Non-bifunctional Mn catalysts based on phosphine–phosphites for the hydrogenation of carbonyl substrates

Abstract

A series of Mn(I) complexes bearing phosphine–phosphite ligands (P-OP) of formula [Mn(Br)(P-OP)(CO)₃] (1) have been prepared and characterised. These compounds exist in solution as a mixture of fac and mer isomers with fac/mer ratios from 18 : 1 to 2 : 1. The relatively low donor ability of P-OP compared to bis(trialkylphosphines) or bis–NHC ligands, typically found in non-bifunctional Mn catalysts reported to date, is evinced in IR spectra. Compounds 1 provide efficient catalysts for the hydrogenation of carbonyl substrates using KOtBu as a base. Notably, these complexes constitute a set of modularly designed precatalysts and optimisation of the P-OP ligand provided appropriate catalysts for the reduction of a wide variety of aldehydes and ketones under mild reaction conditions (S/C = 250–500, 20 bar H₂, RT to 40 °C). A mechanistic study combining spectroscopic and computational results provides support for a catalyst activation process based on the nucleophilic attack of the alkoxide on a coordinated CO in [Mn(OtBu)(P-OP)(CO)₃], as well as a non-bifunctional inner-sphere hydrogenation pathway in the representative hydrogenation of benzophenone. The catalytic cycle is also characterised by a rather stable pentacoordinated alkoxide [Mn(k¹-O–OCHPh₂)(P-OP)(CO)₂] which determines the energetic span of the reaction.

---

Introduction

The search for cost-effective alternatives to well-established noble metal catalysts is one of the more important research topics in homogeneous hydrogenation.¹ In this regard, Mn catalysts have already shown particularly promising results in terms of broad substrate scope and catalyst activity.² This field emerged a decade ago with prominent examples of bifunctional catalysts,³ which boosted the research on this type of systems.^4,5 Alternatively, non-bifunctional ones have remained much less explored and are notably restricted to derivatives of highly donating bidentate ligands. The corresponding catalyst precursors have the formula Mn(X)(L–L)(CO)₃ where X denotes an anionic ligand such as Br, alkyl or triflate, while L–L corresponds to bis(trialkylphosphine) or to NHC based ligands. For instance, complex A (X = Br, R = nPr; Fig. 1)^6a provides an active catalyst for the hydrogenation of ketones and nitriles, while for the latter a triflate derivative (R = iPr) reported by the group of García operated under milder conditions.⁷ Moreover, Kirchner and coworkers have demonstrated that alkyl compound B has a remarkable reactivity in hydrogenation and is capable to reduce without the need of base ketones, alkenes, alkynes and CO₂.^6b–e On the other hand, bis–NHC derivatives C (ref. 8) provide outstanding results in the challenging hydrogenation of esters.^9,10 An example of this type developed by Beller and coworkers (R = R′ = Me) exhibits a broad scope and also hydrogenates a wide variety of ketones, nitriles and olefins.¹⁰ Finally, the group of Sortais has reported an interesting case of a complex based on a NHC–thioether ligand (D) which leads to an active catalyst for the hydrogenation of olefins and ketones.¹¹ These precedents raise the question of whether it is possible to expand catalyst activity in hydrogenation reactions to less donating ligands (e.g. arylphosphines, phosphoramidites, phosphites).¹² This is an area with a high potential due to the vast structural diversity of these P-ligands, which span a wide range of electronic and steric properties, features that are most useful for comprehensive catalyst optimisation.¹³

Fig. 1 Structure of compounds A–D.

In our laboratory we have been interested in the study of phosphine–phosphite ligands (P-OP) in Rh, Ir and Ru catalytic hydrogenation,¹⁴ while applications in other catalytic processes such as hydroformylation (Rh),¹⁵ conjugate addition or alkylation (Cu)¹⁶ cycloaddition (Au)¹⁷ or hydrocyanation reactions (Ni),¹⁸ have also been described. These ligands are characterized by an easily tuneable structure and a π-acceptor phosphite fragment which makes them less donor than ubiquitous diphosphines,¹⁹ while the different electronic properties of their P fragments cause important directing effects in catalysis.²⁰ In this contribution we report the first application of P-OP ligands in Mn catalysed hydrogenation. Complexes prepared show good activity in the hydrogenation of a wide variety of ketones and aldehydes under mild reaction conditions, while complementary spectroscopic and computational results provide support for a low-barrier catalyst activation pathway and a non-bifunctional inner-sphere hydrogenation mechanism.

Results and discussion

Synthesis and characterization of Mn phosphine–phosphite complexes

Compounds [Mn(Br)(P-OP)(CO)₃] (1a–1d) were obtained in high yields from [Mn(Br)(CO)₅] and a stoichiometric amount of P-OP ligand (L1–L4, respectively; Scheme 1). Moreover, 1a reacts with AgOTf to give [Mn(OTf)(L1)(CO)₃] (2a), while the latter reacts with KBH₄ to give the corresponding hydride [Mn(H)(L1)(CO)₃] (3a). For these complexes, four isomers can be distinguished: two chiral facial isomers, fac1 and fac2 (Fig. 2), which differ in metal configuration, and two meridional isomers mer1 and mer2, which can be differentiated by the relative positions of the X and P atoms. The fac isomers can be notated as OC-6-44-C and OC-6-44-A, respectively,²¹ although for simplicity, C (fac1) and A (fac2) will be used to refer to metal configuration. It is important to note that a stereogenic axis in the biaryl is generated upon phosphite coordination. However, a fast atropisomerization of the biaryl fragment at room temperature is expected, as observed before in Rh (ref. 22) and Ru (ref. 20b) complexes containing these P-OP ligands.

Scheme 1 Synthesis of complexes 1–3.

Fig. 2 Structure of isomers of complexes 1–3.

Characterisation in solution of compounds 1 show a mixture of two isomers, with the fac isomer being predominant, and fac/mer ratios from 18 : 1 (1c) to 2 : 1 (1d). This is an interesting difference with diphosphine derivatives [Mn(Br)(P–P)(CO)₃] which solely exist in solution as fac isomers.^6a,23 For compounds 1 comparison of ³¹P{¹H} NMR spectra show distinct data for the two isomers, thus J_PP coupling constants are between 12 and 17 Hz higher in the fac isomer. In addition, there is a marked low field shift (Δδ = 27–30 ppm) of the phosphite signal of the mer isomer compared to the fac one (Fig. 3a). Due to the relatively high concentration of mer-1d, full characterization in solution from ¹H and ¹³C{¹H} NMR (Fig. 3b) has been possible for this isomer. In contrast, for the triflate derivative 2a, the isomer ratio is 24 : 1, while for hydride 3a only the fac isomer was observed. Moreover, the observation of three strong bands in the carbonyl region of IR spectra of complexes 1–3 is also consistent with the fac isomer (e.g. 2030, 1969 and 1943 cm⁻¹ for 1a). Nevertheless, less intense bands attributable to minor mer isomers have also been detected (e.g. 1987 and 1960 cm⁻¹ for 1a). With the available spectroscopic data, it is not possible to ascertain the stereochemistry of the mer isomer, although it is reasonable to propose a mer1 structure in which CO and phosphite avoid mutually trans positions. These assumptions will be further supported by DFT calculations (see below). Compound 1a has been studied by ¹H and ³¹P{¹H} NMR at low temperature with the aim of freezing phosphite atropisomerization. However, down to −80 °C, no significant lineshape changes in these spectra were observed, indicating that atropisomerization is still fast at this temperature. This compound has also been characterized by single crystal X-ray crystallography (Fig. 4). In the crystal lattice a racemic mixture composed by C,S_ax and A,R_ax isomers was observed. Thus, stereoisomers differing in the relative metal and biaryl configuration (i.e. C,R_ax and A,S_ax) should be less stable.

Fig. 3 (a) ³¹P{¹H} NMR (C₆D₆, 202 MHz) spectrum of 1d and (b) aliphatic region of ¹³C{¹H} NMR (C₆D₆, 125 MHz) spectrum of 1d.

Fig. 4 ORTEP view of C,S_ax isomer of complex 1a (H atoms have been omitted for clarity).

An interesting difference between compounds 1 and non-bifunctional catalyst precursors found in literature concerns the lower electron density at the metal centre in the former. This is evidenced by the position of υ(CO) bands in IR spectra. Thus, the bands of 1a appear at higher frequencies than those of A (X = Br, R = nPr; 2003, 1935 and 1903 cm⁻¹),^6a C (R′ = R′′ = Me; 1995, 1908 and 1868 cm⁻¹)¹⁰ and D (2010, 1925 and 1894 cm⁻¹).¹¹ A similar trend is observed when by comparing 2a or 3a with their 1,2-bis(diphenylphosphino)ethane analogues.²⁴

To complete the study, we have investigated the relative stability of the isomers of representative complexes 1a (Fig. 5) and 1c by DFT methods.²⁵ For the sake of clarity, the S_ax configuration at the biaryl was selected throughout the study. In the case of 1a, the most stable structure corresponds to fac1-1a, matching the structure observed by X-ray crystallography, while diastereomeric fac2-1a is 2.8 kcal mol⁻¹ higher in free energy. In addition, mer1-1a is 3.7 kcal mol⁻¹ less stable than fac1-1a, while mer2-1a is the least stable isomer (6.4 kcal mol⁻¹). For 1c, the preferred structure is fac2-1c, while the fac1, mer1 and mer2 isomers lie 3.2, 2.9 and 7.0 kcal mol⁻¹ higher in free energy, respectively. Likewise, analysis of the isomers of 3a showed that the fac1 isomer is the more stable, with the fac2, mer1 and mer2 structures lying 2.1, 3.8 and 5.2 kcal mol⁻¹ above fac1-3a, respectively. These results are consistent with the experimental preference for the fac isomer, while the small energy differences observed between fac1 and fac2, besides a fast atropisomerization, suggest a mixture of these facial isomers in solution, along with minor amounts of mer1. In contrast, mer2 isomers can be disregarded due to their significantly higher energies. Similarly, a difference of 7.9 kcal mol⁻¹ between A (X = Br, R = nPr) and the corresponding mer isomer has been reported.^6a

Fig. 5 Optimized geometries of isomers of 1a. Numerical values correspond to free energies relative to fac1 isomer (kcal mol⁻¹, hydrogens have been omitted for clarity).

Catalytic hydrogenation of carbonyl compounds

To examine the potential of compounds 1 in catalytic hydrogenation, we first carried out a series of hydrogenations of acetophenone with these complexes under relatively mild reaction conditions in iPrOH using KOtBu as a base (20 bar H₂, RT, S/C/B = 500/1/10). We were pleased to observe that complex 1a provided nearly full conversion under these conditions (entry 1; Table 1), while the remaining complexes gave moderate values (65, 53 and 61 %, for 1b–1d, entries 2–4 respectively). From these results, we selected 1a for further optimisation. First, the use of alternative solvents was examined, with satisfactory results obtained in toluene (entries 5, 7), while lower activity was observed in CH₂Cl₂ (entry 6). We then attempted to lower the catalyst loading. In this regard, it was observed that increasing the base concentration led to a decrease in conversion (entries 9–12). Notably, increasing the reaction temperature proved detrimental in both toluene (entries 13, 14) and iPrOH (entry 15). Finally, a mercury drop test indicated that the reaction is homogeneous (entry 16). In contrast to 1a, the triflate complex 2a showed reduced activity (entries 17, 18), while a reaction with hydride 3a in the absence of base lead to no conversion (entry 19). We next examined the substrate scope for ketone hydrogenations using 1a (Scheme 2). Notably, this compound displayed a broad substrate scope, reducing a variety of alkyl/aryl, dialkyl and diaryl ketones with conversion values exceeding 95 %. Worth to note, these results have been obtained with a lower catalyst loading and milder reaction conditions than those reported for other Mn(Br)(L–L)(CO)₃ precursors of non-bifunctional catalysts in these reactions (1–3 mol %, 30–50 °C, 50 bar H₂).^6a,11

Table 1 Hydrogenation of acetophenone with catalyst precursors 1–3^a

Entry	Complex	Solvent	S/C/B	Temp. (°C)	Conv. (%)
a Reactions performed under 20 bar H2 with solvent and temperature specified with KOtBu as a base, conversion was determined by ^1H NMR.b Mercury test.
1	1a	iPrOH	500/1/10	RT	99
2	1b	iPrOH	500/1/10	RT	65
3	1c	iPrOH	500/1/10	RT	53
4	1d	iPrOH	500/1/10	RT	61
5	1a	Toluene	100/1/10	RT	100
6	1a	CH2Cl2	100/1/10	RT	52
7	1a	Toluene	500/1/10	RT	100
8	1a	Toluene	500/1/5	RT	100
9	1a	Toluene	1000/1/5	RT	77
10	1a	Toluene	1000/1/10	RT	61
11	1a	Toluene	1000/1/20	RT	49
12	1a	Toluene	1000/1/50	RT	34
13	1a	Toluene	500/1/10	40	82
14	1a	Toluene	1000/1/10	40	28
15	1a	iPrOH	500/1/10	60	19
16^b	1a	Toluene	500/1/10	RT	100
17	2a	Toluene	500/1/10	RT	17
18	2a	Toluene	500/1/0	RT	0
19	3a	Toluene	500/1/0	RT	0

---

Scheme 2 Conversion (isolated yield in parentheses) values obtained in ketone hydrogenation reactions with 1a.

Due to the limited applications of non-bifunctional catalysts in the hydrogenation of aldehydes,²⁶ we also investigated the performance of compounds 1 in these reactions (Scheme 3). Hydrogenation of benzaldehyde under mild conditions (20 bar H₂, RT, toluene) using 1a showed full conversion after 24 h. Product analysis indicated the preferential formation of benzyl alcohol, along with benzyl benzoate, resulting from a Tishchenko-type condensation,²⁷ with an alcohol/ester ratio (A/E) of 4.1 (entry 1, Table 2). Using 1b led to a slightly lower conversion but a significant increase in the A/E ratio (entry 2). Increasing the temperature to 40 °C resulted in slightly higher selectivity (entry 3). We reasoned that the formation of the ester could be disfavoured by lowering the substrate concentration. Thus, the A/E ratio improved up to 23.0 (entry 4). The superior performance of 1b was further confirmed by repeating the reaction with 1a under the same conditions (entry 5). A blank reaction under these conditions showed nearly full conversion, but the ester as the main product (entry 6). Comparison of reactions of entries 4 and 6 clearly illustrates the significant impact of the catalyst on product selectivity. To further explore the utility of this catalytic system, we examined the hydrogenation of various aldehydes. Hydrogenation of several aryl aldehydes showed full conversion within 24 h and a good selectivity towards the desired alcohol, with A/E ratios up to 50.0. A couple of α,β-unsaturated aldehydes were also tested. In these cases, the preferential formation of the corresponding allyl alcohols was observed, with good selectivity.

Scheme 3 Hydrogenation of aldehydes using complex 1b. All reactions proceeded with full conversion. For aryl aldehydes the numerical values correspond to the alcohol/ester ratio. In the case of cinnamaldehyde and geranial, the values indicate the ratio between the allyl alcohol and the fully hydrogenated product.

Table 2 Hydrogenation of benzaldehyde with catalyst precursors 1a and 1b^a

Entry	Complex	S/C/B	[S]^b	Temp. (°C)	Conv. (%)	Select (A/E)^c
a Reactions performed in toluene under 20 bar H2, temperature specified with KOtBu as a base. Conversion and selectivity were determined by ^1H NMR.b Substrate molar concentration.c Alcohol/ester ratio.
1	1a	500/1/10	1.1	RT	100	4.1
2	1b	500/1/10	1.2	RT	96	9.3
3	1b	500/1/10	1.2	40	100	11.5
4	1b	250/1/10	0.3	40	100	23.0
5	1a	250/1/10	0.3	40	100	6.8
6		250/0/10	0.3	40	98	0.1

---

Mechanistic spectroscopic results

An interesting aspect of catalytic hydrogenation using non-bifunctional [Mn(X)(L–L)(CO)₃] (X = Br, OTf, alkyl) complexes concerns their activation to generate the unsaturated hydride [Mn(H)(L–L)(CO)₂], which is responsible for catalysis. For complex B, it has been clearly shown that activation proceeds via migratory insertion followed by hydrogenolysis, leading to butyraldehyde and [Mn(H)(dippe)(CO)₂].^6d Alternatively, it has been proposed that CO dissociation in [Mn(O^sBu)(dippe)(CO)₃], followed by hydrogen coordination and alkoxide protonation can also lead to the active species.⁷ In addition, the hydride [Mn(H)(H₂)(L–L)(CO)₂] has been proposed as the key species in reactions involving complex D, although the route leading to it is not fully understood.¹⁰

In light of the above, we investigated the activation of compounds 1 under hydrogenation conditions (Scheme 4). We observed that compound 1a reacts with KOtBu (3 equivalents) under 0.5 bar H₂ at room temperature in toluene-d₈. However, this reaction proceeds very slowly due to the low solubility of the base. The addition of 20 μL of THF to the NMR sample increased the reaction rate, showing ca. 30 % conversion after 1 h at room temperature. Alternatively, a faster reaction was observed is the base is dispersed by sonication (15 min; Fig. 6a and b; see SI for additional experiments). The reaction between 1a and KOtBu also took place in the absence of hydrogen, but resulted in a complex mixture, while no reaction between 1a and H₂ was observed without base. Under H₂, the reaction with base is apparent from a colour change in the solution from yellow-orange to pale-yellow. In the hydride region of the ¹H NMR spectrum, a very broad hump composed by two overlapping broad signals at −8.6 and −9.4 ppm, and a doublet of doublets for 3a in a 6 : 1 ratio, were detected (Fig. 7a). Lowering the temperature to −80 °C did not reach the slow exchange regime, although at −40 °C it was observed that the broad signal split into three still broad resonances centred at −7.6, −8.2, and −10.3 ppm in an approximate 1 : 3 : 2 ratio. In the ³¹P{¹H} NMR spectrum at 25 °C (toluene-d₈), a broad signal appears in the phosphite region (201.8 ppm), along with two broad signals in the phosphine region (70.2 and 66.9 ppm), and doublets corresponding to 3a, with unreacted 1a present in a 6 : 1 : 3 ratio, respectively (Fig. 7b). At −40 °C, the phosphine region of the ³¹P{¹H} NMR experiment also showed a splitting into three broad signals. These observations indicate the formation of a mixture of fluxional hydrides (Hf) upon the reaction of 1a with KOtBu under H₂. Exposure of this mixture to a CO atmosphere leads to conversion into 3a. After 15 min, the ratio of Hf to 3a is 2 : 1, which evolves to 0.3 : 1 after 2 h. Alternatively, broad resonances centred at −9.1 and 1.0 ppm are observed in the ²H NMR, in an approximate 1 : 1 ratio. The latter signal can be assigned to tBuOD, likely involved in a coordination/decoordination process. These experiments suggest that Hf corresponds to a mixture of isomers based on the Mn(H)(L1)(CO)₂ fragment which can be stabilized by solvent coordination. As well, computational studies show that these species can also be stabilized by an agostic interaction involving a tBu group (see below).

Scheme 4 Generation and reactions of key unsaturated Mn(H)(L1)(CO)₂ (Hf).

Fig. 6 ³¹P{¹H} NMR (202 MHz, toluene-d₈): spectra of 1a (a), activated 1a by reaction with H₂ and KOtBu after 15 min sonication (b), reaction of the latter with benzophenone (5 equiv., c).

Fig. 7 ¹H (400 MHz, a) and ³¹P{¹H} NMR (162 MHz, b) spectra of activated 1a in toluene-d₈.

Furthermore, Hf reacts with benzophenone (3–5 equiv.), turning the solution deep red. This reaction is evinced by the disappearance of the hydride signal in the ¹H NMR and the appearance of a very broad resonance at 5.65 ppm (5.44 ppm in the ²H NMR for the deuteration experiment). In the ³¹P{¹H} NMR, a set of narrow doublets centred at 201.4 and 28.3 ppm with ²J_PP coupling constant of 87 Hz is observed (Fig. 6c). Upon reintroducing H₂ (0.5 bar), the solution fades to pale orange, the fluxional hydride signals reappeared in the ¹H NMR, and catalytic hydrogenation proceeded, yielding Ph₂CHOH. The aforementioned doublets in the ³¹P{¹H} NMR remain visible until the latter stages of the reaction. Similarly, reaction of Hf with 4,4′-difluorobenzophenone yields two doublets in the ³¹P{¹H} NMR at 201.3 and 28.5 ppm (²J_PP = 88 Hz), a broad signal in the ¹H NMR at 5.38 ppm, and two broad signals in the ¹⁹F{¹H} NMR at −115.4 and −117.0 ppm (Fig. 8a and b). At low temperature, no appreciable changes were observed in the former signal while the latter splits into two singlets which appeared narrow at −40 °C at −116.5 and −117.0 ppm, corresponding to two non-equivalent ¹⁹F nuclei (Fig. 8e), which parallel data reported by Baratta and coworkers for an Os–OCH(4-F-Ph)₂ alkoxide.²⁸ In turn, the signal at ca. −115 ppm should correspond to (4-F-Ph)₂CHOH generated from the residual H₂ present in solution. This signal appears broadened with respect to that of the free alcohol attributable to an exchange between free and coordinated alcohol. The signal at −117.0 ppm disappeared upon substrate consumption by reaction with H₂, raising the signal of the alcohol. Monitoring of the hydrogenation replenishing the H₂ atmosphere showed a clean hydrogenation (Fig. 8c and d). These results agree with the formation of alkoxides Mn(OCHAr₂)(L1)(CO)₂ (Ar = Ph (Ak), 4-F-Ph (Ak^F)) via reaction of Hf with the diarylketone, and subsequent hydrogenation to generate the corresponding diarylcarbinols. Attempts to isolate alkoxide intermediates were unsuccessful due to sample degradation during work-up.

Fig. 8 ¹⁹F{¹H} NMR (282 MHz, toluene-d₈) spectra of reaction between activated 1a and 4,4′-difluorobenzophenone (1 equiv.) after 15 min (a); after 45 min (b). Reaction with 4,4′-difluorobenzophe-none (5 equiv.): immediately after ketone addition (c); after 6 h reaction under 0.5 bar H₂ (d). ¹⁹F{¹H} NMR (376 MHz, toluene-d₈) VT spectra (e) of reaction between activated 1a and 4,4′-difluorobenzophenone (5 equiv.).

Computational study

Catalyst activation. To support the experimental observations, we carried out DFT studies of the key steps involved in the activation of complexes 1. As a reasonable starting point, we considered the formation of Mn(OtBu)(L1)(CO)₃ (4) upon reaction of 1a with KOtBu.^29,30 According to literature precedents proposing either CO dissociation^7,23b or alkoxide migration³¹ in related alkoxide–diphosphine derivatives, we examined in detail both activation routes in alkoxide 4. We studied the fac1, fac2 and mer1 isomers of 4.³² Among these, the fac2 isomer was found to be the most stable, with fac1 and mer1 lying 2.9 and 6.5 kcal mol⁻¹ higher in free energy, respectively.

In the CO dissociation route (Fig. 9), release of the CO ligand trans to the phosphite in fac1-4 leading to a pentacoordinated alkoxide intermediate I1a is almost thermoneutral (ΔG = 0.9 kcal mol⁻¹) and proceeds with a moderate free energy barrier of 16.3 kcal mol⁻¹ (TS1a). Hydrogen coordination to Ia1 can take place either trans to CO or to the phosphite ligand. The former pathway, which locates the alkoxide trans to the phosphite fragment is energetically preferred and leads to a dihydrogen complex I2a which lays 19.0 kcal mol⁻¹ above fac1-4. This complex features a rather unactivated H₂ ligand (d(Mn–H) = 1.926, 1.967 Å; d(H–H) = 0.798 Å). The subsequent alkoxide protonation step (TS3a) represents the highest overall barrier in this pathway (ΔG^‡ = 22.6 kcal mol⁻¹), yielding a hydride–tBuOH complex I4a (ΔG = 1.5 kcal mol⁻¹). IRC analysis of TS3a leads back to a dihydrogen complex I3a (ΔG = 20.2 kcal mol⁻¹) nearly isostructural to I2a but with a more activated H₂ ligand, as shown by Mn–H and H–H distances (d(Mn–H) = 1.712, 1.725 Å; d(H–H) = 0.828 Å). The small energy difference between I2a and I3a (ca. 1.2 kcal mol⁻¹) suggests facile interconversion between these species. Finally, hydride–alcohol complex I4a is expected to exchange readily the tBuOH ligand for a ketone. In the case of benzophenone, the equilibrium favours the k¹-O ketone adduct by 1.1 kcal mol⁻¹. The activation profile via CO dissociation parallels that reported by Beller and co-workers for tert-butoxide generated from C (R = H, R′ = Me), although the barriers found for fac1-4 are around 5 kcal mol⁻¹ lower.¹⁰

Fig. 9 General scheme for the activation of fac1-4 by CO dissociation. Numerical values correspond to relative free (red) and electronic energies (blue, in parentheses) in kcal mol⁻¹ relative to fac1-4.

For the fac2-4 isomer, CO dissociation also proceeds with a moderate barrier (TS1b, ΔG^‡ = 15.7 kcal mol⁻¹) in an endergonic step (5.4 kcal mol⁻¹), while for the least stable alkoxide mer1-4, the CO dissociation barrier is even lower (TS1c, ΔG^‡ = 8.0 kcal mol⁻¹) and the process slightly exergonic (−0.9 kcal mol⁻¹). Worth to note, the release of CO is accompanied by a significant shortening of the Mn–O bond (e.g. 2.206 Å in fac1-4 vs. 1.839 in I1a). In contrast, significantly high barriers (ΔE > 30 kcal mol⁻¹) have been computed for the dissociation of CO in either fac1 or mer1 isomers of 1a and 3a. As in the profile depicted for fac1-4, alkoxide protonation remains the rate-limiting step in the cases of fac2-4 and mer1-4, with corresponding TS3b and TS3c lying 2.6 and 2.3 kcal mol⁻¹ in free energy above TS3a.

On the other hand, a lower-energy pathway was identified for the activation of 4 involving nucleophilic attack of the coordinated OtBu on a CO ligand (Fig. 10).^33,34 In fac1-4, this attack to the CO ligand trans to phosphite (TS1a′) proceeds with a low barrier (ΔG^‡ = 9.1 kcal mol⁻¹), leading to a Mn–C bound carboxylate intermediate I1a′ (ΔG = 4.6 kcal mol⁻¹). Hydrogen coordination yields a dihydrogen complex I2a′ (ΔG = 5.4 kcal mol⁻¹), which undergoes protonation at the carboxylate carbon to form an H-bound formate complex I3a′ (ΔG = 10.5 kcal mol⁻¹). This intermediate undergoes fast isomerisation to the O-bound formate I4a′ (ΔG = −0.8 kcal mol⁻¹). The highest overall barriers in this pathway are associated with hydrogen coordination TS2a′ and protonation TS3a′ (both 13.0 kcal mol⁻¹), which compete favourably with those in the CO dissociation route. The alternative OtBu migration pathway from the most stable alkoxide isomer fac2-4 is less favourable, with a TS for carboxylate protonation (TS3b′) 4.7 kcal mol⁻¹ higher in free energy than TS3a′. Finally, dissociation of tert-butyl formate from I4a′ by tBuOH or benzophenone is thermodynamically favourable (ΔG = −2.9 and −4.1 kcal mol⁻¹, respectively). However, no tert-butyl formate was detected experimentally. We have observed that under reaction conditions used for catalyst activation, this ester reacts with KOtBu to produce tBuOH and CO (see SI for further details),³⁵ the latter potentially responsible for the formation of Mn(H)(L1)(CO)₃ observed in experiments described above.

Fig. 10 General scheme for the activation of fac1-4a by OtBu migration. Numerical values correspond to relative free (green) and electronic energies (blue, in parentheses) in kcal mol⁻¹ relative to fac1-4.

Benzophenone hydrogenation. We next examined the mechanism of ketone hydrogenation catalysed by the present system. To reduce the number of diastereomeric intermediates, we focused on the hydrogenation of benzophenone rather than acetophenone, as both substrates are readily hydrogenated by 1a. As a reference, we followed the key features of the inner-sphere mechanism previously reported for acetophenone hydrogenation catalysed by Mn(H)(dippe)(CO)₂.^6d

Initially, we considered isomers of the unsaturated hydride Mn(H)(L1)(CO)₂ (5, Fig. 11). Based on the activation mechanisms considered, formate dissociation in I4a′ (or tBuOH release in I4a) leads to the pentacoordinated hydride 5a. Given the expected facile phosphite atropisomerisation, diastereotopic isomer 5b was also considered.³⁶ A plausible isomerisation route involves hydride migration to the equatorial plane. This process is indeed feasible, with relatively low free energy barriers for the conversion of 5a to 5d (TSad: 9.8 kcal mol⁻¹) and 5b to 5c (TSbc: 11.7 kcal mol⁻¹). These results suggest a facile exchange between these unsaturated hydrides, consistent with spectroscopic observations. Among the isomers, 5d is the more stable one as the result of an agostic interaction generated by a C–H of a tert-butyl substituent. Hydrides 5 adopt a distorted square planar geometry, with an available coordination site trans to either phosphite or a carbonyl ligand. Therefore, these species should readily coordinate an alcohol or carbonyl substrate to form the corresponding octahedral complexes. The hydrogenation pathway was examined for hydrides 5a–5d, leading to four routes a to d (Fig. 12, Table 3). The reaction begins with formation of a hydride–ketone precomplex I5, which evolves via TS5 to a k¹-O–benzophenone complex I6. This complex may isomerise through TS6 to a k²-O,C complex I7, which undergoes hydride insertion (TS7) to yield an agostic alkoxide I8 (path b). Alternatively, hydride insertion can occur directly from I6 (paths a, c and d). Release of the agostic interaction via TS8 affords a pentacoordinated alkoxide I9, which is considerably more stable (ΔG = −2.5 to −15.2 kcal mol⁻¹) than the preceding agostic species. I9 is suitable for hydrogen coordination (TS9), forming a dihydrogen–alkoxide complex I10. Finally, alkoxide protonation via TS10 yields the hydride–alcohol complex I11. A notable feature of the mechanism is the remarkable stability of alkoxides I9, being I9c the most stable, ΔG° = −21.3 kcal mol⁻¹ relative to the starting materials. According to the energy span model³⁷ analysis of these profiles, the turnover-determining intermediate (TDI) and turnover-determining transition states (TDTS) are I9 and TS10, respectively, across all four cycles. The more favourable TS for alkoxide protonation corresponds to path c (TS10c), which also contains the most stable alkoxide I9c. Moreover, considering the small structural differences found among the I9 alcoxides and their expected highly dynamic nature, it is reasonable to expect that they can easily interconvert. Upon these considerations, it can be assumed that the dominant pathway in the catalysis will correspond to the formation of I9c and subsequent steps via path c. Accordingly, this alkoxide would be the resting state of the catalytic cycle, in agreement with the persistent observation of Ak/Ak^F species upon NMR monitoring of the catalytic hydrogenation.

Fig. 11 Optimized geometries for Mn(H)(L1)(CO)₂ isomers (5a–5d). Numerical values correspond to relative free energies in kcal mol⁻¹ relative to 5d. Hydrogens have been omitted for clarity except those of the methyl involved in an agostic interaction in 5d.

Fig. 12 Free energy profiles for benzophenone hydrogenation catalysed by Mn(H)(L1)(CO)₂.

Table 3 Calculated Energies for benzophenone hydrogenation catalysed by Mn(CO)₂(H)(L1) by paths a to d^a

Species	ΔG (ΔE) (a)	ΔG (ΔE) (b)	ΔG (ΔE) (c)	ΔG (ΔE) (d)
a Values in kcal mol^−1 from 5c + benzophenone.
5 + Ph2C=O	−1.6 (−1.3)	1.6 (1.9)	0.0 (0.0)	−7.0 (−9.9)
I5	−4.4 (−22.8)	−0.6 (−20.5)	1.4 (−17.3)	-2.4 (−21.3)
TS5	−0.4 (−18.0)	4.1 (−14.2)	4.0 (−14.1)	3.4 (−13.5)
I6	−10.4 (−30.2)	−8.5 (−29.0)	−8.3 (−28.2)	−7.4 (−26.9)
TS6		−4.2 (−23.0)
I7		−3.5 (−22.8)
TS7	5.5 (−18.0)	5.0 (−13.7)	4.1 (−19.4)	−0.6 (−20.5)
I8	−7.3 (−30.1)	−8.9 (−29.5)	−6.1 (−26.4)	−5.2 (−28.8)
TS8	−7.0 (−27.4)	−8.1 (−29.5)	−8.9 (−29.5)	−6.4 (−26.3)
I9	−18.2 (−38.4)	−11.4 (−32.5)	−21.3 (−41.6)	−14.2 (−34.3)
TS9	2.1 (−27.2)	−6.7 (−35.5)	−7.6 (−36.7)	−0.7 (−29.8)
I10	4.8 (−30.9)	−3.2 (−37.6)	−3.6 (−39.5)	0.9 (−34.6)
TS10	10.4 (−20.1)	6.0 (−32.1)	1.5 (−29.9)	4.2 (−26.8)
I11	−9.4 (−43.8)	−13.3 (−48.2)	−16.8 (−51.8)	−11.5 (−45.4)

---

Comparison with the dippe-based catalyst and alkoxide trap. Given the promising catalyst activity exhibited by 1a, we compared its hydrogenation cycle with that of reference dippe-based catalyst. Using the same computational methodology, we calculated the hydrogenation profile of benzophenone catalysed by Mn(H)(dippe)(CO)₂ (see SI for details). The barrier for hydride migratory insertion is indeed found to be lower for dippe (1.1 kcal mol⁻¹) than for L1 catalyst (4.1 kcal mol⁻¹). Moreover, the TS for the alkoxide protonation step is somewhat lower in free energy in the case of dippe (−0.6 vs. 1.5 kcal mol⁻¹), as well as the relevant pentacoordinated alkoxide (−21.7 vs. −21.3 kcal mol⁻¹). Overall, the energetic span³⁸ for the cycle of the dippe catalyst amounts 21.1 kcal mol⁻¹, only 1.7 kcal mol⁻¹ lower than for L1 catalyst (22.8 kcal mol⁻¹). From this small difference, within the accuracy of the DFT method,³⁹ it seems reasonable not to expect a remarkable difference in catalyst activity between both catalysts.⁴⁰

Due to the strong metal–oxygen bonds, alkoxides can exhibit high stability, which may pose a severe drawback for catalysis. This has been related to donation from the lone pair on the alkoxide oxygen to a suitable unoccupied metal d orbital, resulting in a strong M–OR multiple bond. This effect is particularly critical in hydrogen borrowing reactions and the term alkoxide trap has been coined to describe it.⁴¹ To explore this effect in our system, we examined the metal–alkoxide interaction in I9c in more detail. This intermediate adopts a distorted trigonal bipyramidal geometry (Fig. 13), with the phosphite, alkoxide and one CO ligand occupying the equatorial plane (P(phosphite)–Mn–O and (CO)_eq–Mn–O angles of 129.4° and 135.1°, respectively), while the apical positions are occupied by the other CO and the phosphine ligands (P(phosphine)–Mn–(CO)_ax angle of 173.4°). Moreover, the Mn–O distance is significantly shorter than in the preceding agostic intermediate I8c (1.8436 Å vs. 1.9717 Å). A natural bonding orbital (NBO)⁴² analysis revealed a Wiberg bond order of 0.51 for I9c, notably higher than the 0.36 found for I8c. In addition, second-order perturbation theory analysis indicated considerable donor–acceptor interactions between the alkoxide oxygen lone pairs of I9c and antibonding orbitals on Mn–C (carbonyl) and Mn–P (phosphite) (Fig. 14), with stabilisation energies (ΔE²) of 25.6, 33.5 kcal mol⁻¹ for the Mn–C and 48.3 kcal mol⁻¹ for the Mn–P.⁴³ These findings agree with a particularly strong Mn–alkoxide interaction in I9c.

Fig. 13 Optimised geometries for alkoxide complexes I8c and I9c (all hydrogens except the alkoxide β one have been omitted for clarity).

Fig. 14 Donor acceptor interactions between alkoxide oxygen lone pairs and Mn–C (carbonyl) (a and b) and Mn–P (phosphite) (c) antibonding orbitals in complex I9c.

Conclusions

In this contribution, we have described the synthesis and characterisation of Mn(I) complexes of the type Mn(Br)(CO)₃(P-OP) (1), bearing phosphine–phospite (P-OP) ligands. These complexes exist in solution as mixtures of the fac and mer isomers, with the fac one being predominant. IR spectroscopy data reflect the lower donor ability of P-OP ligands compared to bis(trialkylphosphines), which are commonly used in previously reported non-bifunctional Mn catalysts. Complexes 1 are effective in the hydrogenation of a broad range of aldehydes and ketones under mild conditions. A combined spectroscopic and computational study support a catalyst activation mechanism involving nucleophilic attack of the alkoxide on a coordinated CO ligand, as well as a non-bifunctional inner sphere hydrogenation pathway.

A key aspect of the catalytic cycle is the formation of a remarkably stable pentacoordinated alkoxide intermediate. This intermediate determines the energetic span of the cycle, rather that barriers associated with the more demanding hydride migration or alkoxide protonation steps. This feature, inherent to catalysts operating by an inner sphere mechanism, seems critical in catalyst activity. These results are of practical significance, as they suggest that a wide variety of potential Mn catalysts precursors can readily be prepared in a single step from Mn(Br)(CO)₅ and commercially available bidentate phosphorous ligands.⁴⁴ This modularity is particularly appealing for catalyst screening in asymmetric hydrogenation,⁴⁵ an area where most applications to date have relied on specifically designed PNP and PNN tridentate ligands for bifunctional catalysts.⁴⁶ Further studies on the scope and enantioselective applications of the Mn–POP catalysts reported herein are currently underway and will be disclosed in due course.

Conflicts of interest

There are no conflicts to declare.

Data availability

Supplementary Information: experimental details of the synthesis and characterization of Mn complexes and procedures for catalytic hydrogenation reactions. NMR spectra of mechanistic experiments, X-ray crystallographic data for 1a, computational details and a coordinate file for structures calculated. See DOI: https://doi.org/10.1039/D5CY00906E.

CCDC 2468788 (1a) contains the supplementary crystallographic data for this paper.⁴⁷

The data supporting this article have been included as part of the supplementary information (SI).

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We thank Ministerio de Ciencia, Innovación y Universidades from Spain (Grant PID2022-139782NB-I00/AEI/10.13039/501100011033/FEDER, UE) for financial support. The use of computational resources of the Centro Informático Científico de Andalucía (CICA, cluster Hércules) and the Galician Supercomputing Centre (CESGA) are also acknowledged.

Notes and references

1. (a) P. Chirik and R. Morris, Acc. Chem. Res., 2015, 48, 2495; (b) Y. Wang, L. Zhu, Z. Shao, G. Li, Y. Lan and Q. Liu, J. Am. Chem. Soc., 2019, 141, 17337–17349; (c) H. Zhao, A. K. Ravn, M. C. Haibach, K. M. Engle and C. C. C. J. Seechurn, ACS Catal., 2024, 14, 9708–9733; (d) M. Beller, Chem. Rev., 2019, 119, 2089.
2. (a) G. A. Filonenko, R. Van Putten, E. J. M. Hensen and E. A. Pidko, Chem. Soc. Rev., 2018, 47, 1459–1483; (b) G. A. Filonenko, R. Van Putten, E. J. M. Hensen and E. A. Pidko, Chem. Soc. Rev., 2018, 47, 1459–1483; (c) Y. Wang, M. Wang, Y. Li and Q. Liu, Chem, 2021, 7, 1180–1223; (d) K. Das, S. Waiba, A. Jana and B. Maji, Chem. Soc. Rev., 2022, 51, 4386–4464; (e) K. Das, S. Waiba, A. Jana and B. Maji, Chem. Soc. Rev., 2022, 51, 4386–4464.
3. (a) S. Elangovan, C. Topf, S. Fischer, H. Jiao, A. Spannenberg, W. Baumann, R. Ludwig, K. Junge and M. Beller, J. Am. Chem. Soc., 2016, 138, 8809–8814; (b) F. Kallmeier, T. Irrgang, T. Dietel and R. Kempe, Angew. Chem., Int. Ed., 2016, 55, 11806–11809.
4. For studies on bifunctional catalysts containing ligands with NH fragments, see: (a) R. van Putten, E. A. Uslamin, M. Garbe, C. Liu, A. Gonzalez-de-Castro, M. Lutz, K. Junge, E. J. M. Hensen, M. Beller, L. Lefort and E. A. Pidko, Angew. Chem., Int. Ed., 2017, 56, 7531–7534; (b) S. Kar, A. Goeppert, J. Kothandaraman and G. K. S. Prakash, ACS Catal., 2017, 7, 6347–6351; (c) D. Wei, A. Bruneau-Voisine, T. Chauvin, V. Dorcet, T. Roisnel, D. A. Valyaev, N. Lugan and J. B. Sortais, Adv. Synth. Catal., 2018, 360, 676–681; (d) H. Li, D. Wei, A. Bruneau-Voisine, M. Ducamp, M. Henrion, T. Roisnel, V. Dorcet, C. Darcel, J. F. Carpentier, J. F. Soulé and J. B. Sortais, Organometallics, 2018, 37, 1271–1279; (e) P. Ryabchuk, K. Stier, K. Junge, M. P. Checinski and M. Beller, J. Am. Chem. Soc., 2019, 141, 16923–16929; (f) Y. Wang, L. Zhu, Z. Shao, G. Li, Y. Lan and Q. Liu, J. Am. Chem. Soc., 2019, 141, 17337–17349; (g) M. B. Widegren and M. L. Clarke, Catal. Sci. Technol., 2019, 9, 6047–6058; (h) W. Yang, I. Y. Chernyshov, R. K. A. van Schendel, M. Weber, C. Müller, G. A. Filonenko and E. A. Pidko, Nat. Commun., 2021, 12, 1–8; (i) T. Vielhaber and C. Topf, Appl. Catal., A, 2021, 623, 118280; (j) S. S. Gholap, A. Al Dakhil, P. Chakraborty, H. Li, I. Dutta, P. K. Das and K.-W. Huang, Chem. Commun., 2021, 57, 11815–11818; (k) D. Wei, R. Sang, P. Sponholz, H. Junge and M. Beller, Nat. Energy, 2022, 7, 438–447; (l) V. Zubar, N. Lichtenberger, M. Schelwies, T. Oeser, A. S. K. Hashmi and T. Schaub, ChemCatChem, 2022, 14, e202101443; (m) Z. Wei, H. Li, Y. Wang and Q. Liu, Angew. Chem., Int. Ed., 2023, 62, e202301042.
5. For examples of catalysts based on a deprotonable CH ligand fragment see: (a) Y. Q. Zou, S. Chakraborty, A. Nerush, D. Oren, Y. Diskin-Posner, Y. Ben-David and D. Milstein, ACS Catal., 2018, 8, 8014–8019; (b) U. K. Das, A. Kumar, Y. Ben-David, M. A. Iron and D. Milstein, J. Am. Chem. Soc., 2019, 141, 12962–12966; (c) R. Buhaibeh, O. A. Filippov, A. Bruneau-Voisine, J. Willot, C. Duhayon, D. A. Valyaev, N. Lugan, Y. Canac and J. Sortais, Angew. Chem., Int. Ed., 2019, 58, 6727–6731; (d) S. M. W. Rahaman, D. K. Pandey, O. Rivada-Wheelaghan, A. Dubey, R. R. Fayzullin and J. R. Khusnutdinova, ChemCatChem, 2020, 12, 5912–5918; (e) E. S. Gulyaeva, R. Buhaibeh, M. Boundor, K. Azouzi, J. Willot, S. Bastin, C. Duhayon, N. Lugan, O. A. Filippov, J. Sortais, D. A. Valyaev and Y. Canac, Chem. – Eur. J., 2024, 30, e202304201.
6. (a) S. Weber, B. Stöger and K. Kirchner, Org. Lett., 2018, 20, 7212–7215; (b) S. Weber, L. F. Veiros and K. Kirchner, Adv. Synth. Catal., 2019, 361, 5412–5420; (c) S. Kostera, S. Weber, M. Peruzzini, L. F. Veiros, K. Kirchner and L. Gonsalvi, Organometallics, 2021, 40, 1213–1220; (d) S. Weber, J. Brünig, L. F. Veiros and K. Kirchner, Organometallics, 2021, 40, 1388–1394; (e) R. A. Farrar-Tobar, S. Weber, Z. Csendes, A. Ammaturo, S. Fleissner, H. Hoffmann, L. F. Veiros and K. Kirchner, ACS Catal., 2022, 12, 2253–2260.
7. J. A. Garduño and J. J. García, ACS Catal., 2019, 9, 392–401.
8. M. Pinto, S. Friães, F. Franco, J. Lloret-Fillol and B. Royo, ChemCatChem, 2018, 10, 2734–2740.
9. (a) K. Azouzi, L. Pedussaut, R. Pointis, A. Bonfiglio, R. Kumari Riddhi, C. Duhayon, S. Bastin and J.-B. Sortais, Organometallics, 2023, 42, 1832–1838; (b) K. Azouzi, R. Pointis, R. Buhaibeh, P. H. Fernández, L. Pedussaut, M. Boundor, A. Bonfiglio, A. Bruneau-Voisine, D. Wei, T. Roisnel, C. Duhayon, M. Á. Casado, D. A. Valyaev, Y. Canac, S. Bastin, C. Raynaud and J.-B. Sortais, J. Catal., 2024, 430, 115334.
10. N. F. Both, A. Spannenberg, H. Jiao, K. Junge and M. Beller, Angew. Chem., Int. Ed., 2023, 62, e202307987.
11. M. Hruzd, S. L. Kleynemeyer, C. Michon, S. Bastin, E. Pollet, V. Ritleng and J. B. Sortais, Chem. Commun., 2025, 2969–2972.
12. For reviews on this topic see: (a) P. W. N. M. van Leeuwen, P. C. J. Kamer, C. Claver, O. Pàmies and M. Diéguez, Chem. Rev., 2010, 111, 2077–2118; (b) J. F. Teichert and B. L. Feringa, Angew. Chem., Int. Ed., 2010, 49, 2486–2528; (c) D. S. Surry and S. L. Buchwald, Angew. Chem., Int. Ed., 2008, 47, 6338–6361; (d) A. L. Clevenger, R. M. Stolley, J. Aderibigbe and J. Louie, Chem. Rev., 2020, 120, 6124–6196.
13. See, for instance: (a) Z. L. Niemeyer, A. Milo, D. P. Hickey and M. S. Sigman, Nat. Chem., 2016, 8, 610–617; (b) J. Tian, G. Wang, Z. Qi and J. Ma, ACS Omega, 2020, 5, 21385–21391; (c) C. Li, R. Dickson, N. Rockstroh, J. Rabeah, A. Spannenberg, M. Bühl and E. Mejía, Catal. Sci. Technol., 2020, 10, 7745–7756.
14. (a) S. Vargas, M. Rubio, A. Suárez, D. del Río, E. Álvarez and A. Pizzano, Organometallics, 2006, 25, 961–973; (b) M. Á. Chávez, S. Vargas, A. Suárez, E. Álvarez and A. Pizzano, Adv. Synth. Catal., 2011, 353, 2775–2794; (c) M. Vaquero, S. Vargas, A. Suárez, S. E. García-Garrido, E. Álvarez, M. Mancera and A. Pizzano, Organometallics, 2012, 31, 3551–3564; (d) M. Vaquero, A. Suárez, S. Vargas, G. Bottari, E. Álvarez and A. Pizzano, Chem. – Eur. J., 2012, 18, 15586–15591; (e) P. Kleman, P. J. González-Liste, S. E. García-Garrido, V. Cadierno and A. Pizzano, ACS Catal., 2014, 4, 4398–4408; (f) P. J. González-Liste, F. León, I. Arribas, M. Rubio, S. E. García-Garrido, V. Cadierno and A. Pizzano, ACS Catal., 2016, 6, 3056–3060.
15. (a) K. Nozaki, N. Sakai, T. Nanno, T. Higashijima, S. Mano, T. Horiuchi and H. Takaya, J. Am. Chem. Soc., 1997, 119, 4413–4423; (b) F. Doro, J. N. H. Reek and P. W. N. M. van Leeuwen, Organometallics, 2010, 29, 4440–4447; (c) H. Fernández-Pérez, J. Benet-Buchholz and A. Vidal-Ferran, Chem. – Eur. J., 2014, 20, 15375–15384.
16. (a) M. Diéguez, S. Deerenberg, O. Pàmies, C. Claver, P. W. N. M. Van Leeuwen and P. Kamer, Tetrahedron: Asymmetry, 2000, 11, 3161–3166; (b) T. Robert, J. Velder and H.-G. Schmalz, Angew. Chem., Int. Ed., 2008, 47, 7718–7721; (c) W. Lölsberg, S. Ye and H.-G. Schmalz, Adv. Synth. Catal., 2010, 352, 2023–2031.
17. Q. Du, J. M. Neudörfl and H. G. Schmalz, Chem. – Eur. J., 2018, 24, 2379–2383.
18. (a) T. Horiuchi, E. Shirakawa, K. Nozaki and H. Takaya, Tetrahedron: Asymmetry, 1997, 8, 57–63; (b) A. Falk, A.-L. Göderz and H.-G. Schmalz, Angew. Chem., Int. Ed., 2013, 52, 1576–1580.
19. A. Pizzano, Chem. Rec., 2016, 16, 2595–2618.
20. (a) M. Rubio, S. Vargas, A. Suárez, E. Álvarez and A. Pizzano, Chem. – Eur. J., 2007, 13, 1821–1833; (b) F. León, A. Comas-Vives, E. Álvarez and A. Pizzano, Catal. Sci. Technol., 2021, 11, 2497–2511.
21. M. F. Brown, B. R. Cook and T. E. Sloan, Inorg. Chem., 1975, 14, 1273–1278.
22. A. Suarez, M. A. Mendez-Rojas and A. Pizzano, Organometallics, 2002, 21, 4611–4621.
23. (a) S. J. A. Pope and G. Reid, J. Chem. Soc., Dalton Trans., 1999, 1615–1621; (b) J. A. Garduño, M. Flores-Alamo and J. J. García, ChemCatChem, 2019, 11, 5330–5338.
24. B. S. Ault, T. M. Becker, G. Q. Li and M. Orchin, Spectrochim. Acta, Part A, 2004, 60, 2567–2572.
25. Calculations have been done using the ORCA program (version 5.0.3) using BP86 functional with def2-SVP basis set along with auxiliary bases def2/J, using CPCM model to include solvent effects (benzene for compounds 1 and 3 and toluene for catalytic studies) and dispersion correction (Grimme D3) in a first stage. Structures obtained were recalculated using B3PW91 functional and def2-TZVP basis set with def2/J auxiliary basis along with mentioned dispersion and solvent corrections. Molecular structures have been generated with Chimera software (see SI for further details).
26. To the best of our knowledge, precedents of these reactions are limited to hydrogenation of a couple of α,β-unsaturated aldehydes reported in ref. 6d.
27. S. Baldino, S. Facchetti, A. Zanotti-Gerosa, H. G. Nedden and W. Baratta, ChemCatChem, 2016, 8, 2279–2288.
28. W. Baratta, M. Ballico, G. Chelucci, K. Siega and P. Rigo, Angew. Chem., Int. Ed., 2008, 47, 4362–4365.
29. (a) C. Liu, R. van Putten, P. O. Kulyaev, G. A. Filonenko and E. A. Pidko, J. Catal., 2018, 363, 136–143; (b) R. van Putten, G. A. Filonenko, A. G. de Castro, C. Liu, M. Weber, C. Müller, L. Lefort and E. Pidko, Organometallics, 2019, 38, 3187–3196.
30. For examples on the formation of tert-butoxide-phosphine complexes upon reaction between halo precursors and KOtBu, see: (a) W. Baratta, M. Ballico, A. Del Zotto, E. Herdtweck, S. M. R. Peloso, K. Siega, M. Toniutti, E. Zangrando and P. Rigo, Organometallics, 2009, 28, 4421–4430; (b) E. Z. Jahromi and J. Gailer, Dalton Trans., 2010, 39, 329–336; (c) M. C. MacInnis, R. McDonald, M. J. Ferguson, S. Tobisch and L. Turculet, J. Am. Chem. Soc., 2011, 133, 13622–13633; (d) A. Adhikary, J. A. Krause and H. Guan, Organometallics, 2015, 34, 3603–3610.
31. S. Weber and K. Kirchner, Acc. Chem. Res., 2022, 55, 2562–2580.
32. For a study of the relative reactivity of fac/mer isomers in Mn(H)(dppm)(CO)₃ see: E. S. Osipova, E. S. Gulyaeva, N. V. Kireev, S. A. Kovalenko, C. Bijani, Y. Canac, D. A. Valyaev, O. A. Filippov, N. V. Belkova and E. S. Shubina, Chem. Commun., 2022, 5017–5020.
33. For a study of nucleophilic additions of alkoxides to metal carbonyls see: R. van Putten, G. A. Filonenko, A. M. Krieger, M. Lutz and E. A. Pidko, Organometallics, 2021, 40, 674–681.
34. Worth to note an analogous activation route from Mn(Br)(N,N′-dimethyl-1,2-cyclohexanediamine)(CO)₃ and KOiPr has been proposed: A. Hashemi, S. Bougueroua, M.-P. Gaigeot and E. A. Pidko, J. Chem. Theory Comput., 2022, 18, 7470–7482.
35. J. C. Powers, R. Seidner, T. G. Parsons and H. J. Berwin, J. Org. Chem., 1966, 31, 2623–2627.
36. Atropisomerization of 5a would lead to a diastereomeric hydride with the same configuration at the metal and R_ax configuration at the biaryl. Instead of the latter, 5b which is enantiomeric with atropisomerized 5a has been considered to keep the S_ax biaryl configuration fixed along the study .
37. S. Kozuch and S. Shaik, Acc. Chem. Res., 2011, 44, 101–110.
38. Calculated with the AUTOF software: A. Uhe, S. Kozuch and S. Shaik, J. Comput. Chem., 2011, 32, 978–985.
39. P. A. Dub, N. J. Henson, R. L. Martin and J. C. Gordon, J. Am. Chem. Soc., 2014, 136, 3505–3521.
40. Our standard conditions use a lower catalyst loading and milder reaction conditions than those reported for A (ref. 6a), although from results reported for B (ref. 6c) it may be inferred a higher activity for Mn(H)(dippe)(CO)₂. Unfortunately, we have not found data at lower catalyst loading for the dippe system to directly compare it with that of 1a.
41. (a) Z. Ye, Z. Yang, C. Yang, M. Huang, X. Xu and Z. Ke, Org. Chem. Front., 2022, 9, 4803–4817; (b) P. Su, Z. Chen, J. Ni, Z. Yang, Y. Li and Z. Ke, ACS Catal., 2023, 13, 12481–12493; (c) A. Vignesh, Z. Wang, Y. Liu and Z. Ke, ChemCatChem, 2024, 16, e202400821.
42. E. D. Glendening, J. K. Badenhoop, A. E. Reed, J. E. Carpenter, J. A. Bohmann, C. M. Morales, P. Karafiloglou, C. R. Landis and F. Weinhold, NBO 7.0., Theoretical Chemistry Institute, University of Wisconsin, Madison, 2018.
43. (a) S. D. Pike, F. M. Chadwick, N. H. Rees, M. P. Scott, A. S. Weller, T. Krämer and S. A. Macgregor, J. Am. Chem. Soc., 2015, 137, 820–833; (b) M. A. Sajjad, J. A. Harrison, A. J. Nielson and P. Schwerdtfeger, Organometallics, 2018, 37, 3659–3669; (c) M. Navarro, A. Toledo, S. Mallet-Ladeira, E. D. Sosa Carrizo, K. Miqueu and D. Bourissou, Chem. Sci., 2020, 11, 2750–2758.
44. M. Renom-Carrasco and L. Lefort, Chem. Soc. Rev., 2018, 47, 5038–5060.
45. (a) M. R. Friedfeld, H. Zhong, R. T. Ruck, M. Shevlin and P. J. Chirik, Science, 2018, 360, 888–893; (b) M. J. Lebowitz, C. S. MacNeil, L. N. Mendelsohn, M. Shevlin, M. V. Pecoraro, G. Hierlmeier and P. J. Chirik, ACS Catal., 2024, 14, 13260–13268.
46. (a) M. B. Widegren, G. J. Harkness, A. M. Z. Slawin, D. B. Cordes and M. L. Clarke, Angew. Chem., Int. Ed., 2017, 56, 5825–5828; (b) M. Garbe, K. Junge, S. Walker, Z. Wei, H. Jiao, A. Spannenberg, S. Bachmann, M. Scalone and M. Beller, Angew. Chem., Int. Ed., 2017, 56, 11237–11241; (c) L. Zhang, Y. Tang, Z. Han and K. Ding, Angew. Chem., Int. Ed., 2019, 58, 4973–4977; (d) F. Ling, H. Hou, J. Chen, S. Nian, X. Yi, Z. Wang, Di. Song and W. Zhong, Org. Lett., 2019, 21, 3937–3941; (e) L. Zhang, Z. Wang, Z. Han and K. Ding, Angew. Chem., Int. Ed., 2020, 59, 15565–15569; (f) L. Zeng, H. Yang, M. Zhao, J. Wen, J. H. R. Tucker and X. Zhang, ACS Catal., 2020, 10, 13794–13799; (g) C. Liu, M. Wang, S. Liu, Y. Wang, Y. Peng, Y. Lan and Q. Liu, Angew. Chem., Int. Ed., 2021, 60, 5108–5113; (h) C. S. G. Seo, B. T. H. Tsui, M. V. Gradiski, S. A. M. Smith and R. H. Morris, Catal. Sci. Technol., 2021, 11, 3153–3163.
47. CCDC 2468788: Experimental Crystal Structure Determination, 2025,  DOI:10.5517/ccdc.csd.cc2mdsmg.

---

Footnote

† These authors have contributed equally to this work.

---