 mL of anhydrous methanol under sonication for 10 min to obtain a precursor solution. Subsequently, 6.93 g of 2-methylimidazole was dispersed in 210 mL of anhydrous methanol. This dispersion was sonicated for 10 min, poured into the precursor solution, and then mixed through magnetic stirring for 10 min. The mixture was kept at room temperature without stirring for a day. The precipitate was collected via centrifugation, washed with water and ethanol several times, and dried at 80 °C to obtain ZIF-67. ZIF-67@SiO 2 was prepared by adding 200 μL of tetraethyl orthosilicate into100 mg of ZIF-67, followed by sonication for 20 s and natural drying. The ZIF-67@SiO 2 was first calcined at 120 °C for 2 h and then at 500 °C for 2 h (both at 1 °C/min) in air. After cooling, the as-obtained samples were reduced at 400 °C in H 2 for 2 h to obtain the core-satellite Co@SiO 2 sample.

Preparation of Co-LNP and Co-SNP
Co-LNP and Co-SNP were prepared through the wetness impregnation method [27]. For Co-SNP, 100 mg of the commercial silica support was dispersed in 5 mL of ethanol and 37.7 mg Co(NO 3 ) 2 ·6H 2 O was added to the above ethanol solution and heated with magnetic stirring until the solution was evaporated to obtain powders. The as-obtained powders were calcined under the same conditions as Co@SiO 2 to obtain Co-SNP catalysts. Meanwhile, Co-LNP was synthesized via the same method except for the amount of 441 mg of Co(NO 3 ) 2 ·6H 2 O.

Characterizations
X-ray diffraction (XRD) was measured on an Empyrean diffractometer with CuKα radiation. Transmission electron microscopy (TEM) images were obtained with an FEI-Tecnai F20 (200 kV) transmission electron microscope. Scanning electron microscopy (SEM) images were acquired on a Zeiss Supra 55 instrument (Carl Zeiss, Germany). Different Co loadings of all samples were measured via inductively coupled plasma source mass spectrometry (ICP-MS). The diffuse reflectance spectra of the different samples were obtained using a Lambda 950 UV/Vis/NIR spectrometer from PerkinElmer and an integrating sphere with a diameter of 150 mm.

Temperature-Programmed Reduction Experiment
The temperature-programmed reduction (TPR) experiment was performed on an automatic chemical adsorption instrument (FINETEC/FINE-SORB-3010). Approximately 10 mg of sample was fixed in a U-shape quartz tube and flushed with Ar (40 mL/min) for 10 min, followed by heating to 300 °C (10 °C/min) for 60 min and then cooling to room temperature in the same Ar flow. Afterward, the sample was exposed to the mixture (5% H 2 /Ar) with a flow rate of 80 mL/min for 20 min at 25 °C for 10 min. Finally, the sample was heated to 600 °C (10 °C/min) in the mixture (5% H 2 /Ar). The temperature and current for the thermal conductivity detector (TCD) were 60 °C and 90 mA, respectively.

Catalytic Measurements
Thermocatalytic CO 2 hydrogenation experiments were conducted in a homemade flow reactor. The reactor was equipped with a circular quartz window (r = 2.75 cm) to allow illumination from the top. The temperature of the reactor was controlled by a homemade heating setup (Fig.  S1). The flow rates of feeding gases were fixed at 4 mL/min for CO 2 , 4 mL/min for H 2 , and 8 mL/min for N 2 . Gas chromatograph (Agilent 7890B) with TCD and flame ionisation detector was used for online analysis of gaseous reactants and product quantities. N 2 in the flow was used as an inner standard. The conditions for the gas-phase photothermal hydrogenation of CO 2 experiments were otherwise identical to those for the thermocatalysis, except that a 300-W Xe arc lamp was used as the irradiation source to illuminate the catalyst without any filter. All catalyst powders (10 mg) were homogeneously dispersed in ethanol and dropped onto a glass fiber filter for the experiment. The response factor of gas reactants and products was calibrated using standard curve methods. The CO 2 conversion rate (R CO2 ), seen as a characteristic of the photothermal catalytic activity, was defined as the sum of the CO formation rate (F CO ) and CH 4 formation rate (F CH4 ), owing to the mere detection of CO and CH 4 in the products. F CO and F CH4 are defined as The selectivity of CO is defined as where n is the yield of products (mmol); ω Co is the loading percent of small-sized Co; m cat is the mass of the catalyst (g), and t is the irradiation time (h).

Results and Discussion
The key to the controlled preparation of the IPR is the encapsulation of ZIF-67 nanocrystals within a silica sheath to realize the discontinuous size distribution [28][29][30][31]. Silica sheath provides the anchoring effect for Co cations at the SiO 2 / ZIF-67 interface, further suppressing the sintering of Co satellites during the pretreatment. Meanwhile, the interior Co cations without the interaction with silica are aggregated into a Co core, which could be attributed to the high surface free energy and low Tamman temperature [32]. Figure 1a illustrates the preparation process of core-satellite structures, which includes the nanocasting of silica onto the surface of ZIF-67 and subsequent calcination treatment.
Experimentally, a straightforward solvent-induced method at room temperature was first used to synthesize ZIF-67 with a rhombic dodecahedral morphology and average edge length of 350 nm (Fig. 1b). Next, the as-obtained ZIF-67 nanocrystals were overcoated with an outer layer of silica to form ZIF-67@SiO 2 . Finally, Co cations in the ZIF-67 were converted into small-and large-sized Co nanoparticles with the protection of silica shells through two-step calcination, first in the air at 120 °C and 500 °C and then in H 2 at 400 °C. During pyrolysis, the ZIF-67 acted as a precursor and structure-directing template. The high-temperature calcination removes the ligand in ZIF-67 and residual organic species to eliminate the possibility of carbon contamination in further catalytic testing [12,18]. In fact, the direct pyrolysis of ZIF-67 without a SiO 2 sheath only leads to the formation of aggregated Co nanoparticles, manifesting the important role of SiO 2 in assisting the fabrication of a coresatellite structure with a discontinuous size distribution (Fig.  S2). Due to the two-step calcination and the presence of a thermally stable silica layer, no collapse of the framework was observed during the calcination treatment (Fig. S3). By contrast, the samples prepared via direct calcination at 500 °C cannot retain the hollow framework, and the largesized central CoO x nanoparticles disappeared (Fig. S4). A core-satellite structure was obtained after reduction, denoted as Co@SiO 2 , in which the loading of Co increased to 30.9% with a relatively large cobalt size of 100 nm and small cobalt satellite size of 11.7 nm. (Fig. 1c, d and Fig. S5). Figure 1e depicts the high-resolution TEM (HRTEM) image of Co@ SiO 2 . The sample displayed small-sized Co nanoparticles regularly embedded in the silica matrix. The lattice spacing along a specific direction was 0.21 nm, which agrees with the (111) crystal plane of the cubic-phase Co [33,34]. The elemental mapping results confirmed the coating of silica and demonstrated the existence of a core-satellite structure (Fig. 1f), which is further confirmed by the line-scan profile (Fig. 1g). These results indicate the successful preparation of core-satellite structured cobalt-based catalysts.
As a comparison, two other Co catalysts with loadings of 40.6% and 3.1% were prepared by the wetness impregnation method, denoted as Co-LNP and Co-SNP. The Co-LNP and Co-SNP catalysts were constituted of the silica support and Co nanocrystals with average particle sizes of 97 and 12 nm, similar to the sizes of the Co core and satellites in the IPR, respectively (Fig. S6). Then, XRD analysis was conducted to confirm the crystalline structure and phase morphology of Co. Fig. S7 depicts the XRD patterns of all oxidized Co samples, and the corresponding peaks match well with the Co 3 O 4 phase [34][35][36]. After reduction at 400 °C, the phases of Co@SiO 2 and Co-LNP were indexed to the mixture of the hexagonal and cubic phases (JCPDS 15-0806 and 01-1254) of Co. The formation of polycrystalline Co particles was also revealed by the selected area electron diffraction pattern of Co@SiO 2 (Fig. S8) [37,38]. In particular, the XRD pattern of the small-sized Co assigned to the CoO phase mainly resulted from the inevitable surface oxidation of Co nanoparticles during the sample preparation. The H 2 -TPR results demonstrate that Co 3 O 4 would completely be reduced to Co under the pretreatment before the subsequent characterization and catalytic reaction (Fig. S9).
The large-sized Co core is the main light absorber, and the small-sized Co satellites expose more catalytic active sites. To evaluate the intrinsic catalytic performance, the real content of small-and large-sized Co nanoparticles in Co@SiO 2 must be determined. H 2 -TPR and ICP-MS were combined to quantitatively distinguish the loading of Co satellites from that of the Co core. The TPR profiles can be deconvoluted into four peaks, as presented in Fig. S9. The first and third hydrogen consumption temperature ranges (A 1 and B 1 ) were attributed to the reduction of Co 3+ and Co 2+ of Co satellites on the silica sheath, respectively. Meanwhile, the second and fourth peaks (A 2 and B 2 ) were ascribed to the transitions of the centered cobalt core: Co 3 O 4 → CoO and CoO → Co [34]. The area of the consumption peak is proportional to the corresponding cobalt content. These results reveal that the relative contents for the centered Co and surrounding Co were 20.0% and 10.9%, respectively.
Highly dispersed active sites and strong absorption of sunlight are two essential factors for high-efficiency photothermal CO 2 catalysis. The IPR is made up of a Co core and surrounding Co satellites and may improve the lightabsorptive ability without sacrificing the highly dispersed active sites. To verify the hypothesis, the intrinsic catalytic performance of Co-based catalysts in catalyzing the CO 2 hydrogenation was investigated in a homemade flow reactor. The feed ratio of CO 2 :H 2 :N 2 was kept at 1:1:2 (volume ratio), and the reaction temperatures varied from 250 to 400 °C under dark conditions. As shown in Table 1, Co-LNP exhibited a very low conversion rate of 1.5 mmol/(g Co ·h) even at 400 °C, indicating that large-sized Co nanoparticles cannot exhibit a catalytic conversion due to the limited number of active sites. Furthermore, the CO 2 conversion rate of Co@SiO 2 normalized by the mass of Co satellites was quite similar to that of Co-SNP within the temperature range. These results demonstrate the important role of Co satellites in the H 2 and CO 2 activations over the core-satellite structure. The CO selectivity of Co-SNP was higher than that of Co@SiO 2 as a result of the narrower size distribution. During the consecutive thermocatalytic testing at 400 °C for 12.5 h, the CO 2 conversion rate and selectivity did not change for Co@SiO 2 , which exhibited superior thermal stability (Fig. S10).
Although small-sized Co nanoparticles have been demonstrated to accelerate CO 2 hydrogenation reactions, the light absorption property and photothermal effect also play an important role in the photothermal CO 2 catalysis. As shown above, Co cores with large sizes are inert in catalyzing CO 2 hydrogenation, but they can contribute to light absorption and photothermal conversion. To test this hypothesis, the diffuse reflectance spectra of different catalysts were measured. As depicted in Fig. 2a, the Co@SiO 2 exhibited a strong light absorption ability with an absorption efficiency of over 90% throughout the solar spectrum. A similar light absorption curve was observed for the Co-LNP, which reveals that large-sized Co nanoparticles exhibit superior light absorption properties. Conversely, the Co-SNP catalyst absorbed a very small number of photons, which could be explained by the small absorption cross sections and low loading amount of Co. Moreover, this trend of light absorption ability is consistent with the surface temperature of different samples monitored by an infrared (IR) camera under the 2.0 W/cm illumination. The stabilized surface temperature of Co@SiO 2 was 264 °C, 38 °C higher than that of Co-LNP ( Fig. 2b-g). The tiny light-absorptive difference between Co@SiO 2 and Co-LNP cannot account for the significant surface temperature difference. Therefore, a silica sheath might contribute to the local heating by the heat preservation and infrared shielding effect, which is consistent with our previous study [12]. These results provide direct evidence for the excellent photothermal conversion performance of Co@SiO 2 .
To verify the great potential of Co@SiO 2 in photothermal CO 2 catalysis, the photothermal catalytic performance of three samples was compared in a homemade flow reactor without external heating under different light intensities [39][40][41]. CO and CH 4 were detected as the major products for all samples. The activities of Co@SiO 2 and Co-SNP were improved without apparently changing the CO selectivity with the increase of the light intensity (Fig. 3a, b). Notably, the CO 2 conversion rate normalized by the Co satellite mass of 711 mmol/(g Co ·h) of Co@SiO 2 was achieved, which is one order of magnitude higher than that of CO-SNP under the same illumination of 2.5 W/cm 2 . This outcome can be explained by the poor light absorption properties of CO-SNP with low surface temperatures. Despite the excellent photothermal conversion performance and high surface temperature of CO-LNP, no products were detected within the limit of the gas chromatography due to the few active sites and low reactivity. By contrast, the Co@SiO 2 catalyst exhibited excellent photothermal catalytic CO 2 performance, which is mainly caused by the enhanced photothermal effects of large-sized plasmonic Co nanoparticles with a great intrinsic catalytic performance of small-sized Co nanoparticles. Previous studies have demonstrated the crucial role of particle size control in optimizing catalytic selectivity. The broad Co size range of Co@SiO 2 was responsible for the production of CH 4 , and the selectivity result was consistent with that of thermocatalytic testing [39,42]. These results fully demonstrate that the synergistic effects of small-and large-sized particles make great contributions to improving photothermal performance. In addition to the nanoscale greenhouse effect, the spatial confinement of a silica sheath in the core-satellite structure also provides an effective way of stabilizing small-sized Co nanoparticles under reaction conditions. To demonstrate the superior catalyst stability of Co@SiO 2 , the photothermal catalytic performance of samples was tested in a continuous run of 8.5 h under a light intensity of 1.7 W/cm 2 . Co@SiO 2 exhibited a very stable conversion rate of 80 mmol/(g Co ·h) and selectivity during the 8.5 h period ( Fig. 3c and Fig. S11). To understand the origin of excellent stability, the spent Co@SiO 2 catalyst after the 8.5 h testing was investigated.
The TEM image confirmed the presence of well-dispersed small-sized Co nanoparticles with the original sizes, and no collapse of the framework was observed (Figs. S12 − 13). Evidently, the Co@SiO 2 catalyst not only exhibited a high CO 2 conversion rate but also showed outstanding thermal stability.

Conclusion
In summary, we developed a core-satellite structured IPR catalyst (Co@SiO 2 ) which was integrated with strong light-absorptive ability, excellent intrinsic catalytic ability, and nanoscale thermal management. Through the combination of complementary advantages of small-and largesized Co nanoparticles, the unique structure exhibited an activity of 711 mmol/(g Co ·h) and good stability in the photothermal hydrogenation of CO 2 . Essentially, subsequent works will be focused on the extension of this strategy to reducible oxides (e.g., TiO 2 and ZrO 2 ) to modulate electronic structures and further facilitate good metallic dispersion to enhance the selectivity of CO. In addition, the simple synthetic method is universal to preparing other MOF-based photothermal catalysts with high loading and dispersity. Our research offers a new avenue for the preparation of highly dispersed photothermal catalysts without sacrificing the photothermal conversion ability. Furthermore, it lays a foundation for the utilization of nonprecious metals in photothermal applications.

Conflicts of interest
The authors declare that there is no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.