Molecular Rhodium Complex within N‐Rich Porous Polymer Macroligand as Heterogeneous Catalyst for the Visible‐Light Driven CO2 Photoreduction

The heterogenization of molecular catalysts within a porous solid acting as macroligand can advantageously open access to enhanced stability and productivity, and thus to more sustainable catalytic process. Herein, a porous organic polymer (POP) made through metal‐free polymerization using bipyridine repeating units is reported. This N‐rich POP is an efficient macroligand for the heterogenization of molecular rhodium complexes. The intrinsic catalytic activity of the heterogenized catalyst is slightly higher than that of its homogeneous molecular counterpart for formic acid production as a unique carbon‐containing product. The heterogenization of the rhodium catalysts enables recycling for a total productivity of up to 8.3 g of formic acid per gram of catalyst after 7 reuses using visible light as the sole energy source.


Introduction
[3][4][5][6][7] Beyond the advantageous easy separation of solubilized product from the solid catalyst and the possible recycling of the latter, the heterogenization of a single-site molecular complex within the porous host allows to control the catalytic activity by the design of the host.As in the case of molecular ligands, the choice of the host material allows to control precisely the electronic environment of the active site at the molecular level. [8]Thus the solid host can be seen as a porous macroligand. [8,9]Such porous macroligands can be built by derivatization of widely used molecular ligands like bipyridine (bpy) [10] as building units for the construction of porous materials like metal-organic frameworks (MOFs), covalent organic frameworks (COFs), and porous organic polymers (POPs). [9,11,12]The concept of porous macroligands benefits from a high synthetic versatility through the choice of co-monomers and the nature of the reticulation node, i.e., organic node in POPs or inorganic node in MOFs, allowing for adjusting the textural properties like porosity as well as the chemical environment of the heterogenized active sites. [9]n the context of renewable energy utilization, the direct reduction of carbon dioxide into valuable C1 molecules, like methanol or formic acid, using sunlight as the sole energy source is of great interest for combining greenhouse gas abatement and solar fuel production.In particular formic acid is considered as a promising renewable feedstock for fine chemical syntheses. [13]16] Among well-defined molecular catalysts that allow for the photoinduced selective transformation of CO 2 into formic acid (HCOOH), the Cp*Rh(bpy) (Cp* = pentamethylcyclopentadienyl) complex stands out by its high catalytic activity and selectivity. [12,17,18]Recently, we have demonstrated that in series of MOF-, COF-, and POP-based macroligands, the highest activity was obtained for the macroligand that causes the less pronounced electron-withdrawing effects to the Cp*Rh(bpy) site.Thus, for this particular catalyst, POPs are the most suitable macroligands to achieve a high catalytic activity in the CO 2 reduction reaction. [8,19]espite their appealing activity, the reported bpy-based POPs, namely BpyMP-1 and BpyMP-2, are still less active than the molecular, nonfunctionalized 2,2 0 -bipyridine-based catalyst Cp*Rh(bpy) Cl 2 (Figure 1). [8]The lower activity was attributed to the electronwithdrawing effect of the ethynyl groups present at the metaposition of the bpy in the BpyMP materials.In addition, these materials were prepared through a Pd-catalyzed Sonogashira crosscoupling, reducing the overall sustainability of the catalysts and DOI: 10.1002/aesr.202300209 The heterogenization of molecular catalysts within a porous solid acting as macroligand can advantageously open access to enhanced stability and productivity, and thus to more sustainable catalytic process.Herein, a porous organic polymer (POP) made through metal-free polymerization using bipyridine repeating units is reported.This N-rich POP is an efficient macroligand for the heterogenization of molecular rhodium complexes.The intrinsic catalytic activity of the heterogenized catalyst is slightly higher than that of its homogeneous molecular counterpart for formic acid production as a unique carbon-containing product.The heterogenization of the rhodium catalysts enables recycling for a total productivity of up to 8.3 g of formic acid per gram of catalyst after 7 reuses using visible light as the sole energy source.
implying a need for palladium traces removal [20][21][22] or for at least the assessment of the inactivity of remaining palladium traces. [8]25] In this work, we take advantage of the high versatility of POP materials to design fully organic N-rich porous macroligands for the single-site heterogenization of active organometallic catalysts.A bipyridine-based POP, N-POP, made by radical polymerization, is used as a porous macroligand for an organometallic Cp*Rh catalyst for the CO 2 reduction into formic acid driven by visible light (Figure 1).The heterogenization within the N-POP allows reaching the intrinsic catalytic activity of its molecular counterpart without any electronic or diffusional restriction, outperforming analogous POP-based catalysts.

Synthesis of Bipyridine-Based POP Catalysts
The bipyridine-based porous macroligand was synthesized via radical copolymerization of 1,4-divinyl-benzene with 4,4 0 -divinyl-2,2 0 -bipyridine to yield bipyridine-containing POP (named hereafter N-POP, Scheme 1, see SI for experimental details).To introduce also mesopores into the framework, the polymerization was carried out in a mixture of water and tetrahydrofuran (THF) in a volumetric ratio of 1-10. [26]he elemental analysis reveals that N-POP contains 3 wt% of nitrogen, corresponding to 1.1 mmol of bpy per gram of N-POP (Table S2, Supporting Information).Fourier transform infrared (FT-IR) spectra of N-POP show strong signals at 2930 and 2859 cm À1 associated with the stretching vibration of bridging -CH 2 , and -CH groups, demonstrating the successful formation of polymeric network.The Cp*Rh catalytic synthon was embedded into a polymer host following an established procedure, [8] by infiltration of a methanolic solution of [Cp*RhCl]NO 3 for 24 h at room temperature, to obtain the Cp*Rh@N-POP heterogeneous catalysts.Various Rh loadings (ω Rh ), between 0.3 and 1.5 wt% (named hereafter x wt% Cp*Rh@N-POP), were prepared.
The inductively coupled plasma optical emission spectroscopy (ICP-OES) analysis of the Cp*Rh@N-POP series revealed Rh loadings of 0.3, 0.6, and 1.3 wt%, corresponding, respectively, to 0.025, 0.059, and 0.125 mmol of Rh per gram of N-POP material (Table S2, Supporting Information).Accordingly, the Rh:bpy ratio within the Cp*Rh@N-POP catalysts can be calculated to be 1:45, 1:20 and 1:10, respectively, a low bpy site occupancy being in line with the site-isolation of the Rh active sites within the material.
The 13 C solid-state magic angle spinning (MAS) nuclear magnetic resonance (NMR) spectra of the pristine polymers show signals ranging from 127 to 155 ppm which are assigned to the aromatic carbons, and the two overlapping signals at approx.40 ppm are assigned to the aliphatic carbon atoms from bridging methylene and methine groups (Figure 2).Signals at 28 and 13 ppm are attributed to ethylene groups of ethyl-vinyl-benzene, the major impurity in commercially available divinyl-benzene.The small peak at 112 ppm in the 13 C MAS NMR spectra is indicative for residual dangling vinyl groups.The additional signal at approx.155 ppm is assigned to carbon atoms next to the bipyridine N atom (Figure 2). [8]he 13 C MAS NMR spectra of the catalyst-loaded material 1.5 wt% Cp*Rh@N-POP, when compared to those of corresponding pristine polymers and molecular complexes, confirm the grafting of Cp*Rh moiety into the polymer's matrix.Two new peaks at 97 and 8 ppm are observed in the 13 C MAS NMR spectra of the polymer-based catalysts which can be assigned to the quaternary carbon atoms at 97 ppm and to the methyl groups at 8 ppm of the Cp* moiety (Figure 2 S6, Supporting Information).
Finally, all materials show a permanent porosity in nitrogen physisorption experiments at 77 K (Figure 3a and S1, Supporting Information).With increasing Rh loading, the apparent surface area decreases from approximately 790 AE 10 to 690 AE 10 m 2 g À1 for pristine N-POP and 1.5 wt% Cp*Rh@N-POP, respectively.Similar trends have been observed for purely microporous materials. [8]However, the decrease of accessible pore volume upon metalation is less prominent in the micro-mesoporous materials reported herein.In the series of N-POPs, the pore volume is reduced by approx.6%, while the apparent surface area is reduced by approx.13% (Table S1, Supporting Information).
Since it is fundamental for the solvent (here acetonitrile) used under catalytic conditions to have good accessibility to all active sites within the solid catalysts, acetonitrile vapor isotherms were recorded at 298 K (Figure 3b and S4, Supporting Information). [8,19,27]The isotherms reveal a hysteresis between adsorption and desorption branch which is most likely caused by a swelling of the polymer when being in contact with acetonitrile.Despite this swelling behavior, typical for porous polymers, the total volume of acetonitrile vapor adsorbed remains similar for pristine and metal-loaded polymers with %0.9 AE 0.1 cm 3 g À1 (Table S1, Supporting Information), revealing the accessibility of the porous network even after metalation.
Furthermore, renewable 2-methyl-tetrahydrofuran (2-Me-THF) can advantageously be used as a synthesis solvent without affecting the properties of the host material (see Figure S3 and S7, Supporting Information), making those materials high-potential platforms for heterogenized molecular catalysis from a perspective of more sustainable catalytic systems. [28]

Catalytic CO 2 Photoreduction
The photocatalytic activity of the series of Cp*Rh@N-POPs in the CO 2 reduction was evaluated under visible light illumination (irradiance %1000 W m À2 ) in a CO 2 -saturated acetonitrile (ACN)triethanolamine (TEOA) solvent mixture (5:1, V:V) in the presence of 1 mM of [Ru(bpy) 3 ]Cl 2 .[Ru(bpy) 3 ]Cl 2 acts as the photosensitizer and TEOA as proton and electron donor (Figure S15, Supporting Information).In the first step, we optimized the heterogeneous catalyst's concentration in milligrams per milliliter (mg mL À1 ) of the reaction mixture.A high mass of solid catalyst in the photoreactor might cause light scattering and consequently might reduce the light-harvesting efficiency of the photosensitizer in solution. [29]hus, this optimization is required to find a balance between light scattering and the mass of catalyst required for high productivity.
Independently of the Rh loading, the highest turnover frequencies (TOF, defined as the moles of formic acid produced per moles of rhodium per hour) were obtained when employing around 0.17 mg of solid catalyst per milliliter of solvent mixture (Figure 4a,b and S9, Table S3, Supporting Information, entries 1-9), which was kept constant in the following reactions.The highest TOF with 286.4 AE 7.2 h À1 was observed for ω Rh = 0.3 wt%.The decrease in TOF with increasing Rh loading might point to mass diffusion limitation when high amounts of Rh are incorporated in the pore network.Similar effects have been observed for other POP-based catalysts. [8]However, as the amount of photosensitizer was kept constant, the ratio of photosensitizer to catalyst changed in the series.At present, we cannot rule out that this change in ratio might affect the observed activity, due to the decomposition of the photosensitizer.The highest production rate R, expressed in mmol HCOOH /g cat /h, was obtained with ω Rh = 1.5 wt% to achieve a production rate of 17.9 AE 3.1 mmol HCOOH /g cat /h (Figure 4b).
Control experiments were carried out to confirm that the formic acid is indeed derived from CO 2 via a photocatalytic process.When using the pristine N-POP in the presence of Ru(bpy) 3 Cl 2 as a photosensitizer, only traces of formic acid were detected (Table S3, Supporting Information, entry 22).Such a low formic acid production has been reported previously for Ru(bpy) 3 Cl 2 containing photosystems and has been attributed to the decomposition of the photosensitizer by the photolabilization of one bipyridine ligand yielding Ru(bpy) 2 (solvent) 2 Cl 2 acting as a weak catalyst. [30,31]n the absence of CO 2 , photosensitizer or of light, no reaction occurred (Table S3, Supporting Information, entries 23-25).
Furthermore, isotopic 13 CO 2 labeling experiments confirm the selective formation of H 13 CO 2 H steaming from 13 CO 2 reduction S12, Supporting Information, for further discussion see ESI).Furthermore, IR spectra of the gas phase after photocatalysis showed no traces of CO or CH 4 (Figure S13, Supporting Information), in line with 13 C NMR spectrum of the liquid phase after catalysis, showing no traces of dissolved 13 CO at 185 ppm, 13 CH 3 OH at % 50 ppm, or 13 CH 4 at À4.6 ppm. [32,33]

Long-Term Productivity and Recycling
To evaluate the productivity of the catalysts during several hours of continuous irradiation, long-term experiments were carried out using 0.3 wt% Cp*Rh@N-POP.The formic acid production increased for the first 2 hours with almost constant TOF values, but decreased to 169.1 h À1 after 4 h of continuous catalysis (Figure 5a).Taking a closer look at the incremental TOF for each one-hour segment, the TOF in the first hour is 240.7 AE 10.3, 261.9 AE 13.7 h À1 in the second, 195.3 AE 14.6 h À1 in the third, and 169.1 AE 3.7 h À1 in the fourth hour.Note that the difference between 1 st and 2 nd hour of catalysis is within the error of multiple experiments.[36] This continuous formic acid productivity clearly highlights the Cp*Rh@N-POP catalyst's stability under catalytic conditions for at least 4 h of continuous irradiation.Likewise, for the homogeneous catalyst, the activation could be re-established to a certain extent, but with a slight decrease after 2 h (Figure S11, Supporting Information).Thus, these results might indicate increased stability of the catalyst due to its site isolation after heterogenization within a porous macroligand.in milligrams per milliliter (mg mL À1 ) using three different Cp*Rh loaded N-POP materials (0.3, 0.6, and 1.5 wt%; red, blue, and magenta, respectively) in the presence of 1 mM Ru(Bpy) 3 Cl 2 in ACN/TEOA (5:1, V:V).b) Product evolution rate R (black) and TOF (blue) as a function of rhodium mass fraction in N-POP using a constant mass of solid catalyst (%0.17 mg mL À1 ).
Interestingly, the incremental TOF increases from 240 h À1 (1 st hour) to 365.9 AE 5.6 h À1 (4 th hour).This increase in activity might be caused by a better accessibility of the catalyst caused by a slow but continuous swelling of the network over the total of 4 h of catalysis.Although ruthenium was not detected in the solid after catalysis, the bpy coordination of decomposed, but catalytically active, Ru species within the N-POP cannot be ruled out and might contribute-albeit on a lower level-to this increase in catalytic activity.Similar behavior has been observed, i.e., for Re(bpy)-based POP photocatalysts. [37]o evaluate the reusability of the heterogeneous catalysts, we conducted a series of use, separation, activation, and reuse experiments consisting of seven successive cycles of 1 h with 0.3 wt% Cp*Rh@N-POP (Figure 5b).Over those cycles, the catalyst exhibited good recyclability, with negligible variations of the activity within the standard deviation of multiple experiments (Figure 5b, Table S4, Supporting Information, entries 1-8), achieving a total cumulative turnover number (TON) 1676 after 7 runs of catalysis.Overall, the long-term and recycling experiments highlight the intrinsic stability of the N-POP-based catalyst.
A leaching test was conducted to demonstrate the heterogeneous nature of the catalysts.After 2 h of photocatalysis, the catalyst was isolated, and the supernatant was saturated with CO 2 and irradiated for another 2 h (Figure S10, Supporting Information).No significant change in the amount of HCOOH was observed, thus excluding the leaching of active species from the material into the solution.
X-ray fluorescence spectroscopy, as well as ICP-OES analysis of the supernatants, confirms that no quantifiable leaching of Rh occurs even after reuse for 7 consecutive catalytic runs.The robustness shown by Cp*Rh@N-POP during the catalytic experiments was further confirmed by 13 C MAS NMR (Figure 6), and FT-IR characterizations of the spent catalyst (Figure S14, Supporting Information).No significant differences between the spectra of the pristine and spent catalyst were observed, indicating the stability of Cp*Rh@N-POP material during the photocatalytic reaction.
Finally, the catalytic efficiency of the 0.3 wt% Cp*Rh@N-POP in the CO 2 -to-formic acid photoreduction has been confronted to that obtained using molecular Cp*Rh(bpy)Cl 2 (1) and using previously reported heterogenized Cp*Rh within BpyMP-1 [8] under the same conditions (Figure 7).
The homogeneous bipyridine complex 1 showed the same TOF with 279.6 AE 6.8 h À1 as 0.3 wt% Cp*Rh@N-POP (TOF of 286.4 AE 7.2 h À1 , Figure 7, Table S3, Supporting Information).Thus, in contrast to most of the literature known porous macroligands, the N-POP system does not affect the intrinsic activity of the heterogenized catalyst.
To allow for a direct comparison of catalytic activity with the best literature known heterogeneous Rh single-site catalyst for the visible-light-driven CO 2 -to-HCOOH transformation, namely, the Cp*Rh@BpyMP-1, [8] we also used this catalyst under the updated experimental catalysis protocol (0.05 mg mL À1 ) using an ABA class solar simulator at 1 sun (irradiance % 1000 W m À2 ).
Under those conditions, Cp*Rh@BpyMP-1 gives TOF and R values of 112.1 AE 11.2 h À1 and 8.4 AE 0.8 mmol HCOOH /g cat /h, respectively, still more than two times lower than the catalysts 1 and Cp*Rh@N-POPs (Figure 7, Table S4, Supporting Information).In contrast to N-POP, the previously reported BpyMP-1 framework is synthesized from 1,3,5-triethynylbenzene and 5,5 0 -dibromo-2,2 0bipyridine through a Pd-catalyzed cross-coupling (Figure 1). [8]he triethynylbenzene node in BpyMP-1 causes, however, a decrease in electron density on the active site, [8] while the methylene/methine substituents on the bpy in N-POP are supposed to slightly increase the electron density on the active site. [38]esides these beneficial electronic properties, the N-POP macroligand also showed a higher porosity to acetonitrile solvent compared to BpyMP-1 (Table S1, Supporting Information), together allowing for a molecular catalyst's heterogenization without reducing its intrinsic catalytic activity.
For a broader comparison with the literature, we focused on productivity which is defined as the number of moles of formic acid produced per mass of solid material and per time.In the series of Cp*Rh@N-POPs, the highest productivity R of 17.9 AE 3.1 mmol HCOOH /g cat /h was reached using 1.5 wt% Cp*Rh@N-POP.Previously reported single-site Mn, Rh, and Ru complexes within crystalline MOF and amorphous POP have reached productivity of up to 6 mmol HCOOH /g cat /h [8,12,34,39] and dinuclear Rh 2 paddlewheel sites within metal-organic polyhedra (MOP) showed productivity of 25-75 mmol HCOOH /g cat /h under similar photocatalysis conditions. [12]The overall lower productivity of heterogenized mononuclear complexes as compared to the di-rhodium MOP-based catalyst can be explained by a much higher density of active sites within the latter system.Note that for all those reports different mass loadings, catalyst loadings, and particle sizes have been used, parameters that all affect the photocatalytic activity (Figure 4a). [36,40,41]However, even when using higher mass loading of 1.5 wt% Cp*Rh@N-POP, this catalyst still outperforms the aforementioned single-site Rh and Ru catalyst with R of 10.2 and 8.5 mmol HCOOH /g cat /h for loadings of 0.35 and 0.5 mg mL À1 (conditions often used in literature), respectively.

Conclusion
In conclusion, we demonstrated that we can address by the design of solid porous macroligand the control over the activity of heterogenized molecular catalysts for the photoinduced CO 2 reduction reaction.A metal-free polymerization protocol allows for increasing the sustainability of the catalyst preparation.The reported Rh-loaded bpy-rich POP catalyst allows reaching the same intrinsic productivity as its parent molecular complex but with advantageous recyclability, to reach a productivity of 1.2 g of formic acid per gram of catalyst for each one-hour run and an overall 8.3 g of formic acid per gram of catalyst after 7 runs, among the highest for Rh-catalyzed CO 2 -to-HCOOH photoreduction.Together with a more virtuous synthesis, the high activity and stability of the reported POP-based catalysts, including the absence of leaching and efficient reuse, allow for increasing the sustainability of the whole catalytic process for greenhouse gas abatement and solar fuel production.Thus the synthetic versatility of porous organic polymers, envisioned as porous macroligands, opens bright perspectives for the development of a novel heterogeneous catalyst for more sustainable molecular catalytic processes.

Experimental Section
Catalyst Synthesis Cp*Rh@N-POP: 242 mg (1.5 mmol, 1 equiv.) of divinyl benzene (80%, passed through a column of basic alumina), 64 mg (0.30 mmol, 0.2 equiv.) of 4,4 0 -divinyl-2,2 0 -bipyridine, and 13.6 mg (55.5 μmol) of 1,1 0 -azobis(cyclohexanecarbonitrile) (ACHN) were dissolved in 4.4 mL of a deaerated 2-Me-THF:H 2 O mixture (10:1, V:V) under Ar.The solution was stirred for 1 h at room temperature.Then, the solution was transferred into a Teflon-lined stainless-steel autoclave and the autoclave was flushed with Ar.The autoclave was sealed and heated to 100 °C for 48 h in an oven (heating rate 10 K min À1 ).After the autoclave was cooled to room temperature, the polymer was transferred into a Soxhlet thimble and purified by Soxhlet extraction using THF for 24 h.The white solid N-POP was dried in an oven for 3 h at 65 °C followed by drying under vacuum at 85 °C and obtained a typical yield of 248 mg (97%).For a typical rhodium infiltration, 50 mg of macroligands (N-POP) were dispersed in 3 mL degassed methanol, then a defined amount of (Cp*RhCl)NO 3 stock solution was added and the suspension was stirred for 24 h at room temperature.The supernatant was removed by centrifugation and the solid was washed with MeOH until the supernatant remained colorless.The solid was dried under reduced pressure first at room temperature, then at 80 °C under vacuum.
Photocatalytic CO 2 Reduction: For a typical catalysis, about 0.5 mg of the catalyst was weighed into a UV-vis quartz cuvette (path length: 10 mm).Then 3 mL of 1 mM Ru(bpy) 3 Cl 2 in a mixture of acetonitrile and triethanolamine (5:1, V:V) were added, and the solution was saturated with CO 2 for 15 min.The cuvette was sealed and the suspension was irradiated under stirring using an ABA class solar simulator (MiniSol model LSH-7320, Newport) at 1 sun (irradiance % 1000 W m À2 ).After 2 h, the solid was removed by centrifugation, and the liquid phase was analyzed by q-1 H-NMR.

Figure 1 .
Figure 1.Nature and catalytic activity of the homogeneous molecular catalyst and its heterogenized derivatives in the photocatalytic conversion of CO 2 into formic acid.

Figure 4 .
Figure 4. a) Optimization of heterogeneous catalyst's concentration (β cat )in milligrams per milliliter (mg mL À1 ) using three different Cp*Rh loaded N-POP materials (0.3, 0.6, and 1.5 wt%; red, blue, and magenta, respectively) in the presence of 1 mM Ru(Bpy) 3 Cl 2 in ACN/TEOA (5:1, V:V).b) Product evolution rate R (black) and TOF (blue) as a function of rhodium mass fraction in N-POP using a constant mass of solid catalyst (%0.17 mg mL À1 ).