Carbon‐Carbon Bond Formation from Carbon Monoxide and Hydride: The Role of Metal Formyl Intermediates

Abstract Current examples of carbon chain production from metal formyl intermediates with homogeneous metal complexes are described in this Minireview. Mechanistic aspects of these reactions as well as the challenges and opportunities in using this understanding to develop new reactions of CO and H2 are also discussed.


Introduction
Carbon oxides, such as carbon monoxide (CO) and carbon dioxide (CO 2 ), offer an intriguing potential to build carbon-carbon bonds directly with high atom efficiency.The ideology is inspired by nature, where enzymes are known to use CO 2 and H 2 O in the production of complex sugars during photosynthesis.Some have suggested that even under prebiotic conditions CO may be one of nature's building blocks for sugars. [1,2] hese "C 1 building blocks" are commonly discussed in terms of their toxicity or negative impact as greenhouse gases; their industrial value can be underappreciated.
CO is an abundant, cheap source of carbon and oxygen atoms.It can be readily obtained from fossil fuel sources including coal and natural gas.[5][6][7] Further, CO can be produced from CO 2 directly by the water-gas shift reaction [Eq.(1)]. [8]Key industrial processes use CO as a building block.For example, the Cativa process produces acetic acid by the iridium catalysed carbonylation of methanol at the 500 000ton scale annually. [9,10] imilarly, the hydroformylation "oxo" process produces aldehydes from reaction of alkenes, CO, and H 2 . [11]The Fischer-Tropsch (FT) process provides liquid hydrocarbons from syngas, mixtures of CO and H 2 , and can be considered as the controlled hydrogenation and homologation of CO.The FT process is typically operated when it is economically viable or necessary due to restricted access to raw materials. [12]For example, the Pearl Gas-to-Liquid Factory converts 150,00À barrels per day of natural gas to liquid hydrocarbons and lubricants. [13] þ The FT reaction consists of an initiation step, a chaingrowth step, and a termination step.The FT process operates under forcing conditions, commonly 473-623 K, 20-45 bar pressure. [13,14]  heterogeneous catalyst comprising a transition metal (M = Co, Fe), [15] a chemical promoter (K 2 O), and a chemically inert structural motif (e.g., SiO 2 , Al 2 O 3 , or MgO) is employed.[16][17][18] Incorporation of main group metals into the heterogeneous transition metal catalyst are known to improve activity and selectivity.[19] The primary reaction products are short to medium chain alkanes [Eq.( 2)] and alkenes [Eq.( 3)] formed alongside other hydrocarbon oxygenates.A Schultz-Flory distribution of hydrocarbons is obtained (C 10 -C 20 ) [20] resulting in the requirement for expensive separation processes and ultimately limiting the proliferation of this process.
The fundamental reactions governing carbon-carbon bond formation remain a challenge to elucidate, despite the FT process being known for over a century.In all cases, the reaction must proceed via C�O and HÀ H bond cleavage, subsequent CÀ C homologation, and eventual dehydration.Surface bound transition metal carbonyls have been proposed as reaction intermediates during the heterogeneous FT process. [17]Hydrogen activation at the metal surface is proposed to give reactive metal hydrides as the other key intermediate in FT reactions. [18]The combination of these two surface bound moieties underpins chain growth at the catalytic site, potentially via transient metalformyl intermediates.
Substantial effort to model the mechanistic processes of FT chemistry have been made using homogeneous metal complexes.These are far more amenable to study than heterogeneous systems, as the nature of the active site and reaction intermediates is readily elucidated by solution and solid-state characterisation methods. [21]While homogeneous systems are yet to prove highly effective as catalysts for FT, [8,21] their study allows elucidation of the fundamental steps for CÀ H and CÀ C bond formation from CO and H 2 (or hydride sites).In this mini-review, we summarise the current examples of carbon chain production from metal formyl intermediates with homogeneous metal complexes.To date, the formation of both unsaturated and saturated C 2 motifs along with cyclic C 3 products has been reported.We discuss the mechanistic aspects of these reactions and the challenges and opportunities in using this understanding to underpin future FT chemistry.

State-of-the-art 2.1. Structural Motifs from CO and H À
The simplest way to combine CO and H À on a metal surface results in the formation of a formyl ligand (CHO À ).
Two related mechanisms can be proposed for formyl generation: an intermolecular insertion of CO into a polarised metal-hydride bond, or an intramolecular migratory insertion of CO into a metal-hydride bond at the same metal centre.These mechanisms are closely related, differing only by the formation of a stable metal carbonyl intermediate.Metal-formyl complexes have been proposed as intermediates in FT processes for both homo-and heterogeneous systems. [8,22] espite their importance, the detailed study of transition metal formyl complexes has been hampered by their low stability. [23]Early studies established that these species undergo facile α-elimination to form the corresponding metal-hydrido carbonyl complex.[26][27] The first structurally characterised formyl complex, [ReCp(CHO)(NO)(PPh 3 )], was reported by Gladysz and co-workers in 1979. [28]We [23] and others [29] have shown that the formyl ligand can be stabilised through resonance with the corresponding oxycarbene structure (Figure 1) or through coordination of the formyl oxygen atom to the metal centre. [22,30] spite their limited stability, metal formyl complexes have been invoked in the formation of several C 2 and C 3 products (Figure 1) through further reactions with CO and H 2 (or hydride sites).The most prevalent products are cisethenediolate complexes.These contain a {C 2 H 2 O 2 } 2À ligand whose coordination mode is dependent upon the nature of the metal centre and supporting ligand.k 1 , k 2 , and η 4 motifs are known (Figure 2).The related transethenediolate complex is markedly less common.Higher C 3 + homologues are rare, with a single example of a C 3 propanetriolate {C 3 H 3 O 3 } 3À ligand reported to date.The known structural motifs have been observed across a series of homogeneous models including transition metal, main group, lanthanide, and actinide complexes.Examples of deoxygenated products from these reactions are also limited; however, ethylene, allyloxide, ethylidene, squarate and hexolate production has been reported from reactions of metal-hydride complexes with CO, with formyl intermediates very likely in all cases.

Ethenediolate Formation
Several reaction mechanisms can be envisioned for ethenediolate formation from 2 CO + H 2 through metal formyl intermediates.Scheme 1 shows the three most commonly discussed routes for carbon chain growth following generation of a metal formyl, these are: Mechanism A) A multistep hydride insertion mechanism involving: (1) hyydrometallation of the formyl intermediate to give bridging oxymethylene intermediate; (2) migratory insertion of a second CO molecule into the MÀ C bond to give an acyl intermediate; (3) 1,2-hydride shift and rearrangement to give ethenediolate complex; [31] Mechanism B) direct dimerisation of formyl ligands, either on two distinct formylcomplexes or a metal bis-formyl; [32] Mechanism C) CO insertion into the metal-formyl ligand, generating a C 2ketene intermediate, followed by insertion of the ketene into a second equivalent of MÀ H to yield the ethenediolate ligand. [33]he first suggestion of metal-formyl complexes as intermediates in CO homologation relevant to the FT process was reported almost simultaneously by Casey and Bercaw in 1976. [34,35] asey and co-workers reported a metal-formyl complex formed by the reaction of a polarised metal-hydrogen bond with a metal carbonyl complex.Reaction of a series of Fe, Cr and W carbonyl complexes with trialkoxyborohydrides [HB(OR) 3 ] À furnished the corresponding metal formyl complexes, identified by characteristic low-field resonances in the 1 H and 13 C NMR spectra.Though no direct CÀ C bond formation was observed, the authors speculated on the role of formyl complexes as potential intermediates in the first step of catalytic CO homologation with H 2 . [35]At the same time, Bercaw and co-workers were investigating dinitrogen zirconium complexes, isolating [Zr(Cp*) 2 N 2 ] 2 (1).The N 2 ligand on 1 was readily displaced with CO or H 2 , acting as an entry point into species of relevance to FT catalysis.Reaction of the zirconium hydride [Zr(Cp*) 2 (H) 2 ] (2) with CO at À 80 °C reversibly formed [Zr(Cp*) 2 (H) 2 (CO)] (3).NMR analysis indicated a symmetric structure of 3, with the CO ligand occupying the central equatorial position  ).The authors suggested a mechanism for ethenediolate formation that proceeded via a transient formyl intermediate (4). 4 was proposed to form by insertion of CO into one of the cis ZrÀ H bonds and undergo a subsequent non-reversible reaction to form the ethenediolate 5 (Scheme 2). [34]The putative intermediate 4 was not observed spectroscopically or characterised.
Subsequent work detailed the reduction of zirconiumcoordinated CO ligands with 2 (Scheme 3). [31]Reaction of [Zr(Cp*) 2 (CO) 2 ] (6) with 2 under an atmosphere of H 2 yielded the cis-isomer of ethenediolate complex 8.It is noteworthy that the analogous reaction in the absence of H 2 gave a mixture of products.The reaction was predicted to proceed via a chelating formyl ligand bridging the two zirconium centres (7).A series of metal carbonyls (M = W, Cr, Mo, Nb) were shown to react analogously with 2, providing isolable bridging formyl complexes but no CÀ C coupled products. [36]n 1981, Marks and co-workers reported the first spectroscopic characterisation of a metal formyl complex as an intermediate in ethenediolate formation (Scheme 4). [37,38] eaction of thorium hydride complex [Th(Cp*) 2 (OR)(H)] (9 a-b) with CO at 25 °C yielded the cis-ethenediolate complex 11 a-b.The cis-geometry of the ethenediolate ligand was identified by NMR spectroscopy techniques and 13 C labelling studies.The rate of reaction   [38,39]

Mechanism A: Hydride Insertion
The most invoked mechanism for ethenediolate formation from formyl intermediates is through the hydride insertion Mechanism A. This mechanism was proposed by Bercaw in their original work but with limited experimental or computational support for the steps following generation of the formyl intermediate. [31,40]  1991, Wolczanski and co-workers reported the reaction of CO with zirconium hydride complex [Zr(NHSi t Bu 3 ) 3 H] (13) to form an ethenediolate species (Scheme 5). [41]While in this case the formyl intermediate 14 could not be observed, its formation is underpinned by the discoveries with closely related Zr and Th metallocene systems described above. [31,37] evertheless, data to support the formation of two downstream intermediates in the hydride insertion mechanism was collected.Specifically, the oxymethylene Andersen and co-workers reported the reaction of monomeric [Ce(Cp') 2 H] (18; Cp' = 1,2,4-trimethylcyclopentadienyl) with CO yielded the corresponding cisethenediolate complex 21 (Scheme 6). [42] ΔG o 298K = À 13.9 kcal mol À 1 ).Onward reaction with an additional equivalent of 23 to give oxymethylene complex 24 was discussed, but further energy profile not calculated in this instance.
In 2008, Labinger, Bercaw, and co-workers reported the reductive coupling of CO mediated by a hydride source and a rhenium carbonyl complex with pendant Lewis acids (Scheme 7).Reaction of a rhenium phosphinoborane complex (25) with CO in the presence of one equivalent of NaHBEt 3 provided rhenium-formyl complex 26 in quantitative yield; 26 was characterised by a downfield singlet in the 1 H NMR spectra (δ H = 13.95 ppm) and single crystal Xray diffraction analysis.Addition of a second equivalent of NaHBEt 3 provided the rhenium cyclic acyl complex 27, wherein the acyl unit is stabilised by coordination of the two Lewis-acidic boryl groups.The structure of 27 was confirmed by NMR spectroscopy and single crystal X-ray diffraction.Rearrangement to the predicted ethenediolate was not observed in this instance (likely due to the stabilising Lewis acid sites), though the fundamental steps In 2015 Hill and co-workers [43] and Jones, Stasch, Maron and co-workers [44] simultaneously reported the reaction of a β-diketiminate stabilised magnesium hydride reagent with CO (Scheme 8).An intramolecular hydromagnesiation proceeds via a modest (ΔH � = 14.8 kcal mol À 1 ) energy barrier to give the bridged oxymethylene complex 33 (ΔH � = À 15.8 kcal mol À 1 ).A second CO insertion and subsequent 1,2-H shift yielded the identified ethenediolate complex through an overall exergonic reaction (ΔH°= À 82.9 kcal mol À 1 from free 28 and CO).

Angewandte
Chemie Minireviews spectra (δ H = 5.00 ppm, δ C = 134.9ppm).An X-ray diffraction analysis of 36 showed a new bonding motif for the ethenediolate ligand.36 is a dimeric calcium species in which two seven-coordinate calcium centres are bridged by a dianionic cis-ethenediolate ligand.The cis-ethenediolate ligand bonds to each calcium centre via symmetric bridging η 4 -interactions.The change in hapticity is proposed to occur because of the larger ionic radius of Ca 2 + compared to Mg 2 + (0.99 and 0.65 Å, respectively).
We recently reported reaction of magnesium hydride dimer 28 with a series of metal carbonyl complexes (M = Cr, Mn, Fe, Co, Rh, W, Ir). [23]In all cases, transition metal formyl complexes were obtained with the formyl ligand trapped as part of a chelating structure.Eight complexes were crystallographically characterised, giving the first series of well-defined metal-formyl complexes.Solution NMR spectroscopy identified the formyl unit (δ H = 13.05-15.11ppm; δ C = = 240-310 ppm), confirmed by HSQC spectroscopy.In most cases, the CÀ H formyl stretch could be observed in the IR spectra (2546-2635 cm À 1 ).
Onwards reaction of chromium formyl complex 49 results in CÀ C bond formation and formation an isolable ethenediolate species 50 (Scheme 10).The ethenediolate ligand in 50 bridges chromium and magnesium centres, binding to chromium through a η 4 -interaction, reminiscent of the bonding in calcium complex 36.DFT calculations were performed on a model chromium anion system.The calculated mechanism is consistent with CÀ C bond formation occurring by stepwise process by Mechanism A. Based on the calculation, either step (1) or ( 3) could be rate determining, proceeding via modest energy transition states: ΔG � 298K = 19.7 and 19.0 kcal mol À 1 respectively.The modest barrier for transformation of the formyl complex 52 to oxymethylene complex 53 is consistent with the isolable nature of the chromium formyl.CÀ C bond formation in step ( 2) occurred via a low transition state (ΔG � 298K = 0.7 kcal mol À 1 ) to yield 54. The overall ethenediolate formation from the chromium formyl anion 51 and magnesium hydride monomer 28 was exergonic (ΔG°2 98K) = À 37.1 kcal mol À 1 ).

Mechanism B: Formyl Dimerisation
In 2020, Okuda and co-workers reported reaction of a cationic calcium hydride supported by an N,N,N,N-type macrocycle, [{Ca(μ-H)(Me 4 TACD)} 2 (THF)][BAr 4 ] 2 (55)  (Me 4 TACD = 1,4,7,10-tetramethyl-1,4,7,10-tetraazacyclododecane; BAr 4 = B(C 6 H 3 -3,5-Me 2 ) 4 ) with CO (Scheme 11). [32]fter 5 minutes at 25 °C in THF solution, full consumption of 55 was observed spectroscopically and formation of cis- 56 was fully characterised and X-ray diffraction analysis showed η 4 -binding of the bridging ethenediolate ligand to each calcium centre.In contrast to the examples discussed above, DFT calculations support a formyl dimerisation mechanism for formation of ethenediolate 59, as outlined in Mechanism B. Dissociation of dimeric calcium hydride 55 and CO insertion into CaÀ H bond is suggested to lead to two equiv. of a mononuclear calcium formyl intermediates 57.Direct coupling of these intermediates by a formyl dimersation is calculated to occur through a modest energy barrier (ΔH � = + 20.2 kcal mol À 1 ) to provide 58 and ultimately 59 in an overall exergonic process (ΔH°= À 100 kcal mol À 1 ).Direct imerization of two anionic formyl ligands is likely stabilised in this instance by the highly Scheme 9. Reaction of β-diketiminate stabilised calcium hydride dimer 35 with CO.DFT calculated pathway (grey) supports formyl, oxymethylene, and acyl intermediates for Group 2 metals Ca, Sr, and Ba.Ar = 2,6-diisopropylphenyl. [45]gewandte Chemie Minireviews electropositive calcium centre, potentially facilitating this unusual reaction pathway.A related pathway has been proposed for alkynide dimerisation at a calcium centre. [47]he dimerisation of formyl intermediates has been investigated computationally for other systems.Mechanisms toward magnesium ethenediolate 31 through bis-(formyl) intermediates were explored computationally (Scheme 12a). [44]Addition of CO to observed bridging formyl hydride complex 30 to give either cis-(60) or trans-( 61) bis(formyl) intermediates were explored computationally.In both cases, the authors note no route to the observed ethenediolate 31 and higher energy barriers compared to the hydride insertion mechanism discussed in Scheme 9.In our study on ethenediolate formation at Cr-Mg complexes, DFT calculations suggest that direct dimerisation of a putative bis(formyl) intermediate ( 62) is a high energy process (ΔG°> 50 kcal mol À 1 ), strongly suggesting that this mechanism is not in operation (Scheme 12b). [23]

Mechanism C: Ketene Generation
In 1984, Bercaw and co-workers reported reaction of 6, 2, and H 2 to give the cis-ethenediolate product 8 (Scheme 3). [33]In efforts to stabilise a proposed formyl intermediate, analogous complexes with less sterically This seminal work is an exceedingly rare example in which formyl and ketene intermediates are identifiable and characterised during carbon chain formation from CO and hydride sites and the fundamental steps are related to Mechanism C for ethenediolate formation. [48]

cis-to trans-Isomerisation of an Ethenediolate
The first example of ethenediolate isomerisation was documented by Evans and co-workers in 1985 (Scheme 14). [49]Reaction of samarium hydride dimer

Minireviews
of several hours to days, with rates depending on solution concentration.A bridging acyl complex was identified as a likely intermediate during isomerisation, with the argument that the cis is the kinetic and the trans the thermodynamic product discussed.Both cis-and trans-isomers were fully characterised by multinuclear NMR and IR spectroscopy, and single crystal X-ray diffraction.Samarium formyl intermediate 72 was proposed as a likely intermediate in ethenediolate formation via a hydride insertion mechanism.
The cerium ethenediolate (21) was shown to nonreversibly convert to the more thermodynamically stable trans-isomer (76), over a period of seven months at 100 or two weeks at 190 °C under vacuum.Both isomers were fully characterised including by single crystal X-ray diffraction analysis. [42]

Ethanediolate Formation
In 1985, Bercaw and co-workers extended their studies to hafnium complexes.[Hf(Cp*) 2 (H) 2 ] (77) was prepared through reduction of the parent chloride complex [Hf-(Cp*) 2 Cl 2 ] with n-butyl lithium under an atmosphere of dihydrogen. [50]At À 41 °C, 77 reacts with CO to form the carbonyl dihydride [Hf(Cp*) 2 (CO)(H) 2 ] (78); as with 3, low temperature solution NMR and IR spectroscopy indicate a symmetric coordination of CO to Hf. Solutions of 78 warmed above À 10 °C yield a complex mixture of products, including: methoxide, cis-and trans-ethenediolate, and ethanediolate hafnium complex (80-83; Scheme 15).In contrast, reaction of carbonyl complex [Hf(Cp*) 2 (CO) 2 ] with metallocene hydrides 2 and 77 formed the expected cis-ethenediolate complexes only.The formation of the ethanediolate species as a minor component of these reactions was unexpected but provides an example of generation of a saturated C 2 product from CO and hydride sites.

Propanetriolate Formation
Formation of C 3 or higher products proposed to result from reactions of formyl intermediates with CO and H 2 (or hydrides) is exceptionally rare.Jones and co-workers have observed the formation of one such species on reaction of CO with a magnesium hydride complex. [44]The steric influence of the β-diketiminate ligands were shown to be crucial in determining the selectivity of the reaction.

Summary and Perspective
[53][54] Metal-formyl complexes have been identified as transient intermediates and mechanisms for ethenediolate formation proposed.Most data support Mechanism A described in Scheme 1 via oxymethylene and acyl intermediates.Despite these findings, examples of well-defined reactions that form C 3 + homologues are uncommon.Equally rare are examples of deoxygenation reactions that increase the ratio of C : O to greater than one.These steps, broadly the propagation sequence in FT chemistry, remain of significant interest and arguably the next focus point for the field.It is likely that the current approach of developing homogeneous models, will ultimately shed light on these remaining questions and some selected examples described briefly below support this assumption and provide motivation for further study.
Longer Carbon Chains: Reactions of metal-hydride complexes with CO to give higher (> C 3 ) coupled products have been achieved using multimetallic lanthanide and transition metal hydride complexes.C 4 squarate [55] and C 6 hexolate [56] ligands have been reported by reductive coupling of CO using titanium-and tantalum-hydride complexes respectively.Though in neither case were formyl or other intermediates identified, the fundamental CÀ C coupling steps likely proceed via formyl intermediates and subsequent loss of dihydrogen.These successes give precedent for future work in this field and point to multimetallic hydride complexes as attractive candidates for further CO homologation. [57,58] hydration/Deoxygenation: A modest number of metal-hydride complexes have been reported to react with CO to yield deoxygenated ligands, wherein the C : O ratio is greater than one.Ethylidene, [59] enolate, [60] allyloxide, [61] and squarate [55] ligands have been reported from reaction of multimetallic hydrides with CO.In the latter three cases, ethylene, [60] propene, [61] and γ-butyrolactone [55] were successfully liberated from the metal centre, respectively (Figure 3).In all these instances, cooperative multimetallic complexes were employed (M = Ta, Ln, Y, Ti).Many of the

Minireviews
known reactions occur at highly electropositive metals and are likely thermodynamically driven through the formation of strong MÀ O bonds.However there is extremely limited mechanistic evidence for how these dehydration or deoxygenation events occur.
In the longer term, understanding the fundamental steps for the coupling reactions of CO and H 2 (or hydride sites) at metal centres has the potential to lead to the rational design of homogeneous FT catalysts.Controlling the selectivity of the FT process, and removing the need for energy-intensive separations, remains challenging.Homogeneous systems have the potential to address this problem while providing access to a diverse range of carbon chain building blocks.

Figure 1 .
Figure 1.Canonical forms of a metal-formyl complex and reported structural motifs formed via metal-formyl intermediates.

Scheme 4 .
Scheme 2. Reductive homologation and hydrogenation of CO by a zirconocene hydride complex 2.
Conducting the reaction in n-pentane solution led to precipitation of the bridging oxymethylene intermediate [{Ce(Cp' 2 )} 2 -(μ-CH 2 O)] (20), isolated as orange crystals.20 was predicted to form from transient formyl intermediate 19.The bridging dimeric structure of 20 was confirmed by X-ray diffraction analysis.Onward reaction of 20 with CO gave ethenediolate 21, identifying it as an intermediate in CO homologation.Density Functional Theory (DFT) calculations, conducted by Eisenstein, were used to gain insight into the system.Although this study focused on onward reaction of 19 with H 2 to give the cerium methoxide complex 22, the initial steps of the calculated mechanism are of the most relevance to this discussion.18 was simplified to the cyclopentadienyl analogue [Ce(Cp) 2 H] (23) for computational cost.Reaction of 23 with CO was calculated to proceed via CO coordination to cerium and subsequent migratory insertion of CO into the CeÀ H bond.A modest energy barrier transition state (ΔG � 298K = 6.1 kcal mol À 1 ) was located for an overall exergonic process to yield the μ 2 -formyl complex [Ce(Cp) 2 (η 2 -CHO)] (24;
Reaction of [Mg{(μ-H){CH{C-(CH 3 )NDipp} 2 }] 2 (Dipp = 2,6-diisopropylphenyl) (28) with CO yielded the cis-ethenediolate complex [{Mg(CH{C-(CH 3 )NDipp} 2 )} 2 (μ-C 2 H 2 O 2 )] (31).31 adopts an unusual asymmetric bridging mode, wherein one of the Mg centres is bound through chelation to both ethenediolate oxygen atoms.The longer MgÀ O bonded oxygen is further bound to the other trigonal Mg centre.31 was characterised by single crystal X-ray diffraction.Running the reaction at À 40 °C provided spectroscopic evidence for the anticipated formyl intermediate.The structure of 30 was proposed based on NMR spectroscopic data: δ H = 14.08; δ C = 358.9ppm.DFT calculations support the generation of a formyl complex as an intermediate in CO homologation and follow the mechanism presented in Mechanism A. Insertion of CO into the MgÀ H bond of 28 suggested the formation of a transient formyl intermediate via a barrierless reaction.

Figure 3 .
Figure 3.Other products formed from reaction of metal-hydride complexes with CO.
Royal Society University Research Fellow, initially at UCL and now back at Imperial.He was appointed as a lecturer in 2011, Senior Lecturer in 2016, Reader in Organometallic Chemistry in 2019, and full Professor in 2021.