Role of base in CO₂ hydrogenation to formate over a Ru^(III) solid micellar catalyst

Abstract

The catalytic conversion of CO₂ to formate is generally performed under basic conditions. While the specific role of amines may not always be clearly defined, they have been the preferred additives to enhance activity and conversion. In this work, we demonstrate that during CO₂ hydrogenation to formate over a Ru^(III) solid micellar catalyst, the base acts as a promoter by improving the sluggish heterolytic H₂ dissociation. Kinetic experiments revealed that it is crucial to direct CO₂ speciation towards bicarbonate to ensure formate generation, and that there is a strong dependence of the hydrogenation rate on H₂ and TEA availability. H₂–D₂ isotope scrambling showed a fourfold increase in the presence of the base. Screening of different bases evidenced the unique promoting effect of TEA.

---

1. Introduction

The catalytic transformation of CO₂ to formate/formic acid is gaining attention due to the potential of formate salts/formic acid for reversible hydrogen storage.^1–3 The synthesis of formic acid from CO₂ and H₂ is endergonic in the gas phase, but the reaction becomes thermodynamically favorable in aqueous medium due to solvation effects.^4,5 Stoichiometric amounts of bases are typically added as co-reagents to promote the stabilization of formic acid as formate salts or adducts.⁶ N-containing organic bases, used as reaction additives, and amine-related functional groups, incorporated as ligands in molecular catalysts, promote CO₂ reduction by acting as proton shuttles, capturing CO₂, and providing electronic stabilization to reaction intermediates.⁷

Proton shuttles are important in CO₂ hydrogenation to formate, as they facilitate H₂ heterolytic cleavage into a hydride and a proton. The hydride reacts with CO₂ to produce formate, while the proton needs to be efficiently transferred to a base to close the catalytic cycle. Amines readily accept a proton, forming ammonium. Filonenko et al. attributed the high hydrogenation rate over a Ru PNP-pincer catalyst in a mixture of N,N-dimethylformamide (DMF) and 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) to the amine-assisted Ru–H bond formation via heterolytic dissociation of a coordinated H₂ molecule.^7,8 In their study, activity was further enhanced by increasing H₂ partial pressure. Rawat et al. demonstrated that the heterolytic H₂ cleavage and hydride transfer over a series of Fe complexes improved due to the presence of an outer-sphere pendant amine.⁹

Amines have been employed as CO₂ trapping agents in reduction reactions over homogeneous catalysts due to their CO₂-capturing ability.¹⁰ In the presence of water, primary and secondary amines (RNH₂ and R₂NH) react with CO₂ to form carbamates (RNHCOO⁻ and R₂NCOO⁻) which can hydrolyze to produce bicarbonate (HCO₃⁻) and protonated amines (RNH₃⁺ and R₂NH₂⁺). Tertiary amines (R₃N) directly form HCO₃⁻ and protonated tertiary amines (R₃NH⁺). In organic solvents, where reactants and products have only one possible form, alkaline aqueous solutions involve multiple equilibria, with reactive species and products varying based on amine type and concentration, pH, temperature and CO₂ partial pressure.^11,12

High CO₂ conversions have been reported over homogeneous catalysts in the presence of amines, with several examples in the literature displaying advancements in catalyst and process development.^13–15 Compared to homogeneous catalysts, heterogeneous catalysts own several advantages for large-scale application in continuous flow, including improved catalyst stability and lower production costs.¹⁶ Single-site catalysts are especially promising for CO₂ hydrogenation to formate due to their resemblance to homogeneous systems: well-defined active sites, high activity and selectivity.

Ru^(III)@MCM is the first example of a SOlid MICellar (SOMIC) catalyst, consisting of Ru^(III) single sites incorporated into the walls of MCM-41 via Ru–O–Si bonds, and stabilized by a cetyltrimethylammonium (CTA⁺). The presence of the CTA⁺ molecules prevents sintering and creates an apolar environment in the pores, favoring the concentration of nonpolar reactants and intermediates while excluding bulk water.^17,18 The Ru^(III) active sites maintain a local environment where water can still interact through coordination or hydrogen bonding, ensuring their accessibility to aqueous-phase reactants.

At 90 °C and 50 bar, a turnover frequency (TOF) of 182 h⁻¹ and a formate concentration of 1 mol L⁻¹ after 15 h was reached in aqueous medium and in the presence of TEA over Ru^(III)@MCM. Catalyst activity and selectivity was maintained over four consecutive CO₂ hydrogenation cycles.¹⁸ DFT modelling suggests that the reaction proceeds via heterolytic splitting of H₂, forming a Ru–H species, followed by hydride transfer to CO₂. H₂ activation is the rate-determining step of the reaction and the Ru–H complex is identified as the resting state of the catalyst. The role of base was not explicitly included in the reaction path, but it was hypothesized that the heterolytic H₂ dissociation could happen over (O)₃SiO⁻, OH⁻, HCO₂⁻, HCO₃⁻ or TEA.

In this work, we investigate the reaction kinetics of CO₂ hydrogenation to formate over Ru^(III)@MCM, focusing on the influence and mechanistic role of the base. Combining the study of the dependence of formate concentration on reaction conditions and H₂–D₂ isotope scrambling, we evaluate the role of pH and CO₂ speciation in the reaction medium, the influence of H₂ partial pressure and TEA concentration on the formate production rate. The performance of a selection of different solvents and bases is evaluated and compared.

2. Experimental

2.1 Chemicals

Hexadecyltrimethylammonium bromide (CTAB, C₁₉H₄₂BrN, ≥98%), ruthenium chloride (RuCl₃, Ru content 45–55%), ammonium hydroxide solution (NH₄OH, ACS reagent 28–30% NH₃ basis), tetraethyl orthosilicate (TEOS, SiC₈H₂₀O₄, reagent grade 98%), triethylamine (TEA, C₆H₁₅N, ≥99%), tripropylamine (TPA, C₉H₂₁N, ≥98%), triethanolamine (TEOA, C₆H₁₅NO₃, ≥98%), N,N-diisopropylethylamine (DIPEA, C₈H₁₉N, ≥99%), propylamine (PA, C₃H₉N, ≥98%), and morpholine (Mor, C₄H₉NO, ≥99%), dimethyl sulfoxide (DMSO, (CH₃)₂SO, ≥99%) and methanol (MeOH, CH₃OH, ≥99.8%) were purchased from Merck Chemicals. Ethanol (EtOH, C₂H₅OH, denaturated with Eurodenaturant) was supplied by Chem Lab. Water was deionized using a Millipore system. H₂, CO₂ and Ar were purchased from Air Liquide and D₂ was purchased from Eurisotop. All chemicals were used without further purification.

2.2 Catalyst synthesis

Ru^(III)@MCM was prepared according to the procedure described by Wang et al.¹⁸ 0.5 g of hexadecyltrimethylammonium bromide (CTAB) powder was added to 96 mL of deionized water and 34 mL of ethanol under stirring at 250 rpm for 30 minutes, until a fully transparent solution was obtained. Then, 0.09 g of ruthenium(III) chloride was added to the mixture and stirred at 650 rpm for 20 minutes to ensure RuCl₃ was fully dissolved. Next, 10 mL of aqueous ammonium hydroxide solution (NH₄OH) was added, followed by the dropwise addition of 2 mL of tetraethyl orthosilicate (TEOS). The resulting solution was stirred for three hours at room temperature. Approximately 1 g of solid product was collected through vacuum filtration, washed with distilled water until the filtrate attained a neutral pH, and finally dried overnight at room temperature.

Ru^(III)@Mic is the homogeneous counterpart of Ru^(III)@MCM. 0.5 g of CTAB powder was added to 96 mL of deionized water and 34 mL of ethanol. The mixture was stirred at 250 rpm for 30 minutes until a fully transparent solution was obtained. Then, 0.09 g of ruthenium(III) chloride was added to the mixture and stirred at 650 rpm for 20 minutes, ensuring complete dissolution of RuCl₃. No TEOS was introduced.

2.3 Catalytic tests

Hydrogenation reactions were performed in a 45 mL stainless steel pressure reactor manufactured by Parr Instrument Company, with a PTFE inner chamber. During reaction, the vessel was placed in an aluminum support placed on a ceramic hot plate stirrer (IKA C-MAG HS7). Temperature was controlled using an electronic contact thermometer (IKA ETS-D5) placed in the aluminum support (Fig. S1, SI).

In a standard procedure, the vessel was initially filled with water, base, fresh catalyst and a magnetic stirrer. Then, the reactor was sealed, purged with CO₂, pressurized with the gaseous reactants and gradually heated to the desired reaction temperature for 30 minutes. After reaction, the vessel was cooled to room temperature, the pressure released and the gas phase analyzed via gas chromatography (GC). The liquid product was separated from the catalyst using a syringe filter. Quantitative analysis of the formate concentration in the liquid product (C_HCOO⁻) was performed via¹H NMR using a Spinsolve 80 MHz phosphorous benchtop NMR spectrometer from Magritek and TSP-d₄ as internal standard (Fig. S2, SI). The standard deviation was calculated based on eqn (1), in which C_HCOO⁻,i is the individual formate concentration measured from each sample, is the average of the concentrations measured for each reaction and n is the total number of measurements.

The turnover frequency (TOF) and base conversion were calculated as defined in eqn (2) and (3), respectively.

where V is the volume of the reaction product, n_Ru is the amount of Ru (mol) in the catalyst used, t is the reaction time (h), n_HCOO⁻ is the amount of formate (mol) in the reaction product, and n⁰_base is the amount of base (mol) added to the reaction feed.

2.4 H₂–D₂ isotope scrambling

H₂–D₂ exchange reactions were performed using the same autoclave employed for the catalytic tests. Prior to each reaction, water, base and fresh catalyst were added to the autoclave. The reactor was sealed, pressurized with H₂ and D₂ with a ratio 1 : 1, and heated to the reaction temperature for 15 min. After reaction, the vessel was cooled to room temperature and connected to a mass spectrometer (MS) for off-line gas analysis. The gas released from the reactor was continuously diluted with argon (Ar) as a carrier gas in a flow line, ensuring a flow rate of 50 Nml min⁻¹. This diluted gas mixture was analyzed using Mass Spectrometry (MS) to monitor the composition until all the gas content was released from the autoclave. During the analysis, the relative intensities of H₂, D₂, and HD were quantified by calculating the signal ratios H₂/Ar, D₂/Ar, and HD/Ar. These ratios were derived from the raw MS signals recorded throughout the analysis period, providing a comparative measure of each gas component relative to the inert Ar reference.

3. Results and discussion

3.1 Kinetic insights

The effect of reaction time on formate concentration, catalyst activity, and TEA conversion was evaluated in a range from 3 to 20 h. The maximum formate concentration reached is 1.23 mol L⁻¹ after 20 h (Table 1, entry 6), aligning with the catalyst performance previously reported.¹⁸ Gas chromatography analysis confirmed the absence of CO and CH₄ after reaction. Catalyst productivity peaks at 2.6 g HCOO⁻ per g_cat. h⁻¹ (TOF = 214 h⁻¹) after 10 h (Table 1, entry 4). The pH remains constant throughout the reaction, and formate production increases linearly during the first ten hours. As the reaction progresses, activity slows down at formate concentrations higher than 1 mol L⁻¹ (Table 1, entries 4 to 6), near full TEA conversion. To guarantee the collection of kinetic data throughout our study, a reaction time of 4 h (X_TEA ∼ 24%, Table 1) was selected to study the dependence of formate concentration on the reaction conditions.

Table 1 Effect of reaction time on formate concentration, catalyst activity and TEA conversion. Conditions: 3.1 μmol Ru, 9 mmol TEA, 6 mL H₂O, 30 bar H₂, 20 bar CO₂, 90 °C

Entry	Reaction time (h)	[HCOO^−] (mol L^−1)	TOF (h^1)	X TEA (%)
1	3	0.08	56	5.9
2	4	0.31	158	24.2
3	5	0.42	174	30.5
4	10	1.03	214	75.9
5	15	1.09	162	85.4
6	20	1.23	137	95.8

---

The catalytic hydrogenation of CO₂ over Ru^(III)@MCM in the liquid phase requires the transfer of CO₂ and H₂ from the gas to the liquid phase, from the liquid phase to the external surface of the catalyst and finally to and within the pores of the catalyst. A study of the effect of stirring speed indicates no external mass transfer limitations above 750 rpm (Fig. S3, SI).

Varying the reaction temperature from 70 to 110 °C (Fig. 1) showed that formate production becomes measurable at 85 °C, and that an increase in the temperature improves CO₂ hydrogenation. The apparent activation energy (E_app) is 56 kJ mol⁻¹ (Fig. S4, SI).

Fig. 1 Effect of temperature on formate concentration. Conditions: 3.1 μmol Ru, 9 mmol TEA, 6 mL H₂O, 30 bar H₂, 20 bar CO₂, 4 h.

The dependence of formate concentration on the gaseous reactants partial pressure was investigated. A limited effect of CO₂ partial pressure is found (Fig. S5, SI). In contrast, there is a significant influence of H₂ partial pressure on the product concentration. An increase from to 10 to 30 bar leads to a tenfold increase of formate concentration (0.31 mol L⁻¹, Fig. 2) in the reaction product. A quasi-second order in H₂ is calculated (Fig. S6, SI), indicating that H₂ adsorption and activation are kinetically relevant.

Fig. 2 Dependence of the formate concentration on H₂ partial pressure. Conditions: 3.1 μmol Ru, 9 mmol TEA, 6 mL H₂O, 20 bar CO₂, 90 °C, 4 h.

The reaction was also studied at different TEA concentrations in the feed (X_TEA < 25%). In water, a very low concentration of formic acid is obtained (Fig. S7, SI). Formate concentration increases linearly from 0.05 to 0.35 mol L⁻¹ when the TEA concentration in the aqueous reaction feed is varied from 0.6 to 1.5 mol L⁻¹ (Fig. 3), all at low TEA conversion (X_TEA < 25%). When the concentration of TEA in the reaction feed exceeds 1.5 mol L⁻¹, catalyst performance drops (Fig. 3). Extending the reaction time from 4 h to 15 h shows a similar effect (Fig. S8, SI).

Fig. 3 Dependence of the formate concentration on the amount of TEA in the liquid reaction feed and pH upon CO₂ pressurization. Conditions: 3.1 μmol Ru, 6 mL H₂O, 30 bar H₂, 20 bar CO₂, 90 °C, 4 h.

The decrease in the Ru^(III)@MCM performance at higher TEA concentrations can be linked to a change in the pH of the liquid phase. Within the pH range of 8–9, bicarbonate (HCO₃⁻) is the main form of carbon in solution (Fig. S9, SI). In a more alkaline medium, carbonate (CO₃²⁻) becomes the dominant species (Fig. S9, SI). KHCO₃ was directly hydrogenated at pH 12 in the absence of CO₂ (entry 2, Table S1, SI, Fig. S10, SI), resulting in no formate production. However, when KHCO₃ hydrogenation was performed under CO₂ pressure at pH 8 (entry 4, Table S1, SI, Fig. S11, SI), the formate concentration increased significantly, surpassing that of the standard TEA-based system (0.31 mol L⁻¹, entry 1, Table S1, SI). In contrast, performing CO₂ hydrogenation at pH 8 in the presence of KHCO₃ but without TEA led to a notably lower formate concentration (entry 5, Table S1, SI, Fig. S12, SI), indicating that maintaining the pH at 8 alone is insufficient for high catalytic activity. Direct hydrogenation of K₂CO₃ at pH 13 (entries 6 to 8, Table S1, SI, Fig. S13 and S14, SI) does not yield formate. CO₂ hydrogenation in the presence of K₂CO₃ generates 0.26 mol L⁻¹ of formate (entry 9, Table S1, SI, Fig. S15, SI), highlighting that K₂CO₃ can serve as a base in the reaction.

These results suggest that at high TEA concentration or during the direct hydrogenation of KHCO₃ and K₂CO₃ the reaction does not proceed because the alkalinity of the medium favors the formation of CO₃²⁻ over HCO₃⁻. In the presence of CO₂, the pH decreases (Fig. 3, Table S2, SI), shifting the equilibrium towards HCO₃⁻. Keeping the pH of the reaction medium in the range 8–9 to maintain HCO₃⁻ as the predominant species in solution seems crucial for formate synthesis over Ru^(III)@MCM (Fig. 3).

As the choice of solvent can impact both the reaction thermodynamics and catalytic mechanism,¹⁹ CO₂ hydrogenation over Ru^(III)@MCM was further studied in DMSO and in methanol (Table S3, SI). Fig. 4 shows that in the absence of a base, the performance of Ru^(III)@MCM in DMSO is similar to the one observed in water. The group of Leitner reported that DMSO stabilizes HCOOH by hydrogen bonding.⁶ Our results demonstrate that replacing water and TEA by a solvent to promote formic acid stabilization did not lead to a higher formate concentration in our system (Fig. 4). Performing the reaction in MeOH and DMSO in the presence of TEA generated formate concentrations three times lower (approximately 0.1 mol L⁻¹) than in aqueous medium (Fig. 4). Similarly, gradually increasing the amount of MeOH and DMSO in mixtures with water did not lead to improved performance (Tables S4 and S5, SI). In aqueous environments, TEA reacts with CO₂ to produce HCO₃⁻. In non-aqueous solvents such as methanol or DMSO, CO₂ interaction with tertiary amines is limited.²⁰

Fig. 4 Performance of Ru^(III)@MCM for CO₂ hydrogenation to formic acid/formate in different solvents. Conditions: 3.1 μmol Ru, no TEA* or 9 mmol, 6 mL solvent, 30 bar H₂, 20 bar CO₂, 90 °C, 4 h.

3.2 Role of the base

The quasi-second order in H₂ (Fig. S6, SI) and the first order in TEA concentration (Fig. 3) suggest that the tertiary amine is involved in H₂ activation. To investigate the involvement of TEA in the heterolytic H₂ cleavage step during the CO₂ hydrogenation to formate, a set of three H₂–D₂ isotope scrambling experiments was carried out (Table 2). H₂–D₂ scrambling involves exposing hydrogen molecules to a catalyst under specific conditions, allowing for the exchange of hydrogen atoms with deuterium atoms. Breaking of H–H bonds and the incorporation of deuterium atoms leads to the appearance of HD as a reaction product. By monitoring the extent of HD in the reaction products, conclusions about the catalyst's capacity to dissociate H₂ can be drawn as well as the involvement of the base.

Table 2 Summary of reaction conditions for H₂–D₂ isotope scrambling experiments and ratios between the MS raw signals of H₂, D₂, HD and Ar after reaction. Reaction conditions: 4 bar H₂, 4 bar D₂, 1 mL H₂O, 50 °C, 1 h, analysis conditions: total flow of 50 mL min⁻¹ fed to the MS (90% Ar)

Entry	Ru (μmol)	TEA (mmol)	Ratio of the MS signals
			H2/Ar	D2/Ar	HD/Ar
1	0	0	0.59	0.19	0.01
2	3.1	0	0.60	0.17	0.08
3	3.1	1.4	0.28	0.10	0.30

---

Fig. S16, SI shows the mass spectrometer signals of H₂, D₂, HD and H₂O during the off-line analysis of the autoclave gas content, after dilution in Ar. In the absence of both the catalyst and base, only a negligible amount of HD is formed (Table 2, entry 1). Introducing Ru^(III)@MCM (Table 2, entry 2) results in an eightfold increase in the HD/Ar ratio, demonstrating the catalyst's critical role in activating H₂ and facilitating hydrogenation. When TEA is added to the system (Table 2, entry 3), the HD/Ar ratio increases another fourfold. The combined presence of Ru^(III)@MCM and TEA creates a synergistic effect, significantly boosting HD production and supporting the hypothesis that the base modifies the reaction mechanism.

To further confirm the involvement of TEA in H₂ dissociation, ¹H NMR spectroscopy was conducted using the homogeneous parent complex of the Ru SOMIC catalyst, Ru^(III)@Mic. Upon pressurization with H₂ and heating at 50 °C for 1.5 hours, distinct changes were observed in the chemical shifts of TEA (Fig. S17, SI (d)) compared to control samples lacking H₂ (Fig. S17, SI (c)) or Ru^(III)@Mic (Fig. S17, SI (e) and (f)). Notably, the CH₂ resonance of TEA shifted downfield from 2.53 ppm (Fig. S17, SI (c)) to 2.78 ppm under reaction conditions (Fig. S17, SI (d)), indicating a change in the electronic environment around the nitrogen atom consistent with protonation.²¹

The proposed pathway for CO₂ hydrogenation in the presence of TEA is illustrated in Fig. 5. The first step concerns H₂ adsorption on the Ru active center. The formation of intermediate 2 (Int 2) is followed by TEA-assisted H₂ heterolytic dissociation via transition state 1 (TS 1), leading to the generation of a Ru hydride species and triethylammonium (Int 3). In the next step, the hydride is transferred to CO₂ (TS 2), which exists in dynamic equilibrium with HCO₃⁻ under basic conditions. The catalytic cycle is closed by formate (Int 4) removal from the active site aided by triethylammonium.

Fig. 5 A proposed mechanism for formate synthesis over Ru^(III)@MCM in the presence of TEA.

The reaction involves the transport of key components—H₂, CO₂, and the amine—through the apolar medium formed by the CTA⁺ surfactant within the pores of the Ru^(III)@MCM structure. The apolar environment within these pores plays a critical role in mediating the interaction of the reactants with the catalytic active sites. Effective base-mediated catalysis depends on the amine's basicity and ability to partition into the nonpolar pore environment, and subsequently interact with the metal site. Different tertiary, secondary and primary amines with varied basicity, lipophilic and steric properties (Table S6, SI, Fig. 6(a)) were screened for CO₂ hydrogenation over Ru^(III)@MCM, and their performance compared to that of TEA (Fig. 6(b)). Lipophilicity was evaluated based on the partition coefficient between n-octanal and water (log P_{Octanal/Water}), chosen as a model system to mimic the hydrophobic environment within the pores of Ru^(III)@MCM and the aqueous phase.

Fig. 6 Screening of amines with different basicity and lipophilicity. (a) for CO₂ hydrogenation to formate over Ru^(III)@MCM (b) and Ru^(III)@Mic (c). Conditions: 3.1 μmol Ru, 9 mmol of amine, 4.5 mmol EtOH (b), 6 mL H₂O, 30 bar H₂, 20 bar CO₂, 90 °C, 4 h (b) or 15 h (c).

TEA combines high basicity (pK_a = 10.75), moderate lipophilicity (log P_{Octanal/Water} = 1.26), and low steric hindrance, and consequently exhibited the strongest promoting effect on formate production among all amines tested (Fig. 6(b)). TPA has a similar basicity to TEA and is more lipophilic (log P_{Octanal/Water} = 2.72), but its bulkier alkyl chains likely hinder diffusion or orientation within the confined pore environment, resulting in approximately 3-fold lower activity. Interestingly, while both DIPEA and TEOA are tertiary amines, they were nearly inactive, comparable to the control reaction without base. This highlights that tertiary structure alone is not sufficient for promotion. DIPEA is strongly basic (pK_a = 10.98) and fairly lipophilic (log P_{Octanal/Water} = 1.89), but its significant steric bulk—arising from branched isopropyl groups—likely hinders its diffusion through the catalyst pores, resulting in poor performance. In contrast, TEOA has lower basicity (pK_a = 7.74) and is highly hydrophilic (log P_{Octanal/Water} = −1.31), though it is small and sterically unhindered. These findings indicate that an optimal combination of strong basicity, intermediate lipophilicity, and minimal steric hindrance is essential to facilitate efficient base promotion within this catalytic system.

To evaluate to which extent poor pore accessibility as result of the bulkier nature of DIPEA might influence formate production, the same tertiary amines were screened over Ru^(III)@Mic (Fig. 6(c)). In these conditions, DIPEA outperformed TPA. The higher TOF of Ru^(III)@MCM compared to Ru^(III)@Mic can be a result of the increased electron density around the metal ion, promoted by the interaction with the supporting silica matrix.

Finally, screening secondary and primary amines, within the same range of pK_a,^8–11 led to productivities ten to twenty times lower than tertiary amines. Since secondary and primary amines typically react with CO₂ to form stable carbamates, we hypothesize these species are less reactive over Ru^(III)@MCM.

4. Conclusions

The study of the liquid phase kinetics of CO₂ hydrogenation to formate over Ru^(III)@MCM provides a deeper understanding of the reaction's dependence on H₂ partial pressure, concentration of TEA, pH, choice of solvent and role of the base.

The observed quasi second-order dependence in H₂ partial pressure reveals the importance of hydrogen activation for the reaction. High initial concentrations of TEA prevent the reaction because the high alkalinity of the medium favors formation of CO₃²⁻ over HCO₃⁻. Performing the reaction in water guarantees HCO₃⁻ formation over carbamates and maintaining a pH between 8–9 is crucial to keep HCO₃⁻ as the predominant species, ensuring formate synthesis over Ru^(III)@MCM.

H₂–D₂ isotope scrambling confirms that the catalyst is able to activate H₂, with the addition of a TEA significantly enhancing formate production. Among the amines screened, TEA has the highest promoting effect. The presence of a base in the vicinity of the Ru^(III) active centers is crucial for H₂ heterolytic dissociation during reaction. Small bases, with a moderate lipophilicity and a pK_a above 10 are preferred as promoters.

These findings provide insights into the role of bases in CO₂ hydrogenation over Ru^(III)@MCM and guide optimization strategies for selective, base-free formic acid synthesis.

Author contributions

Catarina M. De Brito Mendes: conceptualization, investigation, formal analysis, validation, visualization, writing – original draft, writing – review and editing. Javiera Rubio: investigation. Francisca Rebelo: investigation. Sara Santos: investigation. Jan Öner: writing – review and editing. Vitaly V. Ordomsky: writing – review and editing. Mark Saeys: conceptualization, funding acquisition, project administration, supervision, writing – review & editing.

Conflicts of interest

There are no conflicts of interest to declare.

Data availability

Supplementary information is available. See DOI: https://doi.org/10.1039/D5CY00662G.

The processed data supporting this article (¹H NMR and MS) has been included as part of the SI. Raw data are available on request from the corresponding author.

Acknowledgements

This research was supported by the Energy Transition Fund, an initiative from the Belgian FPS Economy, under the Molecules at Sea (MuSE) project. The author V. V. O. acknowledges the financial support of the French National Research Agency (DEZECO, Ref. ANR-22- CE05-0005).

References

1. J. Eppinger and K.-W. Huang, Formic Acid as a Hydrogen Energy Carrier, ACS Energy Lett., 2017, 2(1), 188–195.
2. K. Grubel, H. Jeong, C. W. Yoon and T. Autrey, Challenges and opportunities for using formate to store, transport, and use hydrogen, J. Energy Chem., 2020, 41, 216–224.
3. S. Chatterjee, I. Dutta, Y. Lum, Z. Lai and K.-W. Huang, Enabling storage and utilization of low-carbon electricity: power to formic acid, Energy Environ. Sci., 2021, 14(3), 1194–1246.
4. W. Leitner, Carbon dioxide as a raw material: the synthesis of formic acid and its derivatives from CO₂, Angew. Chem., Int. Ed. Engl., 1995, 34(20), 2207–2221.
5. R. Sun, Y. Liao, S.-T. Bai, M. Zheng, C. Zhou and T. Zhang, et al., Heterogeneous catalysts for CO₂ hydrogenation to formic acid/formate: from nanoscale to single atom, Energy Environ. Sci., 2021, 14(3), 1247–1285.
6. K. Rohmann, J. Kothe, M. W. Haenel, U. Englert, M. Hölscher and W. Leitner, Hydrogenation of CO₂ to Formic Acid with a Highly Active Ruthenium Acriphos Complex in DMSO and DMSO/Water, Angew. Chem., Int. Ed., 2016, 55(31), 8966–8969.
7. J. B. Jakobsen, M. H. Rønne, K. Daasbjerg and T. Skrydstrup, Are Amines the Holy Grail for Facilitating CO₂ Reduction?, Angew. Chem., Int. Ed., 2021, 60(17), 9174–9179.
8. G. A. Filonenko, R. van Putten, E. N. Schulpen, E. J. M. Hensen and E. A. Pidko, Highly Efficient Reversible Hydrogenation of Carbon Dioxide to Formates Using a Ruthenium PNP-Pincer Catalyst, ChemCatChem, 2014, 6(6), 1526–1530.
9. K. S. Rawat, A. Mahata and B. Pathak, Catalytic Hydrogenation of CO₂ by Fe Complexes Containing Pendant Amines: Role of Water and Base, J. Phys. Chem. C, 2016, 120(47), 26652–26662.
10. R. E. Siegel, S. Pattanayak and L. A. Berben, Reactive Capture of CO₂: Opportunities and Challenges, ACS Catal., 2023, 13(1), 766–784.
11. G. Kovács, G. Schubert, F. Joó and I. Pápai, Theoretical investigation of catalytic HCO3⁻ hydrogenation in aqueous solutions, Catal. Today, 2006, 115(1), 53–60.
12. R. Zhang, R. Liu, F. Barzagli, M. G. Sanku, L. Ce and M. Xiao, CO₂ absorption in blended amine solvent: Speciation, equilibrium solubility and excessive property, Chem. Eng. J., 2023, 466, 143279.
13. J. Kothandaraman, A. Goeppert, M. Czaun, G. A. Olah and G. K. Surya Prakash, CO₂ capture by amines in aqueous media and its subsequent conversion to formate with reusable ruthenium and iron catalysts, Green Chem., 2016, 18(21), 5831–5838.
14. T. Schaub and R. A. Paciello, A process for the synthesis of formic acid by CO₂ hydrogenation: thermodynamic aspects and the role of CO, Angew. Chem., Int. Ed., 2011, 32(50), 7278–7282.
15. K. R. Ehmann, A. Nisters, A. J. Vorholt and W. Leitner, Carbon Dioxide Hydrogenation to Formic Acid with Self-Separating Product and Recyclable Catalyst Phase, ChemCatChem, 2022, 14(19), e202200892.
16. C. Kim, K. Park, H. Lee, J. Im, D. Usosky and K. Tak, et al., Accelerating the net-zero economy with CO2-hydrogenated formic acid production: Process development and pilot plant demonstration, Joule, 2024, 8(3), 693–713.
17. Q. Wang, C. M. De Brito Mendes, O. V. Safonova, W. Baaziz, C. A. Urbina-Blanco and D. Wu, et al., Tunable catalysis by reversible switching between Ru^(III) single sites and Ru⁰ clusters in solid micelles, J. Catal., 2023, 426, 336–344.
18. Q. Wang, S. Santos, C. A. Urbina-Blanco, W. Y. Hernández, M. Impéror-Clerc and E. I. Vovk, et al., Solid micellar Ru single-atom catalysts for the water-free hydrogenation of CO₂ to formic acid, Appl. Catal., A, 2021, 290, 120036.
19. S. A. Burgess, A. M. Appel, J. C. Linehan and E. S. Wiedner, Changing the Mechanism for CO₂ Hydrogenation Using Solvent-Dependent Thermodynamics, Angew. Chem., Int. Ed., 2017, 56(47), 15002–15005.
20. L. B. Hamdy, C. Goel, J. A. Rudd, A. R. Barron and E. Andreoli, The application of amine-based materials for carbon capture and utilisation: an overarching view, Mater. Adv., 2021, 2(18), 5843–5880.
21. D. B. G. Berry, I. Clegg, A. Codina, C. L. Lyall, J. P. Lowe and U. Hintermair, Convenient and accurate insight into solution-phase equilibria from FlowNMR titrations, React. Chem. Eng., 2022, 7(9), 2009–2024.

---