Manganese(I) Catalyzed CO2 Reduction Processes in Homogenous Phase

The catalytic activation of carbon dioxide (CO2) can be a promising tool for the use of this abundant, non‐flammable and non‐toxic gas as feedstock for C1 chemical synthesis, in particular for bulk products such as formic acid (HCOOH) and methanol (CH3OH). The key for successful, widespread use of CO2 is the design and application of efficient, thermally robust and cheap catalysts. In the last decade, the use of earth‐abundant transition metal complexes, in particular those of 3d metals, has shown promising results. In this review article, a comprehensive summary of the main catalytic systems described in the literature for homogeneous Mn(I)‐catalyzed CO2 reduction processes (hydrogenation, hydroboration and hydrosilylation) will be described, with particular attention to the effect of stabilizing ligands, reaction conditions and need for additives, together with mechanistic details often obtained by a combination of experimental data and DFT calculations.


Introduction
In the quest for more sustainable and environmentally friendly approaches to chemical synthesis, scientists are facing various issues, among which the choice of abundant substrates, mild conditions, high process efficiency and selectivity to the desired products, use of green solvents etc.Many industrial processes, especially those applied for the synthesis of bulk chemicals, still use harsh reaction conditions (high pressures and temperatures) and/or toxic gases such as carbon monoxide (CO), for example in the synthesis of methanol (CH 3 OH) or formic acid (HCOOH) from methyl formate.Another environmental issue is the need to curb CO 2 emissions, which have overcome the alert threshold of 400 ppm concentration in the atmosphere, to help in keeping global warming under control.Many research groups have demonstrated over the years the possibility to use CO 2 as abundant, cheap, non-flammable and non-toxic feedstock for C1 chemical synthesis. [1]O 2 utilization processes can be principally divided in i) CO 2 addition, either into unsaturated carbon-carbon bonds, such as C=C double bonds (alkenes) and C�C triple bonds (alkynes), or in highly strained organic molecules containing heteroatoms (epoxides, aziridines), in which case the carbon oxidation state does not change, and ii) CO 2 reduction, ranging from 2 to 8 electrons, leading to HCOOH, HCHO, CH 3 OH, CH 4 , respectively.These products find applications in the chemical industry both as bulk chemicals and in the synthesis of commodities, and are so far generally obtained from fossil-based feedstocks.Therefore, practical ways to obtain them from CO 2 are highly desired.In order to overcome the thermodynamic stability of CO 2 and promote C=O bond activation for further transformations under feasible reaction conditions, the use of a catalyst is needed.Thus, homogeneous, heterogeneous, photo-and electrocatalysts have been designed, extensively studied and applied for CO 2 utilization processes.[2] Homogeneous catalysts offer some important advantages, namely easier synthesis by reaction between suitable metal precursors and stabilizing ligands, easier tuning of the steric and electronic properties by simple ligand design, and in general the possibility to achieve higher product selectivity under milder reaction conditions.Traditionally, homogeneously catalyzed CO 2 reduction processes have been the realm of precious metal catalysts, in particular Ru and Ir.With the use of appropriate stabilizing ligands, mainly based on phosphorus and/or nitrogen, extraordinary turnover numbers (TONs) and turnover frequencies (TOFs) have been obtained, for example in CO 2 hydrogenation to HCOOH or formate.[3] The major drawbacks of platinum group metals are however their high costs, and the low abundance in the Earth's crust (Ru: 1×10 À 4 ppm, Ir: 5×10 À 5 ppm), [4] thus for these reasons earth-abundant metals are being studied as possible replacements.In the last decade, complexes of 3d metals such as Fe, [5] Co, [6] and Ni [7] have been studied for CO 2 hydrogenation, and only since 2016 manganese (Mn), the third most abundant metal in the Earth's crust, has received attention for this kind of process.In the present review article, the discussion will be focused on the state-of-the-art for homogeneous Mn(I)-catalyzed CO 2 reduction processes.Thus, heterogeneous, photo-and electrocatalytic approaches will not be included.The literature data will be grouped based on the type of reaction.CO 2 hydrogenation was reported either as direct (Scheme 1a), i. e. the conversion of CO 2 /H 2 gas mixtures to HCOOH, formates or CH 3 OH, and indirect hydrogenation to CH 3 OH (Scheme 1b), that involves the use of amines or alcohols as shuttles to obtain formamides or esters, and the following hydrogenation of these intermediates to the desired product.Next, CO 2 hydroboration processes, leading to boryl-protected HCOOH, HCHO and CH 3 OH will be highlighted, showing how ligand effects and tuning of the reaction conditions can affect the process selectivity (Scheme 1c).Then, in the following section the few existing reports on Mn-catalyzed CO 2 hydrosilylation processes will be summarized, focusing on the proposed reaction mechanisms leading to silylformates, bis(silyl)acetals and methoxysilanes (Scheme 1d).

CO 2 hydrogenation
CO 2 hydrogenation can be formally described as the addition of hydrogen (H 2 ) to CO 2 .Depending on the reaction conditions, type of catalyst, use of base or additives such as Lewis acids, and in particular on the partial pressure ratios between CO 2 and H 2 , different products can be obtained.In the case of the 2electrons CO 2 hydrogenation to formic acid (HCOOH) or formate salts (HCOO À ), generally run under a 1 : 1 CO 2 /H 2 gas ratio (Scheme 1a, left), the reaction is endoergonic in gas phase (ΔG°= 32.8 kJ mol À 1 ), mainly due to the large negative entropic factor (ΔH°= À 31.5 kJ mol À 1 , ΔS°= À 216 J mol À 1 ).In the presence of an added base, for example NH 3 , it can be turned to exoergonic (ΔG°= À 9.5 kJ mol À 1 , ΔH°= À 84.3 kJ mol À 1 , ΔS°= À 250 J mol À 1 ), and the formation of the corresponding formate salt constitutes the thermodynamic driving force. [8]In the case of 6-electrons CO 2 hydrogenation to methanol (CH 3 OH), the required gas ratio is formally CO 2 /H 2 = 1 : 3 (Scheme 1a, right), although in practical tests a higher H 2 partial pressure is often used.The gas-phase reaction is exoergonic (ΔG°= À 9.5 kJ mol À 1 , ΔH°= À 131 kJ mol À 1 , ΔS°= À 409 J mol À 1 ), thus addition of base is not required to drive the equilibrium to the desired product. [9]side the "direct" hydrogenation of CO 2 to CH 3 OH (Scheme 1a, right), another approach reported in the literature leading to CH 3 OH from CO 2 involves two steps, i. e. the initial reduction of CO 2 to HCOOH, that in the presence of an amine or an alcohol is transformed in situ into a formamide or formate ester, followed by the hydrogenation of these intermediate products to CH 3 OH with regeneration of the amine/alcohol, that are thus used as formyl moiety shuttles (Scheme 1b).This approach is generally indicated as "indirect" CO 2 hydrogenation.
For the purpose of this review article, the literature on Mncatalyzed CO 2 hydrogenation will therefore be discussed separately in "direct" and "indirect" CO 2 hydrogenations.Table 1 summarizes the main reaction conditions and catalytic results for all the described complexes.Drawings of the molecular structures of the Mn complexes related to the chapter are shown in Figure 1.Noteworthy, pincer-type complexes are predominant in CO 2 reduction processes.Pincer ligands can impart the ideal geometry, electronic and steric properties to the metal center, and can be easily tuned by synthetic variations of donor atoms and their substituents.

Direct hydrogenation
3b,10] Recent results show that very high productivities (TON = 4650000,TOF = 64593 h À 1 ) can be achieved also with base metals such as Ni in the presence of a stabilizing NP 3 tetradentate ligand. [7]he first report on Mn(I)-catalyzed CO 2 hydrogenation appeared in the literature in 2016 as a predictive DFT computational study (DFT = density functional theory). [11]The authors studied the energy profile of CO 2 hydrogenation starting from [MnBr(bipy)(CO) 3 ] (bipy = bipyridine), previously reported as an active electrocatalyst for the electrocatalytic reduction of CO 2 in the presence of weak Brønsted acids.In detail, the focus of the   investigation was to understand the role of the ratio of σ-donor (PMe 3 ) and of π-acceptor (CO) ligands in substituted complexes [MnX(bipy)(L 1 )(L 2 )(CO)] (L 1 = L 2 = CO; L 1 = CO, L 2 = PMe 3 ; L 1 = L 2 = PMe 3 ; X=CH 3 CN) on the rate-limiting step in the catalytic cycle, in the presence of NH(CH 3 ) 2 as base.It was shown that the key steps of the mechanism are heterolytic H 2 cleavage followed by hydride transfer, and that the latter is rate-determining for all the three complexes.It was further demonstrated that this step can be significantly improved by using a proper combination of σ-donor and π-acceptor ligands by subtle ligand tuning, for example replacing PMe 3 with PF 3 .In 2017, the first experimental study on Mn(I)-catalyzed efficient CO 2 hydrogenation to formate was disclosed by Gonsalvi, Kirchner and coworkers. [12]Two well-defined Mn(I) hydridocarbonyl complexes where synthesized from the reaction of [MnBr(CO) 5 ] with 2,6bis(diaminodiisopropylphosphine)pyridyl pincer type ligands (PNP) and then treated with LiHBEt 3 .The so-obtained complexes [MnH(PNP H -iPr)(CO) 2 ] (1 a) and [MnH(PNP Me -iPr)(CO) 2 ] (1 b) are structurally similar and differ only by the substituent on the amine N atom in the backbone (H and Me, respectively).They were both tested as catalysts in the presence of DBU (DBU = 1,8-diazabicycloundec-7-ene) as added base, screening the main reaction parameters such as solvent, temperature, pressure, catalyst concentration and need for Lewis acid cocatalysts, showing overall a higher performance of complex 1 a.The initial test was run at 80 °C in THF/H 2 O, 80 bar H 2 /CO 2 = 1 : 1, 24 h, using a 1 a/DBU ratio of 1 : 1000, obtaining (DBUH)(HCO 2 ) in quantitative yields with TON = 1000.At lower catalyst loadings (1 a/DBU = 1 : 10000; [1 a] = 0.18 μmol/mL), the yield in formate decreased to 55 %, corresponding to TON = 5520.At a catalyst concentration of 0.036 μmol/mL (1 a/DBU = 1 : 50000), TON increased significantly (9100), but lower yields were obtained (16 %).Further optimization showed that by extending the reaction time to 48 h and using a 1 a/DBU ratio of 1 : 10000, high yields of formate (86 %) and high TON (8600) could be reached.The catalyst performance could be dramatically increased by addition of Li triflate as Lewis acid co-catalysts to the reaction mixture.Addition of LiOTf (0.5 mmol; 1 a/DBU = 1 : 10000; 1 a/LiOTf = 1 : 1000) gave quantitative yields of formate after 24 h of reaction with TON = 10000.The system was further optimized raising the temperature to 100 °C.After 24 h, with 1 a/DBU = 1 : 50000, 1 a/LiOTf = 1 : 5000 ratios, a TON of 31600 (63 % yield) was finally obtained (Table 1, entry 1).
The reaction mechanism involving 1 a was then elucidated using NMR spectroscopy and DFT calculations, that showed that two interconnected mechanisms can equally contribute to the overall catalytic reaction.In the first one (Scheme 2, left), the reaction pathway involves a purely metal-centred mechanism without participation of the NÀ H bond of the PNP ligand.This path starts with nucleophilic attack of the hydride present in 1 a to CO 2 in an outer sphere way, to give the intermediate complex [Mn{k 1 -H-(OCHO)}(PNP H -iPr)(CO) 2 ] (1 c) with a barrier of 9.6 kcal mol À 1 .From 1 c, the pathway may proceed either by formato ligand isomerisation to give [Mn{k 1 -O-(OCHO)}(PNP H -iPr)(CO) 2 ] (1 d) via a transition state with energy of 13.2 kcal mol À 1 , or by exchange of the formate in 1 c with a H 2 molecule to give [Mn(η 2 -H 2 )(PNP H -iPr)(CO) 2 ] (1 e), that is only 7.0 kcal mol À 1 higher in energy than 1 a + CO 2 .This step corresponds to the highest energy transition state of the pathway (ΔG ¼ 6 = 17.9 kcal mol À 1 ).Next, η 2 -H 2 ligand deprotonation by either free formate or DBU base gives back 1 a and the desired product (DBUH)(HCO 2 ).The overall pathway has a maximum barrier of 27.6 kcal mol À 1 , and is thermodynamically favorable with a free energy ΔG = À 11.2 kcal mol À 1 .
In the second pathway (Scheme 2, right), a bifunctional mechanism with participation of the NÀ H bond of the PNP ligand is active.The first step is in common with the previous one, leading to 1 c.Next, DBU deprotonates the acidic NÀ H group of the PNP ligand, releasing the formate ligand as (DBUH)(HCO 2 ).The five-coordinate Mn neutral complex [Mn- (1 g), 4.2 kcal mol À 1 more stable than 1 a + CO 2 in the calculations.The final step involves solvent-assisted HÀ H bond deprotonation from 1 f and ligand protonation, yielding back the initial 1 a.This step has the highest energy barrier of this pathway, at 20.0 6 kcal mol À 1 .The overall barrier for the entire second pathway was calculated as 29.7 kcal mol À 1 , ca. 2 kcal mol À 1 higher than that of the first pathway, i. e. in reach under experimental conditions.Indeed, this observation is in line with the results of the catalytic runs, showing that 1 a (bearing a NH bond in the ligand scaffold) is more active than 1 b, that has instead a NÀ Me moiety and it is thus unable to activate a metal-ligand cooperation (MLC) mechanism.
Another interesting observation from the DFT calculations is that in the two pathways the base (DBU) has different roles.In the first case, DBU can either react with HCOOH produced from CO 2 hydrogenation yielding (DBUH)(HCO 2 ) in an acid-base process, or directly deprotonate the η 2 -H 2 coordinated ligand.In the second mechanism, DBU deprotonates the NÀ H bond in the PNP ligand and forms (DBUH)(HCO 2 ) with the free formate anion.Finally, it was established that the LiOTf Lewis acid cocatalyst has the effect to stabilize the k 1 -H-coordinated formate over the k 1 -O-coordinated formate, that represents a thermodynamic sink common to both mechanisms, by decreasing the free energy difference between the two isomes by ca.3.4 kcal mol À 1 .
Almost at the same time of the report described above, Nervi, Khusnutdinova and coworkers disclosed the use of P-free, Mn(I)-complexes for CO 2 hydrogenation to formate. [13]The authors elaborated on the concept of use of Mn bipy-type proposed earlier by Pathak and coworkers, [11] synthesizing and testing a series of Mn(I) tricarbonyl complexes [MnBr(R 1 R 2bipy)(CO) 3 ] (2 a-e) and [Mn(MeCN)(R 1 R 2 -bipy)(CO) 3 ] OTf (3 a-c), bearing differently 4,4'-and 6,6'-substituted bipy-type ligands (R 1 or R 2 = H, OH, OMe, NH 2 ) shown in Figure 1.In particular, the choice of o-OH bis-substituted bipyridine as a ligand was inspired by natural enzymes, i. e. the ortho-hydroxypyridine structural fragment present in [Fe]-hydrogenase.The main purpose of the synthetic modifications was to test whether the presence of proton-shuttling groups in proximity of the metal center could help one of the crucial steps in the catalytic cycle, i. e. heterolytic hydrogen activation.The best results were obtained with the o-OH substituted complexes 2 a and 3 a, whereas only traces or moderate yields of formate were observed with the unsubstituted or p-substituted bipy-type complexes.Solvent screening showed that in 1,4-dioxane, with DBU as a base, 60 bar of H 2 /CO 2 (1 : 1), 80 °C, 24 h, 97 % yield of formate salt was obtained with 2 a (TON = 1224).Using CH 3 CN, similar results were obtained at 65 °C using 5 μmol of 2 a (TON = 1313, 96 % yield) and 3 a (TON = 1299, yield = 99 %).The catalytic system involving 2 a was then optimized, by decreasing the catalyst amount to 1 μmol and running the tests at 65 °C in CH 3 CN, reaching TON = 6250, 98 % yield.As another example of CO 2 utilization process (amine formylation), in the same study 2 a was tested as catalyst in the presence of diethylamine and 70 bar of CO 2 /H 2 (2 : 5) in 1,4-dioxane, and diethylformate was obtained with a TON of ca.588 after 24 h at 80 °C.
The catalytic mechanism of CO 2 hydrogenation with this class of complexes was then studied by DFT calculations in 2018 by Das and Pati. [14]In keeping with the earlier work by Pathak and coworkers, [11] the authors confirmed that also for substituted bipy-Mn complexes the key steps are base-assisted heterolytic dihydrogen cleavage and hydride transfer to CO 2 , and that both OH groups are deprotonated in the presence of DBU, as experimentally demonstrated. [13]Interestingly, it was observed that the role of the pendant oxygen atoms in ortho position is not, as it was reasonable to expect, to promote dihydrogen activation, but rather to help in reaching the correct geometry making the final hydride transfer step more feasible.
In detail, it was observed that the pendant oxygen and the H 2 molecule occupy equatorial and axial sites in the active species, in turn causing geometric distortions during the watermediated, base-free H 2 activation step.This distortion is strongly reduced by isomerization, causing these two ligands to lie in proximity on the same equatorial plane, thus giving the correct geometry for energetically more accessible H 2 activation.A simplified mechanism is shown in Scheme 3.
The use of Mn-PNN pincer-type complexes was tested by Milstein and coworkers in 2019. [15]The pincer ligands were designed in such a way to allow for metal-ligand cooperation (MLC) in substrate activation.Complexes [Mn(PNN tBu )(CO) 2 ] (4 a) and [Mn(PNN NEt2 )(CO) 2 ] (4 b), obtained by treatment of the parent bromide compounds with K t OBu, differ by the number and type of substituents on the amino group (one t Bu, deprotonated secondary amine vs. two Et groups, tertiary amine), respectively (Figure 1).Additionally, the pyridine ring in  In 2021, Saouma and coworkers tested the reactivity of the PNP pincer complex [MnH(PNP tBu )(CO) 2 ] (5 a, PNP tBu = 2,6-bis(ditert-butylphosphinomethyl)pyridine) for CO 2 hydrogenation. [16]s in the case of Milstein's PNN systems described above, also for complex 5 a an unexpected addition of CO 2 to the ligand backbone was observed.Detailed mechanistic and thermochemical studies showed that the MnÀ H bond in 5 a is not sufficiently hydridic to reduce CO 2 .Deprotonation of 5 a generates as expected complex [MnH(PNP À )(CO) 2 ] (5 b, PNP À = monodeprotonated 2,6-bis(di-tert-butylphosphinomethyl) pyridine).For 5 b, the MÀ H bond hydricity was calculated to be enhanced by ca. 17 kcal mol À 1 compared to 5 a, but in spite of the sufficient hydricity to promote CO 2 hydrogenation, also in this case this reactivity was not observed.Instead, CO 2 preferentially binds to the ligand backbone in a [1,3]-fashion to generate the anionic hydride complex [MnH(PNP-CO 2 )(CO) 2 ] (6), that was isolated by reaction of 5 b with CO 2 (0.85 bar).Complexes 5 a and 6 (5 μmol) were tested as catalyst for CO 2 hydrogenation using KO t Bu (100 equiv.)as base, THF, 18 h.After a screening of CO 2 /H 2 pressure and temperature, best TON = 58 and TON = 49 were obtained for 5 a and 6, respectively, using a pressure of 30 bar (1 : 1 gas ratio) and 100 °C.A detailed study of stoichiometric reactivity and thermodynamic stability led to the proposal of the catalytic mechanism shown in Scheme 5, with two interconnected cycles centered on complex 5 b, leading to the formation of the active species that promote HCO 2 K release.
In applied catalysis, it is often desirable to use simple, bench stable, commercially available ligands, to avoid the need for costly and time-consuming prior synthetic procedures.For this purpose, in 2021 Gonsalvi, Kirchner and coworkers studied the catalytic properties of a series of textbook Mn(I) alkylcarbonyl complexes in CO 2 hydrogenation to formate (Scheme 6). [17]n previous studies, it was observed that complex [Mn-( n Pr)(dippe)(CO) 3 ] (7, dippe = 1,2-bis(diisopropylphosphino) ethane) can act as long shelf-lived precursor for the catalytically active 14-electrons species [MnH(dippe)(CO) 2 ] (7 a), that is formed in situ from 7 under a pressure of H 2 and elimination of butanal.Complex 7 was tested under mild reaction conditions (DBU, 40-60 bar pressure, 80 °C) for the hydrogenation of CO 2 using 1 : 1 to 1 : 3 CO 2 /H 2 ratios.In the absence of additives, it was possible to obtain (DBUH)(HCO 2 ) with TON = 1077 after 24 h under 60 bar CO 2 /H 2 = 1 : 1, albeit with low yields (21 %).The catalytic system was substantially improved by addition of LiOTf as Lewis acid co-catalyst and by changing the gas ratio to CO 2 /H 2 = 1 : 2 (total pressure = 75 bar).In this case a maximum TON = 1988 (yield > 98 %) was obtained after 24 h, 80 °C.The mechanism was then studied by DFT calculations, showing that in the initial step, in the presence of CO Reproduced from Ref. [15] with permission from the Royal Society of Chemistry.
The importance of metal-ligand cooperation in Mn-catalyzed CO 2 hydrogenation has been studied by many authors, and in particular the beneficial effect of the presence of a free N atom in the ligand backbone, able to participate in substrate activation, has been evaluated and described both by experimental and computational means.In 2017, a predictive DFT calculations study by Yang and coworkers [18] examined the potential of a series of manganese and cobalt complexes of the type CH 3 , NH 2 , OH, CHO, COOH, COCH 3 , and COOCH 3 ) as catalysts for the direct hydrogenation of CO 2 to methanol.These complexes, extensively studied by Bullock, DuBois and coworkers in particular for dihydrogen activation, [19] feature a multidentate PNNP ligand with a pendant amine arm, having the correct basicity to facilitate dihydrogen heterolytic splitting by accepting a proton from the metal-coordinated η 2 -H 2 ligand.In the computational study, the calculations showed that the overall process proceeds via three cascade catalytic reactions, namely i) the hydrogenation of CO 2 to HCOOH, ii) the hydrogenation of HCOOH to formaldehyde (HCHO) with the formation of water, and iii) the hydrogenation of HCHO to methanol.
In the case of complex [Mn(k 1 -O-OCHO)( tBu P 2 tBu N 2 )(CN) 2 ] (Scheme 7), a rather accessible overall energy barrier (27.7 kcal mol À 1 ) was calculated.The key difference for the Mn-based system compared to the Co-based analogue is that instead of the step-by-step release of CH 2 O and transfer of the proton on N to the hydroxyl for the formation of H 2 O for Co, in the case of Mn the release of CH 2 O and transfer of the proton on N1 to the hydroxyl group to give H 2 O is a simultaneous process from a single transition state.In turn, this implies that the hydrogenation of HCHO to CH 3 OH is fast and with a low energy barrier (12.7 kcal mol À 1 ).
Among PNP pincer-type ligands, the bis-(phospino)amine [HN(C 2 H 4 P i Pr 2 ) 2 ], commonly denoted i Pr-MACHO, has often been used by many research groups worldwide to obtain the corresponding precious and earth-abundant metal complexes.This simple aminophospine ligand often brings about metalligand cooperation by reversible protonation/deprotonation at the nitrogen atom, in such a way that amino/amido species are involved in the catalytic processes.In 2021, Leitner and coworkers applied the amido complexes [Mn{N(C 2 H 4 PR 2 ) 2 }(CO) 2 ] (R = i Pr, 8 a; Ph, 8 b; t Bu = 8 c) as catalysts for the direct hydrogenation of CO 2 to MeOH, in the presence of Lewis acids such as Sc(OTf) 3 and Ti(O i Pr) 4 and a mixture 1,4-dioxane/EtOH as solvent. [20]ontrol experiments showed that the role of the Lewis acid is to act as an esterification promoter to remove the formate ligand from the Mn coordination sphere of the Mn-formato complexes, that otherwise represent the resting state of the catalysts.Under optimized conditions, i. e. H 2 /CO 2 ratio > 30, 150 °C, 68 h, Ti(O i Pr) 4 , a TON max = 87 was obtained in the presence of 8 b (Table 1).Mechanistic details were obtained in 2022 by the same authors, by a combination of experimental and computational data. [21]Quantum chemical computations at the density functional theory and DLPNOCCSD(T) level of theory confirm the previous hypothesis of a formate resting state as the most stable intermediate.Using CH 3 OH as solvent in the calculations, in the presence of 8 a the reaction sequence considered involves i) CO 2 hydrogenation to HCOOH, ii) methanol esterification to methyl formate, iii) methyl formate hydrogenation, iv) hemiacetal decomposition to give free HCHO and v) HCHO hydrogenation to methanol (Scheme 8).
The initial MnÀ H active species is obtained by a Dub-type mechanism involving proton shuttle-assisted activation of dihydrogen, centered on the activated Mn complex, by the alcohol solvent.Next, CO 2 inserts into the MnÀ H bond to give the Mn-formato species by hydride transfer without the need for pre-coordination.The Lewis acid then promotes the esterification of HCOOH to formate ester.Next, hydride transfer to the formate ester gives a hemiacetal anion that decomposes to HCHO by cleavage of its CÀ O bond.The following hydrogenation of formaldehyde has a very low barrier and proceeds by a concerted hydride and proton transfer to the carbonyl carbon.The rate-determining states (RDSs) of the reaction belong to the hydrogenation of a formate ester to methanol step, as determined by DFT calculations and experimental concentrationÀ time profiles.The rate-determining step was finally confirmed to be hydrogen splitting, by measuring kinetic isotope effects (KIE) using H 2 /D 2 and EtOH/EtOD mixtures.
It is interesting to observe that, at a predictive DFT calculation level, even small changes in the ligand architecture can have a profound influence on the overall reaction mechanism.In 2021 Lei and coworkers used a density functional theory (DFT) calculations approach to explore the reaction mechanism of the three cascade cycles for the hydro-genation of carbon dioxide to methanol catalyzed by the manganese PNP pincer complex [Mn{(Ph 2 PCH 2 SiMe 2 ) 2 N}(CO) 2 ], that differs only slightly from the MACHO-type ligands, by the presence of SiMe 2 moieties in place of a methylene group of the backbone. [22]This complex was previously experimentally used as efficient catalyst for hydroboration reactions, including the conversion of CO 2 to methoxyborane, by Leitner and coworkers in 2018 (see Section 3). [23]In the case of CO 2 hydrogenation to methanol, the study showed a remarkable solvent effect changing the RDS.Under solvent-free conditions, the calculations showed that hydrogen activation is the ratedetermining step of each of the three catalytic cycles, namely CO 2 hydrogenation to HCOOH, formic acid reduction to HCHO and hydrogenation of formaldehyde to methanol, with an overall highest barrier of 27.1 kcal mol À 1 .By introducing different solvents such as water, toluene and THF in the calculations, the RDS of this reaction corresponds to the hydrogen transfer step of the formic acid hydrogenation stage, with energy spans of 21.3, 20.8 and 20.4 kcal mol À 1 for the three solvents, respectively.The initial MnÀ H species is generated as expected by MLC-type heterolytic dihydrogen splitting accompanied by ligand protonation at the N atom.The NÀ H moiety then also assists in the outer-sphere hydrogenation of formic acid to methanediol and the decomposition of methanediol to formaldehyde (Figure 2).  10 c). [24]The catalytic protocol, based on the use of aminoacids such as lysine (Lys) and the corresponding potassium lysinate (LysK) salt, represents a sustainable hydrogen storage and release method, bringing together reversible hydrogenation of CO 2 to HCOOH and carbon capture processes.Under standard reaction conditions, i. e. Lys (5.0 mmol), Mn catalyst (0.02 or 0.10 μmol), H 2 O/THF (1 : 1, 10 mL total), CO 2 /H 2 (1 : 3, 80 bar total), 85-145 °C, 12 h, the best results were obtained with 9 a (TON = 40000, 145 °C) and 10 b (TON = 230000), the latter representing the current state-of-theart TON for Mn-catalyzed CO 2 hydrogenation to formate to date.Complex 10 b also resulted to be a very efficient catalyst for H 2 evolution from formic acid (5.0 mmol) in the presence of Lys (1.0 equivalent) at 90 °C, 12 h, reaching TON = 29400 and > 99 % HCOOH conversion.By recharging the reaction vessel, up to 10 consecutive runs were possible for CO 2 hydrogenation, leading to a total TON (TTON) of ca.2000000.To avoid the drawback of CO 2 recharging in consecutive runs, and with the aim of providing a practical system for a chemical hydrogen battery, the use of LysK in place of Lys was tested, as this salt is known for its property to capture CO 2 .After one cycle of hydrogen release and storage, formate (92 %) was obtained Scheme 8. Stepwise reaction pathway for CO 2 hydrogenation to CH 3 OH in the presence of 8 a. Reproduced (adapted) from Ref. [21].
applying 5 mmol LysK.In a scaled-up prototype, the H 2 evolution process applying LysK was carried out without drop of the efficiency in at least ten charge-discharge cycles, without the need to feed CO 2 to the reactor after the first run.Mechanistic details were obtained by NMR experiments, that allowed the authors to propose a catalytic cycle for reversible CO 2 hydrogenation in the presence of pre-catalyst 10 b (Scheme 9).
The first step involves pre-catalyst activation by LysK via NÀ H deprotonation and dearomatisation of the triazine moiety, giving a bimetallic MnÀ K active species.Then, in the CO 2 hydrogenation pathway (H 2 storage), dihydrogen is activated via heterolytic cleavage to give the corresponding Mn-hydrido complex, with protonation of LysK to Lys.CO 2 insertion into the MnÀ H bond follows to give the Mn-formato species.Finally, HCOOH is liberated by converting 1 equivalent of Lys to LysK, regenerating the active Mn species after dearomatisation of the triazine moiety.
Other reaction pathways and mechanisms involving Mn-MACHO pincer complexes were evaluated.Base-free CO 2 hydrogenation to formate in the presence of 8 a was studied by DFT calculations by Rawat and Pathak in 2017, focusing in particular on the role of substituents on the P donor atoms. [25]The calculated catalytic cycle (Scheme 10) starts from H 2 coordination to 8 a to give a Mn(η 2 -H 2 ) complex bearing deprotonated PNP ligand.Then, by MLC-type heterolytic H 2 cleavage, a MnÀ H species with protonated PNP is formed.The hydrido complex then undergoes CO 2 insertion into the MnÀ H bond to give a k 1 -H-formato complex.HCOOH is formed by proton transfer from the ligand N atom to the O À atom of thecarboxylic group, and it is then released to regenerate 8 a.The calculations showed that heterolytic H 2 cleavage (RDS) is favored by strong π-acceptor Scheme 10.DFT calculated catalytic mechanism for the base-free hydrogenation of CO 2 to HCOOH in the presence of 8 a. Reproduced from Ref. [25]  with permission from the Royal Society of Chemistry.charged with the catalyst, solvent, amine, base, CO 2 /H 2 gas pressure and then heated to the desired temperature for the chosen time.After this time, analysis of aliquots of the reaction mixture allowed to quantify the yields in formamides.The experiments were repeated, and after the end of the formylation step the gas inside the reactor was vented, then fresh H 2 (pressure) and base were added to the reactor containing the reaction mixture and the catalyst.After heating for a set time, the crude mixtures were analyzed to evaluate the amount of methanol formed during the formamide hydrogenation step.K 3 PO 4 and t BuOK were used as bases, whereas the best amines, after a due screening, were found to be morpholine, benzylamine and N-methylbenzylamine.For the first step, it was shown that pre-catalyst 9 a (0.5 mol %) was able to promote the formation of 4-formylmorpholine in 77 % yield, using K 3 PO 4 (2.5 mol %) CO 2 /H 2 (1 : 1, 60 bar total pressure), 110 °C in THF after 24 h.Upon optimization, a maximum yield = 95 % was obtained, using 2.0 mol % of catalyst and t BuOK (10 mol %), 36 h.Using pure 4-formylmorpholine as substrate, this was effectively reduced to morpholine and CH 3 OH (64 % yield, TON = 128) at 0.5 mol % catalyst loading in THF after 24 h under 70 bar of H 2 pressure and 150 °C in the presence of 9 a.On the other hand, in the sequential formylation-hydrogenation experiment, adding an additional 80 bar of H 2 to the mixture used for morpholine formylation under optimized conditions, running the test at 150 °C, 36 h, CH 3 OH was obtained in high yield (71 %) corresponding to TON = 36.Attempted one-pot reduction of CO 2 to CH 3 OH using 9 a, 5 bar CO 2 and 85 bar H 2 , 150 °C, 24 h in the presence of morpholine was however unsuccessful, with only 5 % yield in 4-formylmorpholine.The yield in formamide was increased to 38 % by a different experimental setup consisting of stepwise temperature increase, but also in this case only traces of CH 3 OH were observed.The reaction mechanism was studied by NMR spectroscopy, and the experimental evidences so obtained suggest a pivotal role of the Mn-formato complex 11 c for both reactions (Scheme 13).
The reaction mechanism was further studied by DFT calculations and microkinetics in 2019 by Pathak and coworkers. [29]The amidation step, involving a metal-free organic reaction between HCOOH and morpholine, was calculated to be highly exergonic (À 10.4 kcal mol À 1 ), and this helps to increase the overall rate of the reaction.It was then observed that the following N-formylmorpholine hydrogenation step may follow two competitive pathways with comparable free energy barriers.As shown in Scheme 14, path a) starts from the activation of N-formylmorpholine by 11 a, bringing about C=O bond hydrogenation, followed by C=N bond hydrogenation to give morpholine and CH 3 OH with the regeneration of 11 a.The highest energy barrier for path a) was calculated as 38.2 kcal mol À 1 .In competitive path b) instead, C=N hydrogenation occurs first, releasing morpholine, then C=O bond hydrogenation releases CH 3 OH regenerating 11 a.For path b), the highest energy barrier was calculated as 38.4 kcal mol À 1 , thus both pathways can equally contribute to the overall catalytic mechanism.
In 2023, Al Thabaiti, Lahiri, Maiti and coworkers reported on the use of a Mn-SNS pincer complex for the indirect CO 2 hydrogenation to CH 3 OH. [30]The authors applied the benchstable complex [Mn(SNS)(CO) 3 ]Br [13 a, SNS = bis(2-(ethylthio)ethyl)amine)] as pre-catalyst for the reaction in the presence of amine shuttles, including morpholine, piperidine, benzylamine, 3,4-dimethylbenzyl amine, 3-methoxybenzyl amine.The tests were run using THF as solvent, CO 2 /H 2 (1 : 1, 20 bar total pressure), 1 mol % of catalyst, K 3 PO 4 as base (1 to 2 equiv.),120 °C, 24 h.Under these mild conditions, formamides were obtained in high yields, i. e. up to 91 % for N-formylmorpholine, using 2 equiv. of base, however no traces of CH 3 OH were observed.By changing the CO 2 /H 2 ratio to 1 : 6 (70 bar total pressure) and heating the reaction mixture to 110-140 °C for 24 h, methanol formation was detected by GC analysis.The authors calculated a TON = 60 for methanol formation, however this value was only estimated by NMR spectroscopy, from the decrease of the signal corresponding to the -CHO group of Nformylmorpholine, and not quantified against an internal standard.The proposed catalytic mechanism (Scheme 15) involves two steps, i. e. amine formylation by in situ generated HCOOH (step i) and formamide hydrogenation to CH 3 OH (step ii), both centered on the putative hydrido complex [MnH(SNS)(CO) 2 ] (13 b).

CO 2 hydroboration
Among possible alternative reductants to replace hydrogen, hydroboranes (R 2 BH) have the advantage of being easier to store and handle, due to their liquid state at room tempearture and pressure.Already under mild reaction conditions, they can bring about CO 2 reduction to various products (Scheme 1c) such as formoxyborane (A), bis(boryl)acetal (B), methoxyborane (C) and bis(boryl)ether (D), due to the much weaker nature of the BÀ H than HÀ H bond, making the formation of new BÀ O bonds much easier.The state of the art for CO 2 hydroboration to formoxyborane (A) was obtained by Hazari and coworkers.The use of [PdH( Cy PSiP)] (0.001 mol %), pinacolborane (HBpin, 0.050 mmol) and CO 2 (1 bar) in C 6 D 6 leads to a very high catalytic performance with a TON of 63500 (TOF = 529 h À 1 ) obtained after 5 days of reaction at 25 °C. [31]In the case of methoxyboranes (C), the best results were obtained by Chen and coworkers using bis(phosphinite) POCOP-type Ni(II) pincer complexes of general formula [NiX{2,6-(R 2 PO) 2 C 6 H 3 }] (R = t Bu, X=SC 6 H 4 -p-OCH 3 ; R = i Pr; X=N 3 ).The catalytic hydroboration of CO 2 was carried out in the presence of catecholborane (HBcat) with 0.2 mol % of catalyst in C 6 D 6 at room temperature to obtain TONs of 490 (TOF = 2400 h À 1 ) and 477 (TOF = 1908 h À 1 ), respectively. [32]n the last decade, part of the chemistry community has focused on the quest for cheaper catalysts based on earthabundant metals to increase process sustainability for CO 2 hydroboration. [33]The results of Mn-catalyzed CO 2 hydroboration are hereby summarized.Drawings of the molecular structures of the competent Mn catalysts are shown in Figure 3. Table 2 summarizes the main reaction conditions and catalytic results for all the described complexes.
The first example of use of a manganese complex in CO 2 hydroboration was described by Leitner and coworkers in 2018. [23]n the presence of complex [MnBr {(Ph 2 PCH 2 SiMe 2 ) 2 NH}(CO) 2 ] ( 14), the selective formation of methoxyborane was obtained with high TON = 883 using pinacolborane (HBpin, 2.76 mmol), CO 2 (1 atm), catalyst (0.036 mol %), base (NaO t Bu, 0.2 mol %), under solventless conditions at moderate temperatures (100 °C).The proposed stepwise reduction mechanism (Scheme 16) begins with the base-assisted activation of 14 to generate the coordinatively unsaturated active species [Mn{(Ph 2 PCH 2 SiMe 2 ) 2 N}(CO) 2 ] (14 a) bearing the deprotonated PNP ligand.This step has been confirmed experimentally by 31 P{ 1 H} NMR, with the observation of a signal at 61.3 ppm due to 14 a.Reaction of 14 a with HBpin leads to the activation of the BÀ H bond of the hydroborane by formation of a four-center coordination moiety bearing a new BÀ N bond and partial hydride transfer to the metal center in intermediate 14 b, characterized by a 1 H NMR signal at À 6.6 ppm and a 31 P{ 1 H} NMR signal at 65.2 ppm.For 14 b, single crystal structure X-ray determination was also obtained.Mechanistic studies showed that heating a sample of 14 b in THF-d 8 caused its slow conversion to a Mn-hydride complex upon full Scheme 15.Proposed stepwise CO 2 hydrogenation to CH 3 OH catalyzed by 13.Reprinted (adapted) with permission from [30].Copyright 2023 American Chemical Society.The reaction mechanism was later studied by other authors by DFT calculations.Wan, Li and Liao in 2021 carried out calculations on the reaction pathway, to pinpoint the unsolved key issues of the mechanism proposed by Leitner and coworkers on the basis of experimental data.As in the experimental paper, the mechanism was divided in four steps, namely i) the addition of HBpin to the unsaturated Mn active species, ii) the reduction of CO 2 to (OCHO)Bpin, iii) the reduction of (OCHO)Bpin to HCHO and iv) the reduction of HCHO to (CH 3 O)Bpin.The calculations confirmed that the common feature of all reduction steps is the ligand-assisted activation of HBpin by σ-bond insertion into the unsaturated MnÀ N bond in complex 14 a, followed by nucleophilic attack of the MnÀ H complex to the electrophilic carbon centers.The hydride transfer to (OCHO)Bpin was found to be the rate-limiting step for the whole catalytic cycle, with a total barrier of 27.0 kcal mol À 1 .It was confirmed that CO 2 activation in the first step occurs by BÀ O coordination to give 14 c, followed by a concerted hydride transfer to give a Mn-coordinated k 1 -Hformato species in an almost barrierless step.Formato ligand rearrangement follows with the formation of a NÀ B-O bond, ending with release of (OCHO)Bpin and regeneration of 14 a.This step is endoergonic by 6.2 kcal mol À 1 .In the following step, the reduction of (OCHO)Bpin to HCHO and (Bpin)O(Bpin), the substrate activation occurs by formation of a BÀ O bond between the N-ligated Bpin in 14 b and hydride transfer to the C atom of formoxyborane.The lowest energy key species in this step is a four-centered MnÀ NÀ BÀ OCH 2 -Bpin intermediate at À 25.0 kcal mol À 1 .The pathway proceeds by cleavage of the BÀ O bond, formation of a novel BÀ OÀ B bond and release of HCHO and 14 a.The final step, the reduction of HCHO to (CH 3 O)Bpin, follows a similar sequence, with the highest barrier at 18.1 kcal mol À 1 .Thermodynamic and kinetic analysis showed that the reduction reactivity follows the order HCHO > CO 2 > (OCHO)Bpin @ (CH 3 O)Bpin. [34]s an alternative study, in 2022 Cao, Lei and coworkers [35] approached the DFT study from a different point of view, i. e. by considering a possible inner-sphere activation mechanism involving prior Mn-CO bond dissociation, that was compared to a competitive outer-sphere carbonyl association mechanism involving N-ligated Bpin as in the previous study. [34]The CO dissociation mechanism involves the release of one CO ligand from 14 a to give a free coordination site that can in turn [a] Products:formoxyborane (A), bis(boryl)acetal (B), methoxyborane (C), bis(boryl)ether (D); type of hydroborane used in brackets; [b] Yield of product C; [c] TON was calculated for product C.
coordinate the incoming substrate.The calculated results show that the highest barrier in the carbonyl association mechanism is 27.0 kcal mol À 1 , corresponding to the hydride transfer step, while the rate-determining step of the carbonyl dissociation mechanism is the initial CO ligand release, with an energy barrier of 35.2 kcal mol À 1 .Thus, the association mechanism was confirmed as the most accessible for this system.Gonsalvi, Kirchner and coworkers described in 2020 the reduction of CO 2 to methoxyboranes catalyzed by the welldefined manganese complexes [MnX(PNP NR -iPr)(CO) 2 ] (R=H, X=H, 1 a; R=Me, X=H, 1 b; R=H, X=Br, 1 h) shown in Figure 3, using HBpin and 9-BBN (9-BBN = 9-borabicyclo[3.3.1]nonane) as reductants and various organoborates as Lewis acids additives. [36]The best results using HBpin (2.24 mmol) or 9-BBN (1.12 mmol), CO 2 (1 atm), catalyst (0.05 mol %), B(OPh) 3 (10 mol %) in THF-d 6 at moderate temperature (60 °C), were obtained in the presence of complex 1 a, with formation of (CH 3 O)Bpin and (CH 3 O)-9-BBN in 78 and 74 % yield, respectively.Complexes 1 b and 1 h led to (CH 3 O)Bpin in lower yields of 8 % and 31 % after 24 h, respectively.The formation of PhO-Bpin or PhO-9-BBN as side products of the reactions between boranes and B(OPh) 3 was observed.The proposed mechanism involves the initial activation of CO 2 by a putative highly electrophilic [Mn(PNP NH À iPr)(CO) 2 ][HB(OPh) 3 ] complex, suggested by the observation of a 31 P{ 1 H} NMR signal at 133.0 ppm during the test, allowing for the first kinetically sluggish reaction to give formoxyborane.Further reductions steps leading to the desired methoxyborane products were supposed to follow a metal-free reactivity in the presence of borane and the borate Lewis acid.
Later in 2023 Rong, Lei, Liu, De Proft and coworkers studied this system by DFT calculations. [37]Two possible mechanisms, involving ligand dearomatization (DA) and hydrogen transfer (HT), were considered (Scheme 17).As a common feature for the two mechanisms, HBpin is first activated in the initial step by metal-ligand cooperation, with the formation of a NÀ B bond and hydride transfer to Mn.For the DA mechanism, the highest barrier (14.6 kcal mol À 1 ) corresponds to the breaking of the BÀ N bond prior to release of (OCHO)Bpin, passing through transition state TS1 d.For the HT mechanism, each cycle involves hydrogen transfer between the substrate and the Bpin-N ligated active species, with the highest barrier at 17.2 kcal mol À 1 corresponding to transition state TS1 e.Thus, the DA mechanism has a slightly lower barrier than HT in the CO 2 hydroboration to formoxyborane step.On the other hand, for the formaldehyde formation and formaldehyde hydroboration steps, HT is more favored than DA.The authors concluded that the overall reaction should be a mixture of these two mechanisms.
Gonsalvi, Kirchner and coworkers reported in 2023 on the selective reduction of CO 2 to boryl-protected MeOH in the presence of pinacolborane (HBpin) and well-defined, bench stable Mn(I) non-pincer-type complexes (7, 7 h-j, Figure 3). [39]emarkably, complete hydroborane conversion and selective reduction to (CH 3 O)Bpin was demonstrated in the absence of any base or additive.The best results were obtained with complexes 7 and 7 j (1.0 mol % with respect to HBpin) in dmsod 6 under very mild reaction conditions (60 °C, 1 bar CO 2 , HBpin 0.224 mmol).Under these conditions, using 7 as a catalyst, (CH 3 O)Bpin was obtained in very high yield (83 %) after 1 h, that increased further after 24 h (99 %).In case of 7 j the reaction is faster and after 3 h full and selective conversion to methoxyborane was observed (yield = 100 %).Formoxyborane was observed to form only in small amounts (ca.6 %) at the beginning of the reaction.In the case of complexes 7 h and 7 i, the reaction was much slower and after 24 h gave (CH 3 O)Bpin in 76 % (40 °C) and 12 % yields (60 °C), respectively.The typical catalytic reaction was monitored by 1 H and 31 P{ 1 H} NMR, and the formation of various manganese complexes (7 k-n, Figure 4) was observed over time.
Complex 7 k was then independently synthesized either by the reaction of 7 with HBpin (2 equiv) and dmso (2 equiv) in THF at 60 °C or upon heating a solution of 7 j in dmso-d 6 to 40 °C.Complex 7 e is characterized by 1 H NMR signals at Scheme 17. Highest energy transition states in DA and HT mechanisms considered for CO 2 hydroboration with 1 a.Reprinted (adapted) with permission from [37].Copyright 2023 American Chemical Society.À 6.90 ppm (t, 2 J HP = 49 Hz) and À 7.02 ppm, (t, 2 J HP = 49 Hz); 31 P { 1 H} NMR, δ = 120.9ppm (d, 2 J PP = 22 Hz), 105.2 ppm (d, 2 J PP = 22 Hz).Complex 7 l was characterized as a singlet at 86.8 ppm in the 31 P{ 1 H} NMR monitoring, accompanied by a broad signal at 7.92 ppm in the corresponding 1 H NMR spectrum.Complex 7 m was observed at 115.2 ppm in the 31 P{ 1 H} NMR spectra, and it was identified by comparison with literature data and independent synthesis.Finally, the 31 P{ 1 H} NMR signal at 89.5 ppm was tentatively attributed to complex fac-[Mn(k 1 -O-OCHO)(dippe)(CO) 3 ] (7 n), that was separately obtained by reacting 7 a with formic acid.

CO 2 hydrosilylation
Another catalytic protocol applied for the homogeneous reduction of CO 2 is hydrosilylation.For this type of reaction, a wide library of hydrosilanes is commercially available, from mono-to bis-and tris-hydridosilanes, having different solubilities, electronic and steric properties.The use of hydrosilanes has the main advantage of replacing an explosive, pressurized gas such as hydrogen, and the lower energy required to activate the SiÀ H bond compared to the HÀ H bond. On the other hand, the principal drawback in CO 2 hydrosilylation is the need for highly chemoselective catalysts, in order to avoid the formation of mixtures of silyl formates, bis(silyl)acetals, methoxysilanes, bis(silyl)ethers, that in turn requires tedious purification and extraction (Scheme 1d).Whereas various reports are available in the literature for the selective synthesis of silyl formates, including many metal catalyzed protocols and recently a metal-free, solvent (DMSO, dimethylsulfoxide) promoted approach, [40] fewer examples describe the selective synthesis of methoxysilanes.These organosilane compounds can be of practical interest as methanol precursors, so their selective and efficient synthesis from abundant feedstocks such as CO 2 can be of interest.
To date, only two reports describe the use of Mn(I) organometallic catalysts for CO 2 hydrosilylation.In 2019, Gonsalvi, Kirchner and coworkers disclosed a detailed study on the application of the Mn-PN 3 P complexes 1 a and 1 b as efficient, chemoselective catalysts for CO 2 hydrosilylation. [41]The selectivity of the reaction could be easily switiched from silyl formates to methoxysilanes by simple tuning of the reaction conditions.After an initial solvent screening, confirming previous reports on the beneficial effects of DMSO for this type of reactions, the tests were run using 13 CO 2 (1 bar), catalyst (0.014 mmol, catalyst:CO 2 = 1 : 10), PhSiH 3 (0.70 mmol, 5.0 equiv.SiÀ H bonds respect to CO 2 , 2.0 mol % catalyst to SiÀ H bonds), DMSO-d 6 (400 μl) in J-Young type NMR tubes, monitoring the reactions by 13 C{ 1 H} NMR.At 25 °C, silyl formates were obtained with both catalysts in nearly quantitative yields after ca. 10 min.At longer reaction times (28-46 h), silyl formates were gradually converted to methoxysilanes in high yields (93 %).The tests were repeated at 80 °C, obtaining selectively methoxysilanes in yields > 99 % in the presence of 1 a after 6 h.
Having shown that highly efficient CO 2 utilization could be achieved, the tests were then run in Schlenk tube conditions using an excess of CO 2 to silanes, to gauge effective silane utilization.In a typical experiment, a Schlenk tube equipped with a Teflon Young valve (total volume = 15 mL) was charged with DMSO-d 6 (400 μL), PhSiH 3 (0.23 mmol; 0.7 mmol "SiÀ H", ca. 1 equiv."SiÀ H" respect to CO 2 ), catalyst 1 a (0.007-0.014 mmol; 1 or 2 mol % respect to "SiÀ H") and CO 2 (1 bar).After 24 h at 80 °C, NMR analysis revealed the complete consumption of PhSiH 3 and 100 % selective formation of methoxysilanes.
The catalytic mechanism involving (1 j).The step has a significant barrier (24 kcal mol À 1 ) but the process is however thermodynamically favored by a free energy balance gain of ΔG = À 18 kcal mol À 1 .Finally, a third silane molecule attacks the Mn-coordinated methoxide in 1 j and, transferring a hydride from Si to Mn, liberates (CH 3 O)SiH 2 Ph and Scheme 18. Proposed catalytic cycle for CO 2 hydrosilylation to methoxysilane (CH 3 O)SiH 2 Ph in the presence of 1 a.Energy barriers (kcal mol À 1 ) in brackets.Reprinted (adapted) with permission from [41].Copyright 2019 American Chemical Society.
regenerates 1 a, with a barrier of only 8 kcal mol À 1 .The calculations showed that the overall reaction is exergonic with ΔG = À 28 kcal mol À 1 .
In 2021, Gonzalez and Garcia reported the use of [MnBr(CO) 5 ] as an efficient pre-catalyst for CO 2 hydrosilylation with Et 3 SiH as silane under mild reaction conditions. [42]sing THF as a solvent, triethylsilylformate Et 3 Si(OCHO) was obtained in 67 % yield after 1 h at 50 °C and CO 2 pressure (4 bar).By changing the solvent to a THF/toluene mixture, the bis(triethylsilyl)acetal (Et 3 SiO) 2 CH 2 was instead formed in 86 % yield.At room temperature and atmospheric CO 2 pressure, the reaction still proceeded with high Et 3 SiH conversions (92 %-99 %), albeit with a decrease in the selectivity, regardless of the solvent used.Radical species were proposed to be involved in the proposed catalytic mechanism (Scheme 19), based on radical trapping experiments.

Summary and Outlook
In summary, in this review article it has been shown that the use of an earth-abundant metal such as manganese can be relevant to the timely field of carbon dioxide reduction processes.Although this metal has received attention for this purpose only in the last decade, a growing number of catalytic studies are appearing in the literature.Key points for the successful implementation of Mn-based homogeneous processes, in view of the efficient replacement of the precious metal analogues, is the design of efficient metal-ligand architectures and choice of ancillary ligands.Pincer ligands are at the moment the most represented class, and outstanding results (total TON > 2000000) have been achieved by an evolutionary approach, for example in passing from PNP to PN 3 P to PN 5 P systems.
The success of pincer ligands resides in the easier synthetic tunability of the electronic and steric properties than other class of ligands, the ability to impart rigid meridional geometry to the complexes, in turn avoiding isomerization eventually leading to less active catalytic species, and the possible ligand participation in MLC-type substrate activation mechanisms.Some drawbacks are still to be solved, namely i) replacing expensive Mn precursors with cheaper analogues for the synthesis of the corresponding organometallic catalysts, and ii) achieving a better integration of predictive DFT calculations with experimental studies, to help avoiding time-consuming synthetic work leading to less interesting catalysts.In our opinion, the future studies should focus on the principal use of green H 2 as reductant, targeting more challenging products of (captured) CO 2 reduction, such as methanol and methane, to provide solutions for renewable, non-fossil based fuels, energy solutions and greenhouse gas control, based on the reuse of CO 2 as an abundant feedstock.

Sylwia
Scheme 1.General scheme of CO 2 reduction processes and possible products.

4 b
Scheme 3. Calculated mechanism for CO 2 hydrogenation to formate in the presence of 2 a. Reproduced from Ref.[14] with permission from the Royal Society of Chemistry.

HCO 2 K
was observed in 23 % and 6 % yields in the case of 4 a and 4 b, respectively.
Scheme 4. Reactivity pathways of complexes 4 a and 4 b with CO 2 and H 2 .Reproduced from Ref.[15] with permission from the Royal Society of Chemistry.

Figure 2 .Scheme 9 .
Figure 2. Calculated reaction pathway of the hydrogenation of HCOOH to methanediol and the decomposition of methanediol to HCHO and water catalyzed by complex [Mn{(Ph 2 PCH 2 SiMe 2 ) 2 NH}(CO) 2 ].Reproduced from Ref.[22] with permission from the Royal Society of Chemistry.

Scheme 13 .Scheme 14 .
Scheme 13.Proposed mechanisms for the synthesis of formamide from CO 2 hydrogenation (a) and formamide hydrogenation to methanol (b) starting from pre-catalyst 9 a.Reprinted (adapted) with permission from[28].Copyright 2017 American Chemical Society.

Figure 4 .
Scheme 17. Highest energy transition states in DA and HT mechanisms considered for CO 2 hydroboration with 1 a.Reprinted (adapted) with permission from[37].Copyright 2023 American Chemical Society.Figure 4. Manganese complexes 7 k-n identified by NMR monitoring during CO 2 hydroboration runs in the presence of 8 a.
1 a was studied by DFT calculations, and the proposed catalytic cycle is shown in Scheme 18.The initial step involves the insertion of CO 2 into the MnÀ H bond, resulting in the metal-coordinated formate [Mn{k 1 -O-(OCHO)}(PNP H -iPr)(CO) 2 ] (1 d).This step requires to overcome a barrier of 8 kcal mol À 1 .Then, a first silane molecule attacks the coordinated O-atom of formate in 1 d, transferring one hydride from Si to Mn, giving free silyl formate (OCHO)SiH 2 Ph with regeneration of 1 a.This step has the highest barrier of the entire path (29 kcal mol À 1 ).From this stage, the reaction follows by MnÀ H attack onto the carbonyl Catom of the silyl formate, generating [Mn{(OCH 2 O)SiH 2 Ph}(PNP H -iPr)(CO) 2 (1 i), bearing a a Mn-coordinated silyl hemiacetal ligand, with a barrier of 20 kcal mol À 1 .The next step involves the attack of a second silane to the non-coordinated O-atom of the hemiacetal and the hydride transfer from the Si-atom to the carboxyl C-atom, liberating the silyl ether (PhSiH 2 ) 2 O and forming the Mn methoxo-complex [Mn(OCH 3 )}(PNP H -iPr)(CO) 2 ]

Table 1 .
Summary of main reaction parameters, type of products and best TONs for Mn(I)-catalyzed CO 2 direct hydrogenation.

Table 2 .
Summary of main reaction parameters, type of products and best yields and TONs for Mn(I)-catalyzed CO 2 hydroboration.