Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Nature Energy volume  7, pages 438–447 (2022 )Cite this article indurstrial grade acetonitrile

Efficient hydrogen storage and release are essential for effective use of hydrogen as an energy carrier. In principle, formic acid could be used as a convenient hydrogen storage medium via reversible CO2 hydrogenation. However, noble metal-based catalysts are currently needed to facilitate the (de)hydrogenation, and the CO2 produced during hydrogen release is generally released, resulting in undesirable emissions. Here we report an α-amino acid-promoted system for reversible CO2 hydrogenation to formic acid using a Mn-pincer complex as a homogeneous catalyst. We observe good stability and reusability of the catalyst and lysine as the amino acid at high productivities (CO2 hydrogenation: total turnover number of 2,000,000; formic acid dehydrogenation: total turnover number of 600,000). Employing potassium lysinate, we achieve >80% H2 evolution efficiency and >99.9% CO2 retention in ten charge–discharge cycles, avoiding CO2 re-loading steps between each cycle. This process was scaled up by a factor of 18 without obvious drop of the productivity.

As the most significant greenhouse gas, carbon dioxide (CO2) has risen from pre-industrial levels of 280 ppm (parts per million) to 419 ppm in the Earth’s atmosphere in February 2022, along with the exponential global energy demand supplied by carbon-rich fossil fuels (Supplementary Fig. 1)1,2. The extensive CO2 valorization and zero-CO2 emission technologies are crucial to mitigate global warming and related climatic deterioration. To accomplish such purpose, 196 countries/parties have signed the 2015 Paris Agreement committing to reach net-zero CO2 emissions around the year 2050. Generally, there are two approaches to realize carbon neutrality: reducing carbon emissions by shifting towards alternative energy technologies and balancing carbon emissions with carbon usage.

In this context, the feasibility study of hydrogen (H2) as a clean alternative energy carrier has inspired growing attention because it could be prepared from renewable resources, for example, by electrochemical water splitting, and produces nothing but water and energy in fuel cells3,4. However, it is troublesome to transport and store hydrogen gas due to its physical and explosive properties in mixtures containing oxygen. This situation can be solved by converting hydrogen gas to solid or liquid organic hydrogen carriers5,6, for example, by catalytic CO2 hydrogenation. Following this concept, besides methanol and Fischer–Tropsch products (hydrocarbons), formic acid (HCO2H, FA) and its formate salts also are readily accessible. Both are stable compounds that can be stored and dehydrogenated on demand to H2 and CO2 under milder conditions compared with other liquid organic hydrogen carriers (Fig. 1a)7,8,9, separating hydrogen storage and release without the restriction of time and place. It should be noted that hydrogen is wasted partly in the form of water when it is stored in methanol, which is not the case for formic acid, even though the hydrogen content in FA (4.4 wt%) is lower than that in methanol (12.6 wt%). Overall, a FA-based H2 storage and release system may also benefit from the CO content in the generated hydrogen (usually less than 10 ppm), which is important for its application in fuel cells.

a, Catalytic CO2 hydrogenation to formic acid and its dehydrogenation to H2 and CO2 in the presence of bases. Cat., Catalyst. b, Concept of the carbon-neutral chemical hydrogen storage and release described in this work. The chemical hydrogen storage and release is based on the interconversion of CO2, amino acid (AA), H2 and FA. Retaining CO2 in the cycle (bold arrows) is advantageous compared with CO2 recycling (dotted arrows). Green arrows represent recharge, while pink ones show discharge.

An electric battery is commonly a source of electric power containing electrochemical cells to power electronic products. Therefore, reusable electric batteries can be discharged and recharged multiple times under electric current. According to this concept, a chemical hydrogen battery is a device where energy is stored in the form of hydrogen that is discharged and recharged as needed. Hydrogen is directly converted to electric energy by using fuel cell technologies after its release out of the chemical hydrogen battery. Obviously, such technology offers substantial potential as a clean energy technology working towards carbon neutrality.

Only limited examples of such chemical hydrogen batteries have been demonstrated to date. Clearly, catalysts active in both hydrogenation and dehydrogenation reactions are essential and only a few molecularly defined transition metal-based complexes are known to fulfil this mission. Most systems contain expensive noble metal-based catalysts, for example, Rh (ref. 10), Ru (refs. 11,12,13,14,15,16,17,18) and Ir (refs. 19,20) (Supplementary Fig. 2). An example of a heterogeneous catalyst was reported in 2016 when a reusable bimetallic catalyst was developed to hydrogenate and dehydrogenate N-heterocycles efficiently21. Besides, it’s difficult to reach a simple and truly rechargeable hydrogen storage and release device when it comes to non-unified reaction conditions in H2 recharge and discharge steps, for example, catalysts, solvents11,22, bases change20, pH control19, reloading of storage media (concerning catalysts, H2 carriers and so on) between each charge–discharge cycle13 or generally low catalytic efficiency (that is, turnover numbers). Furthermore, in all currently known examples, the release of simultaneously produced CO2 in the dehydrogenation process results in not only undesirable carbon emissions but also inferior H2 purity for subsequent fuel cell applications13,15. Consequently, there is an actual demand for carbon-neutral hydrogen storage and release technologies that integrate the carbon capture (recycling the released CO2, dotted arrows in Fig. 1b) or in an ideal condition avoid CO2 release in the dehydrogenation reactions (bold arrows in Fig. 1b).

The homogeneous catalyst development for the two individual steps of chemical hydrogen storage and release—CO2 hydrogenation to FA23 and its dehydrogenation processes7—focused traditionally on noble metals. However, non-noble metal-based complexes have been proven to be valuable, more recently24,25. In this context, apart from iron9,26, manganese especially is of important interest, owing to its abundant, non-toxic, biocompatible and environmentally friendly features27. However, Mn pincer complexes were first recognized for catalytic hydrogenation reactions in 201628,29,30,31,32. Thus, examples of homogeneous Mn catalysts for CO2 hydrogenation to FA33,34,35,36,37 and its dehydrogenation38,39,40,41 are rather limited and often come with far lower catalytic efficiency compared with their noble-metal counterparts (Fig. 2). Most important of all, Mn catalysts are rarely reported so far to be efficient in both CO2 hydrogenation to FA and its dehydrogenation under unified reaction conditions (same catalyst, base, solvent and so on). Recently, a ruthenium-catalysed CO2 fixation to formates in the presence of amino acids (AAs) was reported by our group42. We speculated that such a system could also be used to develop state-of-the-art carbon-neutral hydrogen storage and release instead of using conventional CO2 absorbents, for example, alkanolamines43. Among the tested AAs, lysine (Lys) offers several advantages for the fixation of CO2 as it is an essential AA, industrially produced from microbial fermentation with >2.2 million tonnes scale per year44.

Hydrogenation reactions are shown on the left side of the figure under the green bar, and dehydrogenation is shown on the right under the pink bar. DBU: 1,8-diazabicyclo(5.4.0)undec-7-ene, DMOA: N,N-dimethyloctylamine, TEA: triethylamine, AA: α-amino acid, FA: formic acid, TONmax: turnover number, TTON: total turnover number.

Herein we provide a sustainable hydrogen storage and release method integrating the reversible hydrogenation of CO2 to FA and carbon capture processes. In our catalytic system, we use the α-amino acid Lys and a specific manganese complex as the carbon absorbent and catalyst. We show that such a system can catalyse the reversible hydrogenation of CO2 to FA in high productivities. With potassium lysinate, we achieve high evolution efficiency of hydrogen and retain CO2 inside of the cycles.

Initially, using several manganese, iron, cobalt, and rhenium complexes with bidentate and tridentate ligands, we performed the Lys-promoted CO2 hydrogenation (Supplementary Table 1). Under standard reaction conditions, CO2 (20 bar) hydrogenation happened in the presence of Lys (5.0 mmol) in H2O:THF (5:5 ml, THF: tetrahydrofuran) with 60 bar H2 at 145 °C and 12 h (Table 1 and Supplementary Figs. 3–5 and 50). Among all the tested catalysts, Mn–PNP–Br complexes 1–3, and 5–7 proved to be the most efficient ones (PNP: PNP pincer ligand). Remarkably, using Mn-1, 2 and 3 (refs. 28,45,46) (0.10 μmol, 20 ppm, based on Lys), up to 4.0 mmol Lys ammonium formate were produced (80% yield based on Lys; turnover number (TON) 40,000, Table 1, entries 1–3). On the other hand, Mn-4, bearing t-butyl substituents at the phosphorus atoms, resulted in no detectable formate. To our delight, Mn–PN5PiPr–Br complexes Mn-5, 6 and 7 (ref. 29) led to improved formate yields up to 88% and TONs up to 44,000 (Table 1, entries 5–7). Decreasing the catalyst loading to only 0.02 μmol (4 ppm), Mn-6 gave formate in slightly lower yield of 60%, but a TON of 150,000 was achieved (Table 1, entry 9). Applying the same quantity of Mn-5 and Mn-7, however, resulted in much lower yields (Table 1, entries 8 and 10). Therefore, we focused on Mn-6 in subsequent experiments. Fortunately, formate was produced even at 85 °C and 115 °C, leading to high yields (80% and 92%, respectively) and TONs (200,000 and 230,000, respectively, Table 1, entries 11 and 12). However, reducing the initial hydrogen pressure from 60 bar to 40 bar drastically decreased both yield and TON (Table 1, entry 13). Utilizing [Mn(CO)5Br] or other Mn pincer and bidentate complexes and the respective iron, cobalt and rhenium analogues as catalysts, formate was detected in <1% yield (Supplementary Table 1, entries 8–18).

CO2 hydrogenation in the presence of 16 other AAs and organic and inorganic bases led to much lower formate yields, revealing the specific role of Lys in CO2 hydrogenation reactions (Supplementary Tables 2–3). Only when using another basic AA arginine (Arg) bearing a guanidino group in its side chain, formate is produced in 62% yield.

Next, we investigated the H2 evolution from formic acid (5.0 mmol) in the presence of Lys (1.0 equivalent) at 90 °C using the Mn catalysts previously active in the hydrogenation step. The Mn-PN(H)P–Br complexes Mn-1, 2 and 3 promoted the H2 formation in yields up to 70% equivalent to 3.5 mmol H2 and TON 17,500 (Table 2, entries 1–3). Again, Mn–PN5PiPr–Br (Mn-6) catalysed the H2 release more efficiently with quantitative H2 yield and a TON of 29,400, while Mn-5 and Mn-7 gave obviously lower productivities (Table 2, entries 4–6). Lowering the quantity of Mn-6 to 0.10 μmol (20 ppm), the conversion was incomplete with much lower productivity (Table 2, entry 7). Decreasing the temperature to 85 °C, 74% yield and TON >21,000 were still obtained (Table 2, entry 8). Compared with Lys, Arg was found to be less active, leading to H2 in 46% yield (Table 2, entry 9). Utilizing the potassium salt of Lys (LysK), a high yield of H2 also was achieved (94%, Table 2, entry 10). In addition, Table 2 shows the H2:CO2 molar ratios in the gas products that contain generally less than 50% CO2, demonstrating again the good CO2 capture effect induced by Lys in the FA dehydrogenation reactions. Notably, LysK led to >99.9% CO2 retention, which offers the possibility for further reuse of CO2.

Additionally, blank reactions reveal that in the absence of Mn-6 or Lys, no product was detected (Supplementary Tables 4 and 9). Applying FA/LysK mixtures in the presence of other organic solvents, for example, 2-methyl-THF, triglyme and ethanol, H2 was produced in up to 88% yield with >99% CO2 retention (Supplementary Table 14 and Supplementary Fig. 38). In addition, H2 evolution from FA/LysK mixtures in an open system (using a manual burette) presented a complete FA conversion within 3 h, which resulted in a higher CO2 ratio (14%, Supplementary Figs. 39–41). Time-dependent experiments of CO2 hydrogenation and FA dehydrogenation applying Lys and Mn-6 were performed (Supplementary Table 16). Slightly decreased formate yield in the CO2 hydrogenation reaction was observed at 90% in 6 h, which dropped to 12% in 3 h. On the other hand, performing the FA dehydrogenation reaction in a shorter time led to decreased H2 yields (72% in 6 h and 4% in 3 h).

To avoid extra addition of catalysts or bases between the H2 storage–release cycles, we investigated the reusability of catalyst and base beforehand (Fig. 3). Based on the biphasic solvent system (Supplementary Fig. 26), we achieved a facile recycling of the Mn catalyst and organic solvent by separation of the organic and aqueous layers. Therefore, catalyst Mn-6 could be reused for ten consecutive runs in hydrogenation of CO2 to formate reactions. About 80% of the initial productivity remains after ten runs, resulting in a remarkable total TON (TTON) of 2,050,000 for formate production (Fig. 3a and Supplementary Table 8). The decrease of formate yield is probably due to the incomplete catalyst separation between each run rather than catalyst deactivation as no free phosphine ligand is detected by 31P NMR in the separated organic phase.

a,b, The reusable performance of Mn-6 catalyst in CO2 hydrogenation to formate (a) and FA dehydrogenation (b). Reaction conditions adopted from Table 1, entry 11, and Table 2, entry 5, respectively. The dotted lines serve as guides to the eye.

Then, we evaluated the reusability of Mn-6 as well as Lys in FA dehydrogenation (Fig. 3b). After each run, a new batch of FA was reloaded into the reaction mixture (first to fifth runs: 5.0 mmol FA and sixth to tenth runs: 20.0 mmol FA). Following this procedure, Mn-6 catalyst and Lys were reused for ten consecutive runs. Fortunately, more than 89% of the theoretical H2 productivity was achieved, resulting in an excellent TTON of 676,700 for hydrogen. Moreover, the H2:CO2 molar ratios in the gas phase are found to be approaching the theoretical value 50:50 after ten runs (Supplementary Table 10 and Supplementary Figs. 29–31). Besides recycling the Mn-6/Lys catalyst system in up to ten runs, the stability in long-term storage was evaluated. Quantitative H2 yield in FA dehydrogenation was obtained after storing the corresponding reaction solution for two weeks at room temperature under argon. Overall, these results demonstrate the high stability and reusability of the Mn-6 catalyst in both the CO2 hydrogenation and FA dehydrogenation processes.

Next, we combined CO2 hydrogenation to formate and its dehydrogenation applied Lys and Mn-6 complex. Due to the CO2 loss in the H2 release step, CO2 was reloaded for each hydrogenation step (Fig. 4a). The H2 storage–release cycles start from the dehydrogenation of FA: Mn-6 (0.17 μmol), FA:Lys (5.0:5.0 mmol), THF:H2O (5:5 ml) in a 100 ml autoclave at 90 °C for 12 h. The inside pressure was then carefully released to the manual burettes at room temperature and analysed by GC. Then, CO2 is reloaded under CO2 pressure (20 bar for 0.5 h or 2 bar for 6 h). Afterwards, the autoclave was filled with H2 (80 bar) and heated at 85 °C for 12 h. After the completion of the H2 storage step, the reactor was subjected to H2 release. Following this procedure, ten consecutive cycles in total were performed with >90% yield of H2 evolution (CO2 reloading at 20 bar). Using lower CO2 (2 bar) pressure, an average hydrogen yield of >82% is still achieved. Notably, we performed this carbon-neutral hydrogen storage–release methodology even with CO2 from air (reloading at ambient conditions for 24 h). Following this concept, over 72% yield of H2 was obtained in ten consecutive cycles (Fig. 4c, Supplementary Table 11 and Supplementary Figs. 32–34). Detailed studies of CO2 capture with Lys under various conditions show that carbamate species are first produced under air, then are converted further to bicarbonate at higher CO2 pressure (Supplementary Tables 5–7 and Supplementary Figs. 6–25).

a,b, Illustration of the carbon-neutral hydrogen storage and release applied with Lys, including CO2 capture from air (a) and the simplified hydrogen storage and release system based on LysK without CO2 reloading (b). c,d, H2 evolution in the H2 storage–release cycles combined with CO2 capture using Lys; CO2 is reloaded in each cycle after H2 release with either pure CO2 (20 bar or 2 bar) or air. H2 evolution using LysM (M = K, Na, Li) without CO2 reloading (c). H2 evolution in upscaled H2 storage–release cycles applying 5.0 mmol, 20.0 mmol and 90.0 mmol LysK without CO2 reloading (d).

Even though we performed the hydrogen storage–release cycles with Lys in excellent efficiency, the development of a practical rechargeable chemical hydrogen battery is not yet achieved. Ideally, such devices consist of closed autoclaves containing the hydrogen storage media (in this case, CO2), where hydrogen is charged and discharged conveniently. Obviously, the reloading of the storage media in this system between each cycle should be avoided as far as possible because it requires additional feedstock, energy input and reaction steps. To overcome such issues and demonstrate our main goal of a stable and practical chemical hydrogen storage and release with efficient H2 storage–release cycles (Fig. 4b), we tested different lysinate salts (LysM, M = K, Na, Li; Fig. 4c and Supplementary Table 12), due to the excellent performance of LysK in capturing CO2 (Table 2, entry 10). Indeed, LysK, LysNa and LysLi led to much higher CO2 retention (>99.9%) during the H2 evolution process compared with Lys (Supplementary Figs. 35–37). Hence, applying LysM instead of Lys, the H2 evolution efficiency reached ≥80% (LysK), ≥60% (LysNa) and ≥46% (LysLi), respectively, in ten charge–discharge cycles. The observed cation-based effect is consistent with the observation of a previous work for the hydrogenation of imines, where potassium tert-butoxide led to a higher hydrogenation rate than the sodium one47. Notably, CO2 reloading was not necessary in H2 storage–release cycles applying LysM, thus greatly simplifying the process. After one cycle of hydrogen release and storage, 4.6 mmol formate (92%) was obtained applying 5 mmol LysK. As a prototype example (0.3 l autoclave), the H2 evolution process applying LysK was scaled up to 90.0 mmol without obvious drop of the efficiency in at least ten charge–discharge cycles (Fig. 4d and Supplementary Table 13), demonstrating its applicability.

To understand both catalyst and CO2 capture efficiency, we tried to detect the main reaction intermediates. Thus, we performed stoichiometric reactions in Young–NMR tubes and analysed them by in situ NMR (Fig. 5a). After adding LysK (1.0 equivalent) to the solution of Mn-5 in THF-d8:D2O (0.4:0.1 ml) at room temperature, a new signal at 131.3 ppm in 31P NMR was observed, which is assigned to I-1 (ref. 48). Next, this mixture was pressurized with H2 (0.5 bar), and a Mn–H species was identified with new signals of 31P NMR at 162.2 ppm and 1H NMR at −5.99 ppm (JH–P = 51.0 Hz, triplet, JH–P: H–P coupling constant), which is in accordance with a similar hydride complex previously reported47,48,49. After adding CO2 (0.5 bar) to the tube, a weak signal at 136.9 ppm (31P NMR) and 8.73 ppm (1H NMR) appeared, corresponding to the Mn–OOCH species that is considered to be poorly soluble33, meanwhile the hydride signals at 162.2 ppm (31P NMR) and −5.99 ppm (1H NMR) disappeared. Alternatively, by adding FA (1.0 equivalent) to the mixture of Mn-5 and LysK, a similar peak was observed (Supplementary Figs. 42–49). Control experiments of CO2 hydrogenation in the presence of various lysine derivatives (Supplementary Table 15) revealed the presence of the α-amino acid group and an appropriate basic side chain in the amine molecule that are crucial for both the Mn-catalysed CO2 hydrogenation reaction and CO2 capture processes.

a,b, In situ 31P NMR studies in Young–NMR tubes (f1: chemical shifts are shown in ppm); a) and the proposed catalytic cycle for reversible CO2 hydrogenation catalysed by Mn–PN5PiPr complexes (b).

On the basis of the in situ NMR studies, control experiments above and the previously reported research work33,47,48,49, we propose the following catalytic cycle on the reversible CO2 hydrogenation catalysed by Mn–PN5PiPr complexes (Fig. 5b). The catalyst precursor is first activated by an excess of LysK via N–H deprotonation and dearomatisation of the triazine moiety, leading to active species I-1, which is described as a bimetallic Mn–K species50,51. During the H2 storage step (green pathway), dihydrogen is activated via heterolytic cleavage resulting in a Mn-hydride species I-2 and transforming LysK into Lys. After insertion of C = O bond in CO2 to Mn–H I-2, the Mn–OOCH intermediate I-3 is formed. Then, HCOOH is liberated by transforming 1 equivalent of Lys to LysK, which regenerates the active species I-1 after dearomatisation of the triazine moiety. The reverse reaction (pink pathway) enables the HCOOH dehydrogenation to H2 and CO2, accordingly.

The present work describes a concept for carbon neutral hydrogen storage and release. Utilizing a molecularly defined manganese complex in the presence of naturally occurring Lys, high efficiency for direct CO2 hydrogenation to formate is achieved (93% yield; 2,000,000 TTON). On the other hand, the same system promotes H2 generation from FA in the presence of Lys with >99% yield and a TTON of 600,000. This reaction system exhibited high stability and reusability. On the basis of these results, the combination of the individual processes was realised. Notably, we performed such hydrogen storage–release cycles without the addition of extra AA, catalyst and especially CO2 due to the excellent CO2 capture effect (99.9%) of LysK.

It is noteworthy that the CO content in the produced H2 gas was below 10 ppm throughout the process. We successfully scaled up this method without obvious drop of the productivity in at least ten charge–discharge cycles. The current methodology represents one of the most productive combinations of CO2 valorization and FA dehydrogenation applying homogeneous non-noble metal-based catalysts. The results inspire further research towards practical applications and pave the way for building up a carbon-neutral chemical hydrogen storage and release set-up by employing non-hazardous AAs and benign catalysts.

Reagents were purchased from commercial suppliers and used without further purification, including L-lysine (fluorochem, 97%), L-arginine (Alfa Aesar, >98%), L-tyrosine (Tokyo Chemical Industry (TCI), >98.5%), L-threonine (TCI, >99%), L-methionine (Alfa Aesar, >98%), L-glutamic acid (TCI, >99%), L-serine (TCI, >99%), L-proline (Acros Organics, >99%), L-cysteine (TCI, >98%), L-histidine (Sigma-Aldrich, >99%), L-tryptophan (TCI, >98.5%), glycine (Merck, >99.7%), L-glutamine (TCI, >99%), L-norleucine (TCI, >99%), 1,5-diaminopentane (TCI, >98%), 6-aminohexanoic acid (Alfa Aesar, 99%), 2,3-diaminopropanoic acid (fluorochem, 95%), Nα-(t-butoxycarbonyl)-Lys (fluorochem, 95%), Nε-(t-butoxycarbonyl)-Lys (fluorochem, 97%), tetramethylguanidine (Alfa Aesar, 99%), L-lysine methyl ester (Alfa Aesar, 99%), pentaethylenehexamine (Sigma-Aldrich, >98%), 1,4-diazabicyclo[2.2.2]octane (TCI, > 98%), 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU, TCI, > 98%), 1-butylamine (Acros Organics, 99.5%), FA (FA, Sigma-Aldrich, 98%), potassium bicarbonate (Sigma-Aldrich, 99.5%), potassium hydroxide (Fisher Chemical, 86.4%), sodium hydroxide (Fisher Chemical, 99.3%), lithium hydroxide (Sigma-Aldrich, 98%), deuterium oxide (Deutero, D-99.9%), tetrahydrofuran-d8 (Deutero, D-99.5%), Mn(CO)5Br (Stream, 98%), H2 (Air Liquide, grade 5.0) and CO2 (Linde, grade 4.8). Organic solvents were collected from an SPS (Solvent Purification System) machine and stored under argon with drying reagent (molecular sieve). Deionized (DI) water was used for CO2 capture and hydrogenation/dehydrogenation reactions. Organometallic complexes Mn-1 (refs. 28,52), Mn-2 (refs. 28,52), Mn-3 (refs. 46,53), Mn-4 (ref. 45), Mn-5 (ref. 29), Mn-6 (ref. 29), Mn-7 (ref. 29), Mn-8 (ref. 54), Mn-9 (refs. 45,54), Mn-10 (ref. 55), Mn-11 (ref. 56), Mn-12 (ref. 34), Fe-1 (ref. 57), Fe-2 (ref. 57), Fe-3 (refs. 58,59), Co-1 (ref. 60) and Re-1 (refs. 61,62) were synthesized according to the corresponding publications and stored under argon with light exclusion. The manganese, rhenium and cobalt complexes were prepared by adding Mn(CO)5Br (137.4 mg, 0.5 mmol), Re(CO)5Br (203.1 mg, 0.5 mmol) or CoBr2 (109.4 mg, 0.5 mmol) in THF (10 ml), respectively, dropwise to 1.1 equivalenent corresponding ligands in THF then heated and stirred at 100 °C for 20 h under argon. The reaction mixture was cooled to room temperature and concentrated in vacuo. The crude mixture was then thoroughly washed with pentane and dried under vacuum, yielding the corresponding complexes as powder (yellow for manganese complexes, white for rhenium complex and purple for cobalt complex). Fe-3 is prepared by adding 1.1 equivalent corresponding ligands in THF to anhydrous FeBr2 (107.8 mg, 0.5 mmol) in THF (10 ml), and the argon inside the flask was replaced with CO by performing freeze–pump–thaw cycles. The solution was stirred at room temperature for 4 h, followed by removal of solvent under vacuum. The resulting residue was washed with pentane and dried under vacuum to generate the corresponding compound as a blue powder. Fe-2 and Fe-3 were prepared by adding 1 and 5 equivalent of NaBH4 (18.9 mg, 0.5 mmol and 94.6 mg, 2.5 mmol), respectively, into Fe-3 (274.5 mg, 0.5 mmol) in EtOH (35 ml). The reaction mixture was then stirred for 5 h at room temperature. The solvent was removed in vacuo and the residue dissolved in toluene. The resulting suspension was filtered through a short pad of celite. The filtrate was concentrated in vacuo and the resulting solid washed thoroughly with pentane to create a yellow powder. Unless otherwise stated, all reactions were conducted under an argon atmosphere. 1H, 13C and 31P NMR spectroscopy were recorded using Bruker AV 300 MHz and Bruker AV 400 MHz spectrometers. All NMR data are expressed as chemical shift in ppm. 1H and 13C NMR chemical shifts were determined relative to the internal standard DMF (7.92 ppm and 165.53 ppm, respectively) or 1,4-dioxane (3.75 ppm and 67.19 ppm, respectively) in D2O. 31P NMR spectra were calibrated with an external H3PO4 standard (0 ppm). 13C NMR-quant measurements were performed with a Bruker AV 400 MHz spectrometer, relaxation delay = 20 s (relaxation delay > 20 s did not change the integration), number of scans = 512, acquisition time = 1.1141 s (refs. 63,64,65,66). NMR spectra were interpreted and processed using MestReNova (version 14.0.1–23559). Gas chromatography (Agilent Technologies 7890 A GC system, Carboxen/TCD) was used to analyse the content of the gas phase with a CO quantification limit of 10 ppm. pH values were measured on a laboratory digital pH meter (Mettler–Toledo AG, SevenEasy pH 8603) at room temperature (24 °C).

The respective amount of catalyst (0.10 μmol or 0.02 μmol) was dispensed from the stock solution; Lys (5.0 mmol), THF (5 ml) and H2O (5 ml) were added to a 50 ml autoclave equipped with a magnetic stir bar. After pressurizing the reactor with CO2 and H2 gas, the reaction mixture was heated and stirred on a pre-heated oil bath for 12 h. Then the reactor was cooled to room temperature and the inside pressure was carefully released. A biphasic reaction mixture was obtained containing a transparent organic upper layer and an aqueous yellow lower layer. Addition of DI water (~3 ml) to the above-mentioned mixture resulted in a homogeneous solution (Supplementary Figs. 3–4) (ref. 42). DMF (250 μl, 3.24 mmol) was added as an internal standard to the reaction mixture. The reaction mixture was then analysed by 1H NMR with a few drops of D2O (~1 ml) to lock the signals34. Yield of formate is calculated by (mmol formate)/(mmol Lys) × 100%. TON of formate is calculated by (mmol formate)/(mmol catalyst). The gas mixtures were analysed by GC, and no concomitant of CO/CH4 was found.

Appropriate amount of catalyst (0.10–0.20 μmol) was dispensed from the stock solution, FA (5.0 mmol), Lys (5.0 mmol), THF (5 ml) and H2O (5 ml) were added to a 100 ml autoclave equipped with a magnetic stir bar. The reaction mixture was then heated and stirred on a pre-heated oil bath for 12 h. The reactor was cooled to room temperature, and the inside pressure was released carefully to the manual burettes. A 5 ml degassed syringe was used to obtain a gas sample analysed by GC. CO is not detectable in all cases (below the CO quantification limit of 10 ppm). Yield of H2 is calculated by (mmol H2)/(mmol HCOOH) × 100%. TON of H2 is calculated by (mmol H2)/(mmol catalyst).

The gas evolution was corrected with the blank volume (18 ml), which corresponds to the gas evolution of the reaction without any catalyst. H2 volume, \(V_{{\mathrm{H}}_{2}}\) , and CO2 volume, \(V_{{\mathrm{CO}}_{2}}\) , are calculated with the following equation:

Moles of H2, \(n_{{\mathrm{H}}_{2}}\) , and moles of CO2, \(n_{\mathrm{CO}_{2}}\) , are calculated with the following equation:

Hydrogen yield \(Y_\mathrm{{H}_{2}}\) is calculated with the following equation:

The turnover number (TON) of H2 is calculated with the following equation:

\(V_\mathrm{obs}\) is the gas evolution volume of the catalytic reaction measured in the manual burettes. \(V_{\mathrm{blank}}\) is the gas evolution volume of the blank reaction measured in the manual burettes. \(\% V_\mathrm{{H}_{2}}\) and \(\% V_\mathrm{{CO}_{2}}\) are the volume ratios of H2 and CO2, respectively, determined by GC. \(n_\mathrm{FA}\) and \(n_\mathrm{cat}\) are the moles of FA and catalyst, respectively. \(V_\mathrm{{m}_{{H}_{2}(25\,^\circ C)}}\) and \(V_\mathrm{{m}_{{CO}_{2}(25\,^\circ C)}}\) are the molar volumes of H2 and CO2 at room temperature (25 °C), respectively, calculated with the Van Der Waals equation (below).

Calculation of H2 molar volume \(V_{m_{{\mathrm{H}}_2(25^\circ C)}}\) :

Where R = 8.3145 m3·Pa·mol−1·K−1, T = 273.15 + room temperature (°C) K, P = 101325 Pa, a = 24.9×10−3 Pa·m6·mol−2, b = 26.7×10−6 m3·mol−1.

Calculation of CO2 molar volume \(V_\mathrm{{m}_{{CO}_{2}(25\,^\circ C)}}\) :

where R = 8.3145 m3 Pa mol−1 K−1, T = 273.15 + room temperature (°C) (K), P = 101,325 Pa, a = 36.5 × 10−2 Pa m6 mol−2 and b = 42.7 × 10−6 m3 mol−1.

Catalyst Mn-6 (0.02 μmol, dispensed from the stock solution), Lys (730.9 mg, 5.0 mmol), THF (5 ml) and H2O (5 ml) were added to a 50 ml autoclave equipped with a magnetic stir bar. After pressurizing the reactor with CO2 (20 bar) and H2 (60 bar) gas, the reaction mixture was heated and stirred on a pre-heated oil bath at 115 °C for 12 h. Then the reactor was cooled to room temperature, and the inside pressure was carefully released. A biphasic reaction mixture was obtained containing a transparent organic upper layer and an aqueous yellow lower layer (Supplementary Fig. 26). After the separation of the two layers, the aqueous lower layer was washed and extracted with additional THF (1 ml × 2), and the combined organic parts were used directly for the next run by adding a new batch of 5.0 mmol Lys. Meanwhile, DMF (250 μl, 3.24 mmol) was added into the aqueous lower layer as an internal standard, then analysed by 1H NMR with a few drops of D2O (~1 ml) to lock the signals34. A total of ten runs of CO2 hydrogenation to formate were performed by using the above-mentioned procedure. Yield of formate is calculated by (mmol formate)/(mmol Lys) × 100%. TON of formate is calculated by (mmol formate)/(mmol catalyst). The gas mixtures were analysed by GC, and no concomitant of CO/CH4 was found.

Catalyst Mn-6 (0.1 mg, 0.17 μmol), FA, Lys (730.9 mg, 5.0 mmol), THF (5 ml) and H2O (5 ml) were added to a 100 ml autoclave equipped with a magnetic stir bar. The reaction mixture was then heated and stirred on a pre-heated oil bath for 12 h. The reactor was cooled to room temperature and the inside pressure was released carefully to the manual burettes, and the content of the gas phase was analysed by GC. CO was not detectable in all cases (below the CO quantification limit of 10 ppm). After each run, a new batch of FA (first to fifth runs: 5.0 mmol FA and sixth to tenth runs: 20.0 mmol FA) was loaded into the reaction mixture, then heated and stirred on a pre-heated oil bath for 12 h. Following this procedure, a total of ten runs of catalytic dehydrogenation of FA in the presence of Lys were performed. Yield of H2 is calculated by (mmol H2)/(mmol HCOOH) × 100%. TON of H2 is calculated by (mmol H2)/(mmol catalyst).

The H2 storage–release cycles start from the dehydrogenation of FA (H2 release): Mn-6 (0.1 mg, 0.17 μmol), FA (188.6 μl, 5.0 mmol), Lys (730.9 mg, 5.0 mmol), THF (5 ml) and H2O (5 ml) were added to a 100 ml autoclave equipped with a magnetic stir bar. The reaction mixture was then heated and stirred on a pre-heated oil bath at 90 °C for 12 h. The reactor was cooled to room temperature, and the inside pressure was released carefully to the manual burettes. The content of the gas phase was analysed by GC. CO was not detectable in all cases (below the CO quantification limit of 10 ppm). Then CO2 is replenished under the following conditions—capture with CO2 (20 bar or 2 bar): after the completion of the above-mentioned FA dehydrogenation, the 100 ml autoclave was charged with CO2 (20 bar for 0.5 h or 2 bar for 6 h). After releasing the overpressure of CO2, the 100 ml autoclave was filled with 80 bar of H2. Then the reaction mixture was heated and stirred on a pre-heated oil bath at 85 °C for 12 h. Afterwards, the reactor was cooled to room temperature, and the inside pressure was released carefully. Then the autoclave was subjected to the H2 release procedure. Following this procedure, a total of ten runs were performed via CO2 recharging under 20 bar or 2 bar of CO2. Capture from ambient air: after the completion of FA dehydrogenation, a biphasic reaction mixture containing a transparent organic upper layer and a pale yellow aqueous lower layer was obtained. After the separation of the two layers, the aqueous lower layer was washed with additional THF (1 ml × 2), and the combined organic layer was reserved for the following H2 storage step. On the other hand, the aqueous layer was subjected to CO2 capture from ambient air at room temperature for 24 h. After bubbling with argon for 0.5 h, this aqueous layer was combined with the above-mentioned organic layer and loaded to a 100 ml autoclave. The autoclave was then filled with 80 bar of H2 and heated and stirred on a pre-heated oil bath at 85 °C for 12 h. The reactor was cooled to room temperature, and the inside pressure was released carefully. Then the autoclave was subjected to the H2 release procedure. Following this procedure, a total of ten runs were performed with CO2 capture from ambient air. It should be noted that no H2 evolution was observed at room temperature in the H2 storage–release cycles for up to 12 h.

The lysinate salts (LysM, 5.0 mmol) was prepared via stirring Lys (5.0 mmol) with 1.0 equivalent of corresponding alkali metal hydroxides MOH (KOH, NaOH, LiOH) for 30 min in 1 ml DI water at room temperature before the H2 storage–release cycles. The H2 storage–release cycles start from the dehydrogenation of FA (H2 release): Mn-6 (0.1 mg, 0.17 μmol), FA (188.6 μl, 5.0 mmol), LysM (5.0 mmol), THF (5 ml) and H2O (4 ml) were added to a 100 ml autoclave equipped with a magnetic stir bar. The reaction mixture was then heated and stirred on a pre-heated oil bath at 90 °C for 12 h. The reactor was cooled to room temperature, and the inside pressure was released carefully to the manual burettes and the content of the gas phase was analysed by GC. CO was not detectable in all cases (below the CO quantification limit of 10 ppm). After the completion of FA dehydrogenation, the 100 ml autoclave was filled with 80 bar of H2. Then the reaction mixture was heated and stirred on a pre-heated oil bath at 85 °C for 12 h. Afterwards, the reactor was cooled to room temperature, and the inside pressure was released carefully. Then the autoclave was subjected to the H2 release procedure. Following this procedure, a total of ten runs were performed with 5.0 mmol LysK, LysNa, and LysLi.

The 20.0 mmol lysine potassium salt (LysK, 20.0 mmol) was prepared via stirring Lys (2,923.8 mg, 20.0 mmol) with 1.0 equivalent of KOH for 30 min in 10 ml DI water at room temperature before the H2 storage–release cycles. The H2 storage-release cycles start from the dehydrogenation of FA (H2 release); Mn-6 (0.4 mg, 0.68 μmol), FA:LysK (20.0:20.0 mmol), THF (20 ml) and H2O (10 ml) were added to a 300 ml autoclave equipped with a magnetic stir bar. The reaction mixture was then heated and stirred on a pre-heated oil bath at 90 °C for 12 h. The reactor was cooled to room temperature, and the inside pressure was released carefully to the manual burettes and the content of the gas phase was analysed by GC. CO is not detectable in all cases (below the CO quantification limit of 10 ppm). After the completion of FA dehydrogenation, the 300 ml autoclave was filled with 80 bar of H2. Then the reaction mixture was heated and stirred on a pre-heated oil bath at 85 °C for 12 h. Afterwards, the reactor was cooled to room temperature, and the inside pressure was released carefully. Then the autoclave was subjected to the H2 release procedure. Following this procedure, a total of ten runs were performed with LysK in 20.0 mmol scale. The 90 mmol lysine potassium salt (LysK, 90.0 mmol) was prepared via stirring Lys (90.0 mmol) with 1.0 equivalent of KOH for 30 min in 20 ml DI water at room temperature before the H2 storage–release cycles. The H2 storage–release cycles started from the dehydrogenation of FA (H2 release); Mn-6 (1.8 mg, 3.05 μmol), FA:LysK (90.0:90.0 mmol), THF (30 ml) and H2O (60 ml) were added to a 300 ml autoclave equipped with a magnetic stir bar. The reaction mixture was then heated and stirred on a pre-heated oil bath at 90 °C for 12 h. The reactor was cooled to room temperature, and the inside pressure was released carefully to the manual burettes and the content of the gas phase was analysed by GC. CO was not detectable in all cases (below the CO quantification limit of 10 ppm). After the completion of FA dehydrogenation, the 300 ml autoclave was filled with 80 bar of H2. Then the reaction mixture was heated and stirred on a pre-heated oil bath at 85 °C for 12 h. Afterwards, the reactor was cooled to room temperature, and the inside pressure was released carefully. Then the autoclave was subjected to the H2 release procedure. Following this procedure, a total of ten runs were performed with LysK at 90.0 mmol scale.

All the relevant data are included in the published article and its Supplementary Information.

Tans, P. Trends in atmospheric carbon dioxide. NOAA/GML https://gml.noaa.gov/ccgg/trends/ (2022).

Keeling, R. Carbon dioxide measurements. Scripps Institution of Oceanography https://scrippsco2.ucsd.edu/ (2022).

Roger, I., Shipman, M. A. & Symes, M. D. Earth-abundant catalysts for electrochemical and photoelectrochemical water splitting. Nat. Rev. Chem. 1, 0003 (2017).

Schneidewind, J., Argüello Cordero, M. A., Junge, H., Lochbrunner, S. & Beller, M. Two-photon, visible light water splitting at a molecular ruthenium complex. Energy Environ. Sci. 14, 4427–4436 (2021).

Yadav, M. & Xu, Q. Liquid-phase chemical hydrogen storage materials. Energy Environ. Sci. 5, 9698–9725 (2012).

Chatterjee, S., Dutta, I., Lum, Y., Lai, Z. & Huang, K.-W. Enabling storage and utilization of low-carbon electricity: power to formic acid. Energy Environ. Sci. 14, 1194–1246 (2021).

Mellmann, D., Sponholz, P., Junge, H. & Beller, M. Formic acid as a hydrogen storage material—development of homogeneous catalysts for selective hydrogen release. Chem. Soc. Rev. 45, 3954–3988 (2016).

Onishi, N., Laurenczy, G., Beller, M. & Himeda, Y. Recent progress for reversible homogeneous catalytic hydrogen storage in formic acid and in methanol. Coord. Chem. Rev. 373, 317–332 (2018).

Sordakis, K. et al. Homogeneous catalysis for sustainable hydrogen storage in formic acid and alcohols. Chem. Rev. 118, 372–433 (2018).

Leitner, W., Dinjus, E. & Gaßner, F. Activation of carbon dioxide: IV. Rhodium-catalysed hydrogenation of carbon dioxide to formic acid. J. Organomet. Chem. 475, 257–266 (1994).

Boddien, A. et al. CO2-“neutral” hydrogen storage based on bicarbonates and formates. Angew. Chem. 50, 6411–6414 (2011).

Papp, G., Csorba, J., Laurenczy, G. & Joó, F. A charge/discharge device for chemical hydrogen storage and generation. Angew. Chem. 50, 10433–10435 (2011).

Boddien, A. et al. Towards the development of a hydrogen battery. Energy Environ. Sci. 5, 8907–8911 (2012).

Filonenko, G. A., van Putten, R., Schulpen, E. N., Hensen, E. J. M. & Pidko, E. A. Highly efficient reversible hydrogenation of carbon dioxide to formates using a ruthenium PNP-pincer catalyst. ChemCatChem 6, 1526–1530 (2014).

Hsu, S.-F. et al. A rechargeable hydrogen battery based on Ru catalysis. Angew. Chem. 53, 7074–7078 (2014).

Kothandaraman, J. et al. Amine-free reversible hydrogen storage in formate salts catalyzed by ruthenium pincer complex without pH control or solvent change. ChemSusChem 8, 1442–1451 (2015).

Sordakis, K., Dalebrook, A. F. & Laurenczy, G. A viable hydrogen storage and release system based on cesium formate and bicarbonate salts: mechanistic insights into the hydrogen release step. ChemCatChem 7, 2332–2339 (2015).

Xin, Z. et al. Towards hydrogen storage through an efficient ruthenium-catalyzed dehydrogenation of formic acid. ChemSusChem 11, 2077–2082 (2018).

Hull, J. F. et al. Reversible hydrogen storage using CO2 and a proton-switchable iridium catalyst in aqueous media under mild temperatures and pressures. Nat. Chem. 4, 383–388 (2012).

Tanaka, R., Yamashita, M., Chung, L. W., Morokuma, K. & Nozaki, K. Mechanistic studies on the reversible hydrogenation of carbon dioxide catalyzed by an Ir-PNP complex. Organometallics 30, 6742–6750 (2011).

Forberg, D. et al. Single-catalyst high-weight% hydrogen storage in an N-heterocycle synthesized from lignin hydrogenolysis products and ammonia. Nat. Commun. 7, 13201 (2016).

Enthaler, S. et al. Exploring the reactivity of nickel pincer complexes in the decomposition of formic acid to CO2/H2 and the hydrogenation of NaHCO3 to HCOONa. ChemCatChem 7, 65–69 (2015).

Jessop, P. G., Joó, F. & Tai, C.-C. Recent advances in the homogeneous hydrogenation of carbon dioxide. Coord. Chem. Rev. 248, 2425–2442 (2004).

Bernskoetter, W. H. & Hazari, N. Reversible hydrogenation of carbon dioxide to formic acid and, ethanol: Lewis acid enhancement of base metal catalysts. Acc. Chem. Res. 50, 1049–1058 (2017).

Alig, L., Fritz, M. & Schneider, S. First-row transition metal (de)hydrogenation catalysis based on functional pincer ligands. Chem. Rev. 119, 2681–2751 (2019).

Wei, D. & Darcel, C. Iron catalysis in reduction and hydrometalation reactions. Chem. Rev. 119, 2550–2610 (2019).

Valyaev, D. A., Lavigne, G. & Lugan, N. Manganese organometallic compounds in homogeneous catalysis: past, present, and prospects. Coord. Chem. Rev. 308, 191–235 (2016).

Elangovan, S. et al. Selective catalytic hydrogenations of nitriles, ketones and aldehydes by well-defined manganese pincer complexes. J. Am. Chem. Soc. 138, 8809–8814 (2016).

Kallmeier, F., Irrgang, T., Dietel, T. & Kempe, R. Highly active and selective manganese C=O bond hydrogenation catalysts: the importance of the multidentate ligand, the ancillary ligands, and the oxidation state. Angew. Chem. 55, 11806–11809 (2016).

Mukherjee, A. & Milstein, D. Homogeneous catalysis by cobalt and manganese pincer complexes. ACS Catal. 8, 11435–11469 (2018).

Wang, Y., Wang, M., Li, Y. & Liu, Q. Homogeneous manganese-catalyzed hydrogenation and dehydrogenation reactions. Chem 7, 1180–1223 (2021).

Azouzi, K., Valyaev, D. A., Bastin, S. & Sortais, J.-B. Manganese—new prominent actor in transfer hydrogenation catalysis. Curr. Opin. Green Sustain. Chem. 31, 100511 (2021).

Bertini, F. et al. Carbon dioxide hydrogenation catalysed by well-defined Mn (I) PNP pincer hydride complexes. Chem. Sci. 8, 5024–5029 (2017).

Dubey, A., Nencini, L., Fayzullin, R. R., Nervi, C. & Khusnutdinova, J. R. Bio-inspired Mn(I) complexes for the hydrogenation of CO2 to formate and formamide. ACS Catal. 7, 3864–3868 (2017).

Kumar, A. et al. CO2 activation by manganese pincer complexes through different modes of metal–ligand cooperation. Dalton Trans. 48, 14580–14584 (2019).

Kostera, S. et al. Carbon dioxide hydrogenation to formate catalyzed by a bench-stable, non-pincer-type Mn(I) alkylcarbonyl complex. Organometallics 40, 1213–1220 (2021).

Schlenker, K. et al. Role of ligand-bound CO2 in the hydrogenation of CO2 to formate with a (PNP)Mn catalyst. ACS Catal. 11, 8358–8369 (2021).

Tondreau, A. M. & Boncella, J. M. 1,2-Addition of formic or oxalic acid to –N{CH2CH2(PiPr2)}2-supported Mn(I) dicarbonyl complexes and the manganese-mediated decomposition of formic acid. Organometallics 35, 2049–2052 (2016).

Andérez-Fernández, M. et al. A stable manganese pincer catalyst for the selective dehydrogenation of methanol. Angew. Chem. 56, 559–562 (2017).

Anderson, N. H., Boncella, J. & Tondreau, A. M. Manganese-mediated formic acid dehydrogenation. Chem. Eur. J. 25, 10557–10560 (2019).

Léval, A. et al. Hydrogen production from formic acid catalyzed by a phosphine free manganese complex: investigation and mechanistic insights. Green Chem. 22, 913–920 (2020).

Wei, D., Junge, H. & Beller, M. An amino acid based system for CO2 capture and catalytic utilization to produce formates. Chem. Sci. 12, 6020–6024 (2021).

Ghayur, A., Verheyen, T. V. & Meuleman, E. Biological and chemical treatment technologies for waste amines from CO2 capture plants. J. Environ. Manag. 241, 514–524 (2019).

Pandey, A. K., Pandey, K. & Singh, L. K. in Innovations in Food Technology (eds Mishra, P. et al.) 211–229 (Springer, 2020).

Papa, V. et al. Efficient and selective hydrogenation of amides to alcohols and amines using a well-defined manganese–PNN pincer complex. Chem. Sci. 8, 3576–3585 (2017).

Borghs, J. C., Lebedev, Y., Rueping, M. & El-Sepelgy, O. Sustainable manganese-catalyzed solvent-free synthesis of pyrroles from 1,4-diols and primary amines. Org. Lett. 21, 70–74 (2019).

Fertig, R., Irrgang, T., Freitag, F., Zander, J. & Kempe, R. Manganese-catalyzed and base-switchable synthesis of amines or imines via borrowing hydrogen or dehydrogenative condensation. ACS Catal. 8, 8525–8530 (2018).

Freitag, F., Irrgang, T. & Kempe, R. Mechanistic studies of hydride transfer to imines from a highly active and chemoselective manganate catalyst. J. Am. Chem. Soc. 141, 11677–11685 (2019).

Kallmeier, F., Dudziec, B., Irrgang, T. & Kempe, R. Manganese-catalyzed sustainable synthesis of pyrroles from alcohols and amino alcohols. Angew. Chem. 56, 7261–7265 (2017).

Zhang, G., Irrgang, T., Schlagbauer, M. & Kempe, R. Synthesis of 1,3-diketones from esters via liberation of hydrogen. Chem. Catal. 1, 681–690 (2021).

Schlagbauer, M., Kallmeier, F., Irrgang, T. & Kempe, R. Manganese-catalyzed β-methylation of alcohols by methanol. Angew. Chem. 59, 1485–1490 (2020).

Elangovan, S. et al. Efficient and selective N-alkylation of amines with alcohols catalysed by manganese pincer complexes. Nat. Commun. 7, 12641 (2016).

Xia, T. et al. Manganese PNP-pincer catalyzed isomerization of allylic/homo-allylic alcohols to ketones—activity, selectivity, efficiency. Catal. Sci. Technol. 9, 6327–6334 (2019).

Kaithal, A., Hölscher, M. & Leitner, W. Catalytic hydrogenation of cyclic carbonates using manganese complexes. Angew. Chem. 57, 13449–13453 (2018).

Perez, M., Elangovan, S., Spannenberg, A., Junge, K. & Beller, M. Molecularly defined manganese pincer complexes for selective transfer hydrogenation of ketones. ChemSusChem 10, 83–86 (2017).

van Putten, R. et al. Non-pincer-type manganese complexes as efficient catalysts for the hydrogenation of esters. Angew. Chem. 56, 7531–7534 (2017).

Chakraborty, S. et al. Iron-based catalysts for the hydrogenation of esters to alcohols. J. Am. Chem. Soc. 136, 7869–7872 (2014).

Alberico, E. et al. Selective hydrogen production from methanol with a defined iron pincer catalyst under mild conditions. Angew. Chem. 52, 14162–14166 (2013).

Werkmeister, S. et al. Hydrogenation of esters to alcohols with a well-defined iron complex. Angew. Chem. 53, 8722–8726 (2014).

Zhou, W. et al. Cobalt-catalyzed aqueous dehydrogenation of formic acid. Chem. Eur. J. 25, 8459–8464 (2019).

Wei, D., Roisnel, T., Darcel, C., Clot, E. & Sortais, J.-B. Hydrogenation of carbonyl derivatives with a well-defined rhenium precatalyst. ChemCatChem 9, 80–83 (2017).

Piehl, P., Peña-López, M., Frey, A., Neumann, H. & Beller, M. Hydrogen autotransfer and related dehydrogenative coupling reactions using a rhenium(I) pincer catalyst. Chem. Commun. 53, 3265–3268 (2017).

Barzagli, F., Mani, F. & Peruzzini, M. A 13C NMR study of the carbon dioxide absorption and desorption equilibria by aqueous 2-aminoethanol and N-methyl-substituted 2-aminoethanol. Energy Environ. Sci. 2, 322–330 (2009).

Barbarossa, V. et al. Efficient CO2 capture by non-aqueous 2-amino-2-methyl-1-propanol (AMP) and low temperature solvent regeneration. RSC Adv. 3, 12349–12355 (2013).

Perinu, C., Arstad, B. & Jens, K.-J. 13C NMR experiments and methods used to investigate amine-CO2-H2O systems. Energy Proc. 37, 7310–7317 (2013).

Kothandaraman, J., Goeppert, A., Czaun, M., Olah, G. A. & Prakash, G. K. S. Conversion of CO2 from air into methanol using a polyamine and a homogeneous ruthenium catalyst. J. Am. Chem. Soc. 138, 778–781 (2016).

D.W., R.S., P.S., H.J. and M.B. acknowledge financial support from the State of Mecklenburg-Vorpommern and European Union (EFRE; project ‘h2cycle’). D.W., H.J. and M.B. acknowledge the Danish government for funding the CADIAC excellence cluster and the Leibniz-Program Cooperative Excellence K308/2020 (project ‘SUPREME’). D.W. thanks X. Liu for providing some starting materials. We thank the analytical team of LIKAT for their kind support.

Leibniz-Institut für Katalyse e.V., Rostock, Germany

Duo Wei, Rui Sang, Henrik Junge & Matthias Beller

APEX Energy Teterow GmbH, Rostock-Laage, Germany

You can also search for this author in PubMed  Google Scholar

You can also search for this author in PubMed  Google Scholar

You can also search for this author in PubMed  Google Scholar

You can also search for this author in PubMed  Google Scholar

You can also search for this author in PubMed  Google Scholar

Investigation (performing the experiments, data collection), formal analysis and writing (original draft), D.W. and R.S.; conceptualization, funding acquisition, methodology, supervision, writing (review and editing), P.S., H.J. and M.B. All authors have read and agreed to the published version of this work. D.W. and R.S. contributed equally to the work.

Correspondence to Peter Sponholz, Henrik Junge or Matthias Beller.

The methods disclosed in this work have been filed as Patent application PCT/EP2022/052967 with APEX Energy Teterow GmbH as the applicant and D.W., R.S., H.J., M.B. and P.S. (employed by APEX Energy Teterow GmbH) as inventors. The status of the patent application is pending.

Nature Energy thanks Sheri Lense, Rhett Kempe and Wan-Hui Wang for their contribution to the peer review of this work.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

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 license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license 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 license, visit http://creativecommons.org/licenses/by/4.0/.

Wei, D., Sang, R., Sponholz, P. et al. Reversible hydrogenation of carbon dioxide to formic acid using a Mn-pincer complex in the presence of lysine. Nat Energy 7, 438–447 (2022). https://doi.org/10.1038/s41560-022-01019-4

DOI: https://doi.org/10.1038/s41560-022-01019-4

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Nature Energy (Nat Energy) ISSN 2058-7546 (online)

CAS 96-29-7 2-Butanone oxime Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.