Nickel and Palladium Complexes with Reactive σ-Metal-Oxygen Covalent Bonds

1 Nickel and Palladium Complexes with Reactive s -Metal-Oxygen Covalent Bonds. Luis M. Martínez-Prieto[a] and Juan Cámpora.*,[b] Published in Israel Journal of Chemistry, https://onlinelibrary.wiley.com/doi/full/10.1002/ijch.202000001 [a] Instituto de Tecnología Química. CSIC – Universidad Politécnica de Valencia, Avda. Los Naranjos, S/N, 46022, Valencia, Spain. [b] Instituto de Investigaciones Químicas, CSIC - Universidad de Sevilla. C/ Américo Vespucio, 49. 41092, Seville, Spain. 2 Table of Contents: Nickel and palladium complexes with terminal hydroxide, alkoxide and other covalently-bound metal-oxygen bonds exhibit a combination of organometallic type reactivity with strong basicity and nucleophilicity. This article reviews results from our research group in the context of the general literature. Short Biographies Juan Cámpora is CSIC Research Professor at the Institute for Chemical Research (IIQ, CSIC-University of Seville) since 2012. He graduated with honors in 1986 and took his Ph. D. degree in Chemistry in 1990 under the supervision of Prof. E. Carmona. He was awarded a Fulbright grant for postdoctoral position with Prof. S. L. Buchwald at MIT, where he worked on the chemistry of titanium and zirconium benzyne and metallacyclic complexes (1990-1992). In 1995 he became permanent researcher at the newly created IIQ, where he is research group leader. He has authored over 100 publications, and supervised 15 Ph. D. thesis. His scientific interests focus on the organometallic chemistry of transition elements, in special, late transition metal alkyl, alkoxo and amido complexes and their applications in homogeneous catalysis. Luis Miguel Martínez-Prieto received his PhD degree in Chemistry from the University of Seville in 2012, working at the IIQ under the supervision of Profs. J. Cámpora and P. Palma on Ni and Pd pincer complexes with alkoxide ligands. He then moved with Prof. B. Chaudret at the “Laboratoire de Chimie de Coordination” (LCC; CNRS) in Toulouse, France for a postdoctoral stay. His research focused on the synthesis of metal nanoparticles (MNPs). In 2015 he joined the lab of Prof. P. van Leeuwen at the “Lab. de Physique et Chimie des Nano-Objets” (LPCNO; INSA), exploring the use of MNPs as catalysts. In 2017, he was awarded with a “Juan de la Cierva” fellowship, that allowed him to start an independent career working on confined metal catalysts at the “Instituto de Tecnología Química (ITQ)” (Valencia, Spain) in the team or Prof. A. Corma. O Ni Me P P iPr iPr iPr iPr R M P P OiPr iPr iPr iPr R M = Ni or Pd 3 Nickel and Palladium Complexes with Reactive s -Metal-Oxygen Covalent Bonds. † Luis M. Martínez-Prieto[a] and Juan Cámpora.*,[b] [a] Instituto de Tecnología Química. CSIC – Universidad Politécnica de Valencia, Avda. Los Naranjos, S/N, 46022, Valencia, Spain. [b] Instituto de Investigaciones Químicas, CSIC - Universidad de Sevilla. C/ Américo Vespucio, 49. 41092, Seville, Spain. Abstract. This article reviews the chemistry of nickel and palladium complexes with terminally bound hydroxide and alkoxide ligands. The research carried out in our group is discussed in the context of the general literature. It is shown that suitable methods of synthesis, combined with the choice of adequate ligands allow the isolation of a range of stable complexes. This has enabled a detailed investigation of the chemical reactivity of the M-O bonds, once believed to be intrinsically weak. The elucidation of trends in thermodynamic stability and kinetic lability is the key for a better understanding the reactivity of this clas of compounds, that combines typical organometallic patterns, like b -hydrogen elimination, with classic properties as bases and nucleophiles. Based on reversible acid-base exchange and CO2 insertion reactions, we discuss how the polarity of the M-OR bonds influence their relative stability, their hydrolytic sensitivity and their tendency to react with electrophiles. Keywords: Nickel; Palladium; O-Ligands; b-Hydrogen Elimination; CO2 fixation. Introduction According to the conventional definition, organometallic compounds are those that contain at least one covalent metal-carbon bond.[1] However, the boundaries of Organometallic Chemistry are fuzzy.[2] For example, hydride or dinitrogen complexes have always been regarded as pertaining to the organometallic domain, even if they are lacking any metal-carbon bonds. In contrast, many other molecular metal compounds, such as titanium(IV) alkoxides or amides, nickel(II) acetylacetonate or palladium(II) acetate, are regarded as inorganic † Dedicated to Prof. S. L. Buchwald in the 25th anniversary of the first report on the aryl amination reaction. 4 coordination compounds, even though the metal is covalently bound to essentially organic ligands. Although some oxygen donors, like the above-mentioned acetylacetonate, or carboxylates are frequently found in many organometallic compounds, their role is usually passive, as “ancillary” or stabilizing ligands. Over the past decades, many studies have shown that many typical organometallic reactions, including reactions like migratory insertion, bhydrogen elimination or reductive couplings, are not exclusive of M-C or M-H bonds, but can also be observed for a range M-Heteroatom bonds. These reactions of covalently bound heteroatom-based ligands mean a wide expansion of the Organometallic Chemistry panorama. Amongst the earliest and most conspicuous cases of "organometallic-type" reactivity stand late transition amides, hydroxides and alkoxides. This article focuses mainly on the reactivity of nickel and palladium complexes with terminal hydroxide and alkoxide ligands. The terminal coordination mode of the M-OR bonds in alkoxides is in many regards analogous to s-M-C bonds of classic organometallic complexes. Alkoxide and amide ligands abound in the chemistry of the early transition metals, either in simple (binary) derivatives or in complexes containing other ligands. When bonded to such metals, alkoxide and amide ligands are relatively unreactive and contribute to stabilize their complexes. However, on moving to the right along the d transition series, this class of ligands become less common.[3] The reason is that, as the metal ions becomes more electron-rich and less electropositive, the M-alkoxo or M- amido linkages become more reactive, to the point that such complexes become increasingly difficult to isolate and characterize. Although late transition metal hydroxide and alkoxide species have been long known to be intermediates in catalytic processes, e. g. palladium-catalyzed Wacker olefin oxidation,[4] or olefin alkoxycarbonylations,[5] well-characterized alkoxides of metals of groups 9 to 11 were chemical oddities until relatively recent time. Most often, the techniques and strategies successfully applied in the early days of organometallic chemistry to the synthesis of organometallic compounds failed to produce stable alkoxides or amido complexes, and when they did, their isolation was hampered by their high sensitivity to moisture. In addition, amides, alkoxides and even the more stable hydroxide complexes exhibit a pronounced tendency to expel other ligands, giving polynuclear species, or even polymeric insoluble solids. In 1988, Bryndza and Tamm reviewed the chemistry of monomeric late transition metal complexes with terminal hydroxide, alkoxide and amide ligands.[3a] Amongst these, just a few examples were true terminal alkoxides of the group 10 elements in their usual +2 oxidation state. The difficulties found in the synthesis of mononuclear late transition metal alkoxides and other related compounds, like amides, was formerly attributed to the inherent instability of the covalent M-O or M-N interactions. In a first approach, this proposal was justified in the context of Pearson’s hard and soft acids and bases (HSAB) theory,[6] as a consequence of the mismatch 5 between the “soft” metal acids and “hard” bases containing N and O donor atoms. In contrast, soft carbon anions would be a better match for late transition metal ions, leading to more stable M-C bonds. This essentially qualitative HSAB argumentation collides with the widespread occurrence of many oxygen-based ligands traditionally regarded as “hard” (e. g. carboxylates, acetylacetonates, etc) in the coordination chemistry of late transition metals. A more subtle argumentation resorted to the theory repulsive metal-ligand nonbonding p interactions.[7] According to this concept, whereas early transition metal alkoxides and amides are stabilized by p-donation from the electron pairs of the metal-bonded N or O atoms into the empty d orbitals of the metal, most d orbitals in late transition metals are filled, resulting in non-bonding, repulsive p-d orbital interactions. Such repulsions are alleviated if the lone pairs of the metal-bound heteroatom are delocalized by conjugation with ligand-based p* orbitals, as in carboxylates or acetylacetonates, or by push-pull interactions involving other ligands with p-acceptor capacity that may be present in the same complex, like CO.[7,8] Both lines of argumentation, HSAB and p p-d repulsions, have been used to justify the difficulty of preparing late transition metal complexes with strongly p-donor ligands, not only alkoxides or amides, but also fluoride, or oxo.[9] However, the relevance of p-d repulsions in the thermodynamic stability of late transition metal-heteroatom bonds is now considered less important.[10] More than three decades passed since Bryndza and Tamm’s review was published and, since that time, the chemistry of late transition metal complexes with alkoxo, amido and other hard, p-donor ligands has developed intensively.[11] The number of applications of the special “organometallic-like” reactivity of transition metal compounds with M-N or M-O bonds has experienced a continuous growth. Some prominent examples of these are the Buchwald- Hartwig aryl amination[12] (now celebrating its 25th anniversary), the Suzuki-Miyaura reaction,[13,14] reductive couplings of carbonyl and unsaturated compounds,[15] reduction of carbonyl compounds,[16] aerobic[17] or acceptorless[18] alcohol oxidation, hydrogen borrowing alkylation,[19] or various dehydrogenative coupling reactions,[20] just to cite a few of them. In parallel, an intense effort has been deployed to understand the fundamental aspects of reactive metal-heteroatom bonds, either as catalytic intermediates or as stoichiometric reagents. Research in this field has been one of the interests in our group for many years, focusing mainly on the chemistry of Ni(II) and Pd(II) complexes containing reactive metal-heteroatom bonds. Although not exclusively, most of our work has concentrated in the use of the chelating diphosphine 1,2-bis(diisopropylphosphino)ethane, and the iPrPCP pincer ligand, shown in Figure 1. Both systems have in common the strong PiPr2 donors, and the joint presence of M-C and M- O bonds, confined into cis and trans geometries, respectively. The recent interest on the substitutive uses of abundant metals from the first transition series instead of precious or scarce elements, has put additional interest on the comparison of the analogous chemistries of Ni and 6 Pd. These considerations moved us to present an account of our investigations on the topic of Ni and Pd complexes containing reactive metal-oxygen bonds. Figure 1. Basic structural types of nickel and palladium complexes with reactive M-O bonds studied in our laboratories. Syntheses of nickel(II) and palladium(II) hydroxide and alkoxide complexes. Out of the group 10 triad, Ni, Pd and Pt, only Ni gives rise to a simple hydroxide and alkoxide derivatives, M(OR)2. Nickel(II) hydroxide is a paramagnetic (high spin), polymeric solid formed by two dimensional layers of octahedra sharing their edges (CdI2-type structure).[21] At the turn of the past century, the literature just mentioned a few nickel alkoxides, obtained by metathetical exchange between nickel halides and alkaline alcoholates. These are insoluble solids of unknown structure,[22] although a number of Ni and Pd hydroxide and aryloxide complexes stabilized by ancillary ligands were already known (see below). At the onset of our exploration of the chemistry of reactive alkoxides of Ni and Pd, one of our interests was to establish whether discrete species containing -OR ligands could be stabilized using bulky enough R groups. Therefore we investigated the reaction of [NiBr2(DME)] (DME = 1,2-dimethoxyethane) and [PdCl2(MeCN)2)] with sodium salts of phenols that had bulky substituents R in their ortho positions Scheme 1).[23] For R = tBu, diamagnetic aryloxides M(OAr)2 1a and 1b were obtained in moderate yields, as deep -colored crystalline solids, respectively, soluble in hydrocarbon solvents. They have unusual p-sandwich structures, with the ligands interacting in h3 mode (pallyl-type) fashion through C2, C3 and C4, but not with the oxygen atom, therefore these compounds cannot be considered true aryloxides. However, solvated intermediates with a stype structure were also isolated.[24] By lowering the steric hindrance of the ortho R groups, normal oxygen coordination becomes prevalent. Thus, for R = iPr a black paramagnetic nickel compound 2 was obtained. Its X-ray diffraction structure reveals a linear array of three Ni atoms linked by four µ-OAr interactions, and two terminal terminal OAr ligands in the Ni atoms of the sides. These atoms are tricoordinated, a highly unusual environment for Ni(II). Complex 2 is a O Ni R’ P P iPr iPr iPr iPr RM P P OiPriPr iPriPr R M = Ni or Pd Chelating Diphosphine (DiPPE) Pincer (iPrPCP) 7 true molecular Ni(II) alkoxide, and remains hitherto the only example of this class of compounds in the group 10 triad, although several monomeric (linear) Ni(II) amides were reported later.[25] Scheme 1. Synthesis of binary Ni(II) and Pd(II) aryloxides with discrete molecular structures. The unusual p-allyl-type coordination found in complexes 1 led us to investigate the synthesis of similar complexes containing a plain h3-C3H5 fragment as an ancillary ligand, hence we explored the reaction of the dimeric allyl complexes [M(h3-C3H5)(µ-X)]2 (M = Ni or Pd; X = Br or Cl) with sodium phenoxides containing differently sized aryloxide anions.[26] For bulky aryloxides (ortho R = tBu), signs of destabilization of the M-O bond were observed again. As shown in Scheme 2, the outcome of such reactions is highly dependent on the metal and the metal/aryloxide ratio. For a metal/NaOAr ratio = 1, and for M = Pd, the “normal” s-aryloxide- bridged dimer 3 was produced, but this is thermally unstable and decomposes in solution giving free ArOH and black palladium. For M = Ni, a monomeric p-aryloxide complex [Ni((h3-C3H5)(h5- OAr)], 4 was obtained. Increasing the Ni/NaOAr ratio led to the “ate” allyl-aryloxide complex 5, solvated with a single molecule of THF. When this complex is mixed with an equimolar amount of 4, the p-aryloxide fragment displaces the solvent to yield 6a, another Ni-Na bimetallic complex with a Na atom linking s- and p-Ni aryloxide units. 6a, and its palladium analogue 6b can be directly obtained by carefully adjusting the M:NaOAr ratio and the reaction time. Shortly after our reporting of these complexes, A. Klein and A. Dogan reported the crystal structure of a cationic palladium p-aryloxide, [Pd(h2:h2-COD)( h5-OC6H2Me3)]+.[27] [MX2L2] R R NaO - 2 NaX M = Ni, X = Br, L = DME M = Pd, X = Cl, L = MeCN solvated aryloxide - S Slow S = solvent M = Ni, n = 1 (1a) R R M(O )2S2] [ R R M(O )2]n [ R = t-Bu M = Pd, n = 1 (1b) M = Ni, n = 3 (2)R = i-Pr O RR O R R M 1a (Ni); 1b (Pd) R R O R R O O RR R R O O RR R R ONi Ni Ni R = i-Pr R = t-Bu 2 8 Scheme 2. Synteses of allyl-aryloxide complexes of Ni(II) and Pd(II). Simultaneously with these studies, we also investigated the syntheses of nickel complexes with terminal alkoxide, hydroxide or other potentially reactive s-bound fragments, stabilized with suitable ancillary ligands. As mentioned above, a small number of molecular group 10 hydroxide and alkoxide complexes were known at this time. Most of them were Pd or Pt complexes, but very little was known about similar Ni derivatives. Scheme 3 summarizes some of the precedents on molecular Ni(II) hydroxides and alkoxides. The pioneering work of H. F. Klein with methylnickel complexes stabilized with monodentate PMe3 ligands revealed that square-planar complexes of the type [Ni(Me)(X)(PMe3)2] are stable if X = Cl, OCOR, or OAr, but not for X = OH or OR.[28] The latter compounds spontaneously loss one PMe3 unit to afford dimers with µ-OH or µ-OR bridging ligands, because such interactions allow dissipating the excess of electron density localized at the oxygen atom. The trinuclear hydroxide 7 is another example of the tendency of basic OH ligands to share its electron density through bridging interactions is.[29] The discovery of this complex in E. Carmona’s group in the early 1980’s led the Sevilla team to investigate the chemistry of nickel hydroxide complexes [Ni(R)(µ-OH)(PMe3)]2 (8),[29b] a work that, years later, inspired us to pursue the investigation of monomeric hydroxides and alkoxides of nickel and palladium. Similar arylpalladium hydroxides [Pd(R)(µ-OH)(PPh3)]2 were reported by H. Alper and V. V. Grushin in 1993.[30] X M X M t-Bu t-Bu ONa Me M = Ni M = Pd M : NaOAr = 1:1 O Pd O Pd Ar Ar O Ni t-Bu t-Bu Me 4 NaOAr NaOAr = O M t-Bu t-Bu Me O M O Ar Ar Na M : NaOAr > 1:2 O Ni O Ar Ar Na 3 THF THF M = Ni M : NaOAr ≈ 1:5 56a, M = Ni 6b, M = Pd 4 M = Ni or Pd 9 Scheme 3. Precedents in the chemistry of Ni(II) alkoxide and hydroxide complexes. The tendency of basic, O-bound covalent ligands to bridge metal centers is very general and has deep effects in the properties of their complexes. For example, the bridging interaction partially cancels the reactivity on the O atom, but reversible M-OR dissociation can open a transient coordination site, providing low-energy pathways for processes such as b-hydrogen elimination,[31] reductive coupling[32] or migratory insertion.[33] In some other cases, the association of reactive species gives results in more complex behaviors. We observed one such situation when attempting to prepare nickel enolate complexes. A metal can bind either the deprotonated carbon atom (C-enolates) or the oxygen of the enolate functionality (O-enolates). Each of these modes has its own, characteristic reactivity, the O mode behaving as a nucleophilic alkoxide, and the C-enolate as a stabilized metal alkyl, but the difference is usually masked by the facile exchange of both coordination modes. Thus, we decided to synthesize a rigid nickelacyclic enolate, in order to prevent the otherwise facile isomerization of the C and O- coordination modes.[34] As shown in Scheme 4, we started from 2-acylaryl nickel complexes (acyl = acetyl or propionyl), and then carried out the deprotonation of the acyl group with potassium t-butoxide. Pre-activation of the acyl fragment by the Ni center, confirmed by X-ray structures, promotes selective formation of the O-bonded isomer, and the strong chelate DiPPE ensures the stability of the O-enolate complex, 9. However, PMe3 as ancillary ligand leads to the unusual binuclear complexes 10. The formation of the latter can be rationalized as an aldol addition of an enolate unit across the acyl fragment of a second molecule of complex. In addition, it requires a molecule of water, that ends up in the bridging hydroxyl and the extra proton bound to the organic ligand. The mechanism of this reaction is shown in Scheme 5.[35] When the substituent R on the enolate unit is Me, 9’ is fairly stable in solution, but it spontaneously evolves to the final product 10 if isolation is attempted. The extra molecule of water in 10 suggests that the product somehow arises from a hydrolysis reaction. However, the reaction is triggered by the loss of volatile PMe3, allowing the appearance of alkoxide bridge interactions, and water Me3P R [Ni(Me)2(PMe3)3] [Ni(Me)(Cl)(PMe3)2] ROH NaOR - CH4 -PMe3 - NaCl Me Ni PMe3 PMe3 OR Isolable for R = Aryl O Ni O Ni Me Me3PMe PMe3 R R stronly favored for R = Alkyl or H - PMe3 + PMe3 Ni Ni Ni H O O H R Me3P R R R = -CH2C6H4-o-Me (o-Xylyl) Carmona, µ3-OH (1985) Klein, µ2-OH and OR (1973) H O Ni O H Ni R Me3PR PMe3 7 Carmona, µ2-OH (1989) R = CH2EMe3 (E = C, Si); o-Xylyl, CH2CMe2Ph 8 16 Scheme 13. Reversible b-hydrogen pre-equilibrium in alkyl-methyl complexes. An important difference between the Pt and Ni-based processes shown in Schemes 11 and 12 is that, whereas the former involves different H/OMe and H/Et couplings, all decompositions in the Ni system entail reductive elimination of methane in the final step. In consequence, any important kinetic differences in the decomposition rates of the Ni complexes could be attributed to the b-hydrogen elimination step. Therefore, a comparative kinetic study of the decay rate of the alkoxide, amido and alkyl complexes 18, 20 and 22 was undertaken. As would be expected for an intramolecular mechanism like that shown in Scheme 13, the decomposition of alkyl 22 obeys first-order kinetics, characterized with a near-zero activation entropy (DS‡ ≈ 0). The amide 20 follows also first-order kinetics, but with a large and positive DS‡ . Therefore, both decompositions probably share essentially similar mechanisms, but these must have some significant difference. We believe that the positive activation entropy found for 20 might be the signature of pre-equilibrium step involving the opening of the diphosphine chelate ring, prior to the b-elimination and methane elimination.[67] The activation enthalpies (DH‡) for the decomposition of 20 and 22 are 48 and 23 Kcal·mol-1, respectively. Thus, discounted the entropy factor, the energy barrier for the decomposition of the amide is still higher for the amide than for the alkyl.[68] In contrast with 20 and 22, the decomposition of alkoxide 18 shows a surprising zero-order kinetic dependency on the starting complex. In addition to point out to a totally different mechanism, the zero-order kinetics is puzzling because apparently there are no other components in the chemical system to control the reaction, that advances at a constant rate. This implies that there must be some undetected substance to exert the control of the reaction. Very likely, this catalyst must be some coordinatively unsaturated nickel complex, that promotes b-hydrogen elimination in the manner of Mechanism C (Scheme 10), but the lack of kinetic dependency on the starting material 18 implies that the rate is determined in an irreversible step that involves exclusively the catalyst. We suggested that the catalyst could be monomeric Ni(I) alkoxide that, on decomposing by b-hydrogen elimination (the rate-determining step), transfers the hydride to 18, which then eliminates methane, as depicted in Scheme 14. P Ni P CH3 CH2CH2Ph P Ni P CH3 H CH2 Ph P Ni P CH3 CH3 Ph 22 22’ - CH4 P Ni P CH2 Ph iPr2 iPr2 iPr2 iPr2 iPr2 iPr2 17 Scheme 14. Proposed mechanism for the decomposition of Ni(II) alkoxides catalyzed by Ni(I) species. In line with the previously discussed investigations for Ir y Pt complexes, our conclusion is that the ease of b-hydrogen elimination methylNi(II) system decreases in the order alkyl > amide >> alkoxide. However, Fryzuk has reported a nickel system closely related to ours, [(tBuPPC)Ni- ECHRR’], that contains a cyclometallated diphosphine ligand (tBuPPC).[69] In this system, bhydrogen elimination for ECHRR’ = alkyl, amido or alkoxide ligands, is followed by irreversible H transfer to the metallacycle, to give Ni(0) olefin, imino and ketone complexes, respectively. Fryzuk reports that attempts to introduce the alkoxo moiety led directly to the Ni(0) aldehyde complex, suggesting that b-hydrogen from the alkoxide is very fast, whereas the b-elimination from an amido ligand was considerably slower. However, since no mechanistic studies were carried out, any conclusions should be taken with caution. We have also investigated the mechanism of the thermal decomposition of Ni and Pd methoxides with iPrPCP pincer ligands, [(iPrPCP)M-OMe] 14a and 14b (Scheme 15).[51] Both complexes decompose in solution, through processes that involve b hydrogen elimination and H transfer to the pincer ligand to yield different M(0) species. The decomposition of the Pd methoxide 14b is noticeable even at the room temperature, yielding formaldehyde and a binuclear Pd(0) complex, [(iPrPCHP)Pd]2, where iPrPCHP stands for the demetallated “pincer” diphosphine. The Ni methoxide 14a is stable in solution at the room temperature, but signs decomposition van be observed above 50 ºC. In this case, the process entails full dehydrogenation of methoxide ligands to yield polynuclear Ni(0) carbonyl complexes with bridiging iPrPCHP ligands, along with H2. Continuous monitoring of these decompositions in C6D6 allowed the detection of the Ni ad Pd hydrides, 23a and 23b in the initial stages of the reaction. This was the first time that hydrides arising from b-hydrogen elimination were identified in the decomposition of well-defined alkoxides of Ni or Pd. Tests with authentic samples of the hydrides showed that, whereas the Pd complex 23b decomposes yielding the same Pd(0) product as the methoxide, the Ni hydride 23a is thermally robust, and the hydride does not migrate to the pincer ligand spontaneously. As represented in Scheme 15, the irreversible evolution of 23a into a P Ni P O CH2Ph P Ni PO CHPh H P Ni P CH3 OCH2Ph P Ni P CH3 OCH2Ph H P Ni P O HPh CH4 catalyst irreversible rate determing step 18 (I) P = P(iPr)2 18 nickel(0) carbonyl is brought about by formaldehyde, with loss of H2. The identification of 23a and 23b enabled to us to gather experimental thermodynamic and kinetic data for the belimination step, and compare sith the results of DFT calculations.[51b] The parameters shown in Scheme 15 confirm that both for M = Ni and Pd, b-elimination is reversible and slightly endothermic. This process is thermodynamically and kinetically more favorable for Pd than for Ni, but the activation barriers are similar,the difference being due to entropy effects. Two main conclusions were drawn from our mechanistic work on the decomposition of the Ni and Pd alkoxides 14a and 14b. First, the Pd alkoxide decomposes more readily than its analogue of Ni; the second is that for both metals, b-hydrogen elimination is reversible, and the equilibrium is slightly disfavored. In order to advance, reversible b-H elimination has to be connected to the irreversible decomposition of the hydrides 23a and 23b. This is facile for Pd, but for Ni requires a further reaction with formaldehyde to advance. Thus, the different stability of the methoxides of Ni and Pd is more connected to the behavior of the hydrides 23 than to the b-hydrogen elimination step itself. It is the reactivity of the hydrides 23 which largely determines the fate of the initial alkoxides 14. This is probably more the general rule than an exception in the chemistry of late transition alkoxides, and helps explaining the critical role that the choice of reagents, solvents and, in general, the experimental conditions have for the success of their isolation as stable compounds. Scheme 15. Mechanism of decomposition and activation parameters for b-hydrogen elimination from methoxide in Ni and Pd pincer complexes. M P P O iPr iPr iPr iPr 14a, M = Ni 14b, M = Pd CH3M P P H iPr iPr iPr iPr β-elimination 23a, M = Ni 23b, M = Pd + CH2=O P P P P Ni Pd H OCH HOCH H H Ni P P -H2 M = Ni O HP P H Ni CO [Ni(CO)2(µ-PCHP)]2 [Ni4(µ-CO)5(µ-PCHP)2] + P P H Pd M = Pd CH2=O [Pd(µ-PCHP)]2 Experimental Thermodynamic Paremeters ΔH‡ ΔHoΔSoΔGo298K Keq (298 K) Ni Pd 13(1) 23(3) +5.9(1) 8 x 10-5 11(2) 26(6) +3.4(2) 2 x 10-3 Experim. vs. Calcd. Activation Parameters ΔS‡ΔG‡298K ΔG‡ (DFT) Ni Pd 23(2) -30(5) 27.4(2) 28.3 27(2) -7(6) 25.0(3) 26.4 ΔH, ΔG, Kcal·mol-1 ΔS, cal·mol-1·K-1 19 Basicity and hydrolytic sensitivity. One of the most characteristic features of late transition metal alkoxides and related compounds, is their facile hydrolysis, even with moisture traces. The hydrolytic sensitivity poses, in practice, one of the main difficulties for their isolation as pure species. The reaction of the monomeric methylnickel hydroxide, t-butoxide and pyrrolidinide with enolizable ketones shown in Scheme 16 provides a good example of this type of reactivity.[53c] The hydroxide complex reacts reversibly with the relatively acidic acetophenone (pKa ≈ 24. 7 in DMSO), giving the corresponding O-enolate, but fails to deprotonate the less acidic ketone pinacolone (pKa = 27.7 in DMSO). The t-butoxide complex reacts quantitatively with acetophenone. It also does with pinacolone, but the reaction is sluggish and slowly approaches an equilibrium situation. In contrast, the pyrrolidinide, an aliphatic amide, not only deprotonates rapidly and quantitatively both ketones, but also weaker carbon acids like ethyl acetate or acetonitrile (pKa = 30 - 31 in DMSO). It is worth mentioning that protonation of the methyl group was never observed, even though the methide anion is a much stronger base than amides or alkoxides. Scheme 16. Reaction of methylnickel hydroxide, alkoxide and amides with methylketones It is illustrative to compare in some more detail the basic reactivity of the hydroxide and the t-butoxide complexes of Scheme 16. The constants for the equilibria between the hydroxide and acetophenone, and the t-butoxide and pinacolone are 0.02 and ≈ 6, respectively. Considering that these values and the pKa data of both ketones in DMSO, it is possible to estimate the equilibrium constant for the reaction of the t-butoxide with acetophenone as ≈ 6 x103. This would be about five orders of magnitude higher than that the equilibrium constant for the hydroxide with the same ketone. This difference is much larger than one would expect from the basicities of the hydroxide and t-butoxide anions. Furthermore, in organic solvents, the hydroxide anion is actually a significantly stronger base than t-butoxide, as can be deduced from the pKa's of water P Ni P CH3 E +H3CPh O P Ni P CH3 O + H-E CH2 Ph EKeq OH 2 x 10-2 = tBuO ✔large NC4H4 ✔large P Ni P CH3 E H3CtBu O P Ni P CH3 O + H-E CH2 tBu OH – ✘ tBuO =~6 NC4H4 ✔large P Ni P CH3 N H3CZ P Ni P CH3 H2 CZ pKa = 24.7 pKa = 27.7 + + Z = CN , CO2Me pKa ≈ 30 +N HNC4H4 ✔large iPr2 iPr2 iPr2 iPr2 iPr2 iPr2 iPr2 iPr2 iPr2 iPr2 iPr2 iPr2 20 and t-butanol in DMSO, 31.2 and 29.4, respectively. Thus, the apparent basicities of the tbutoxide and hydroxide complexes would require some explanation. These are some: i) The tbutoxide complex is sterically destabilized as compared to the hydroxide, which renders the former more reactive; ii) the Ni-O bond in the hydroxide is significantly more stable than the tbutoxide; and iii) a combination of i) and ii). In order to provide a deeper insight in the acid-base relationships of alkoxide complexes of Ni and Pd, we examined water / alcohol proton exchange equilibria involving pincer hydroxo complexes 12a or 12b and several common alcohols, with not too different molecular sizes: methoxide, ethoxide, butoxide, 2-methoxyethoxide and ipropoxide, as shown in Scheme 17.[49] The water / alcohol exchange equilibrium constants (Kex) were determined in dry C6D6 by titration of the hydroxides 12a or 12b with the alcohols. As can be seen, the equilibria are strongly shifted to the side of the hydroxide. This is consistent with the high hydrolytic sensitivity of the alkoxides, whose tendency to react with traces of water represents one of the major difficulties for the isolation and handling of pure samples of this type of compounds. The equilibrium constants for Ni are about one order of magnitude smaller than those of Pd, implying that nickel alkoxides are more sensitive to hydrolysis than those of palladium. This may be counterintuitive, because Pd alkoxides tend to be less stable and more difficult to isolate. Scheme 17. Alcohol-water exchange equilibria involving Ni and Pd pincer complexes. At first glance, the trends shown in Scheme 17 are consistent with a "common sense" view. Since alcohols are usually seen as much weaker acids than water, they would not be expected to protonate a hydroxide to a large extent. As discussed in the preceding example, the difficulty with this reasoning is that H2O is actually a much weaker acid in nonprotic organic solvents than it is as bulk liquid water. According to the pKa’s listed in Scheme 17 water behaves in DMSO as an acid of comparable force or, in some case even weaker, than alcohols. Thus, it could be expected that the equilibria shown in Scheme 17 should be more shifted to the right than they appear in our experiments. However, such an interpretation misses the fact that we are not dealing with H+ transfers between a neutral alcohol and a free OH- anion (or, conversely, between water and OR-), but exchanges between electroneutral, covalent species. Therefore, in order to give a correct interpretation to the values of Kex, or, what is the same, the [(iPrPCP)M-OH] + ROH [(iPrPCP)M-OR] + H2O Kex ROH pKa, ROH (DMSO) Kex M = Ni M = Pd C6D6 H2O27.5 1 1 MeO(CH2)2OH 26.5 5.7 x 10-3 5.0 x 10-2 MeOH 27.9 1.6 x 10-2 4.5 x 10-2 EtOH 28.2 1.4 x 10-3 2.6 x 10-2 iPrOH 27.9 ~ 3 x 10-4 n. d. M = Ni, 12a M = Pd, 12b 21 corresponding DGoex, it is necessary to evaluate not only the proton transfer capacities of the alcohol and water, but also the strengths of the M-OH and M-OR bonds. A simple approach to analyze neutral exchanges like those of Scheme 17 is to compute the reaction energy as the algebraic sum of the bond energies (or Dbond) that are being formed and broken. In these exchanges, M-OH and RO-H bonds break to form new M-OH and H-OR bonds (Eq 1). In this expression it is assumed that the entropy variation for the exchange is negligible, therefore DGºex ≈ DHºex. Eq. 1 can be reorganized as shown in Eq. 2, and the differences within brackets can be regarded as "relative bond energies" (DD) in scales referred to DM-OH or DH-OH, these becoming the common reference values. As the data for H-O bonds can be obtained from the literature, or computed by theoretical methods, the experimental DGºex provide an estimate of the DDM-OR values, i. e. the gauge for M-OR bond strengths that was being sought. These DDM-OR are more susceptible of interpretation than the crude Kex constants. In a pioneering work by Bryndza and Bercaw,[70] this type of analyses was applied to exchanges between hydroxide or alkoxide complexes of Ru and Pt with various weak acids, HA (including not only alcohols). These authors found that the constants for such equilibria were close to unit, hence DGº ≈ 0. Thus, according to Eq 2, the relative metal-ligand bond energies (in this case, DDM-A) approximately equal to the corresponding values of DDH-A. The conclusion is that, in these particular systems, the metal fragment “senses” the effect of the in the same way the influence of R, or in a graphic image, that the Ru and Pt centers behave as "large hydrogens". However, later work by Andersen, Bergman, and Holland on similar protic exchange reactions on Cp-Ni anilido complexes with HA acids showed that the thermoneutral condition (DGº ≈ 0) is not a general feature of this type of exchange reactions.[71] The degree of departure from the thermoneutral condition depends largely on the metal fragment considered. A survey of constants for the exchange equilibria of various late transition metal hydroxide complexes with methanol gave values spanning over three orders of magnitude.49 Such deviations can be rationalized in terms of the different degree of polarization of the M-ligand bond and H-ligand bonds.[71,72] In Pauling’s classic description of the energies of polar covalent bonds, these have a significant covalent and ionic contributions, the latter a Coulombic attraction force between the charged fragments. The sum of the two components give rise to the overall bond energies. Accordingly, it has been shown that relative metal-ligand bond stabilities are satisfactorily fitted ΔGºex ≈ ΔHºex = DM-OR + DH-OH - DM-OH - DH-OR ΔGºex ≈ (DM-OR - DM-OH) - (DH-OR - DH-OH) = ΔDM-OR - ΔDH-OR (1) (2) 22 with Drago’s empirical E-C model that accounts for both the electrostatic and covalent contributions of the bond.71 The analyses of the spectroscopic and structural features on Ni and Pd hydroxides and alkoxides [(iPrPCP)M(OH)] and [(iPrPCP)M(OR)] confirm that the M-O bond is strongly polar, to the point that they can be regarded as partially ionic.49 Based on 13C NMR data for the [(iPrPCP)M] moiety, we deduced that the s-donor capacity of the hydroxide and alkoxide groups is very similar, hence that the covalent bond energy is relatively independent of R and nearly the same for hydroxides and alkoxides. Thus, the effect of the R groups of alkoxides on the Kex values is mostly due to electrostatic effects. Whereas the attractive force between the cationic metal fragment and the small and electronegative hydroxide ligand is particularly strong, the effect of large, polarizable R groups is to disperse the negative electric charge concentrated on the oxygen atom, reducing the magnitude of electrostatic attraction. The consequence is that, being the M-OR bonds are destabilized by R to a larger extent than the H-OR bonds, the difference DDM-OR-DDH-OR is always positive (DGºex > 0), and the water / alcohol exchanges are shifted to the side of the hydroxide. The effect is larger for the R groups with a larger capacity to disperse the charge (e. g., the iPr). This concept provides a rationale to the experimental order of the hydroxide-alcohol exchange constants, and also explains the fact that the effect of the R groups is more marked for Ni than Pd. This can be attributed to the smaller size of the Ni, and its slightly more electropositive character as compared to Pd. Nucleophilicity and insertion chemistry. In addition to their characteristic basicity of hydroxides and alkoxide complexes, the strong polarization of the metal-oxygen bond induces a significant nucleophilicity that results in a rich chemical reactivity. The role of bond polarization is especially evident in the case of enolate complexes. These can exist as either C-bound and O-bound isomers, each of which should exhibit a characteristic reactivity. The C-enolate should behave as alkyl alkyl, and the O-enolate is a special type of alkoxide. The effect of the coordination mode would be noticed if the chemical reactivity of the isomers could be probed independently, but the difference is usually masked due to the facile C/O exchange. As mentioned before, we have studied a series of nickel metallacyclic O-enolates, 9, stabilized with the chelating diphosphine DiPPE. Nickel O-enolates are competent nucleophiles that selectively react with aldehydes, enones, or CO2.[34,53b,73] As shown in Scheme 18, the cyclic structure of the complex restricts C/O isomerization. This takes place only when 9 is heated in solution above 50 ºC, with the thermodynamic isomer ratio becoming frozen when the solutions are rapidly cooled to the room temperature. When an aldehyde is added to a mixture of the O- and C- enolates, only the O-enolate isomer reacts with 23 aldehydes, leaving the C-isomer unreacted.[34] Due to the charge delocalization on the enolate functionality, the attack takes place invariably at the remote and less hindered CH2 site. In general, the nucleophilic reactivity of this type of enolate complexes can be assimilated to that of a conventional carbon nucleophile, hence it shall not be discussed here in detail. An exception is the reaction of the enolates with carbon monoxide, as it does take place at the metal site. CO becomes incorporated in ester-type products, because insertion comes immediately followed by reductive ring closure, as shown in Scheme 18. Scheme 18. Differential reactivity of isomeric O- and C-bound enolates with aldehydes Reductive elimination is a common outcome for the carbonylation of Ni or Pd complexes that simultaneously contain M-C and M-O bonds. For example, alkyl-aryloxides of the type [Ni(R)(OAr)( DiPPE)] react with CO to afford the corresponding esters RCOOAr and [Ni(CO)2(DiPPE)].[52] This sort of reactivity conceals the actual insertion site of CO, and prevents establishing whether it takes place into the Ni-O or the Ni-C bonds. CO insertion into Pd-OR bonds has been known for a long time to be involved in the catalytic carbonylations.[74,75] Stable Pt(II) alkoxycarbonyl complexes (LnM-CO-OR) arising from CO insertion into Pt-OR bonds were isolated by Bennett,[76] Michelin[77] and Bryndza[78] in the 1970’s and 80’s. Carbonylation of complexes of complexes containing both Pt-C and Pt-OR groups invariably yields the alkoxycarbonyl products arising from CO insertion into the Pt-O rather than the Pt-C bonds (Scheme 19). Stable Pd(II) alkoxycarbonyl complexes bearing inert C6F5 groups resist reductive elimination. These were obtained from the reaction of hydroxide complexes with CO in alcohol solvents.[79] Computational (DFT) analysis of the mechanism of CO insertion in simplified Group 10 models of the type [M(Me)(OMe)(PH3)2] (M = Ni, Pd, Pt) confirmed that CO insertion into the M-OMe bonds proceeds through the migratory mechanism.[80] Insertion in the M-OMe bond is thermodynamically favored over the M-CH3, although the balance in favor of the M-OMe insertion decreases in the order Pt (12 Kcal·mol-1) > Pd (7 Kcal·mol-1) > Ni (3 Kcal/mol). The Ni O P P iPr2 iPr2 CH2 Ni H2 C P P iPr2 iPr2 O Δ R-CHO room temp. No reaction Ni O P P iPr2 iPr2 CH OH R CO O O OH R [Ni(CO)2(DIPPE)] + 99’ 24 24 energy barriers for CO insertion, either in the M-C or M-O bonds are significant only for Pt, and much lower for Pd and Ni. For Pt and Pd, CO insertion into the M-O bonds is both thermodynamically and kinetically favored over M-C, but tor Ni, the kinetic preference is reversed. However, the size of the energy barriers for Ni is so low that the kinetic control probably would be overridden by the thermodynamic preference, were the insertion product stable enough to be detected. The preference for the carbonylation of the alkoxide linkage arises largely from the formation of the particularly stable C(O)-OMe bond, that is largely independent of the metal. The strength of the ester-type C-O bonds is due to the interaction of the oxygen lone pairs with the empty p* orbital of CO. This interaction is already present in the transition state of the migratory insertion (Figure 2), where it contributes to lower the energy barrier in comparison to the M-C insertion, that lacks this stabilizing factor. Independently of the participation of the oxygen electron pairs, the carbonylation mechanisms of M-OR bonds in square-planar d8 metal complexes resembles the situation for similar alkyls, clearly favoring an associative pathway with a 5-coordinate carbonyl intermediate for Ni, and 4-coordinate intermediates for Pd and Pt.[80] Further DFT studies on more realistic [Ni(Me)(OR)(HN=CHCH=NH)] and [Pt(Me)(OR)(PH2CH2CH2PH2)] as models confirmed the main predictions from the above study and showed the influence of the O-R group on the thermodynamic balance of the reaction, the carbonylation of the M-O bonds becoming thermodynamically disfavored when OR is an aryloxide group.[81] Scheme 19. Experimental and theoretical preference for CO insertion in group 10 alkyl-alkoxo complexes Figure 2. Transition states for the migratory insertion of CO into M-OMe and M-Me bonds. P Pt OMe P R CO P Pt COOMe P R stable L M OR’ L R CO L M COOR’ L R RCOOR’ + [M(CO)2L2] CO Thermodynamically M = Ni, Pd favored (DFT) MCO O Me MCO Me 25 The widespread application of pincer ligands in the 1990's opened new opportunities in the study of the fundamental reactivity of Ni-OR and Pd-OR bonds. When we synthesized the first monomeric Ni and Pd hydroxides with the iPrPCP ligand (12a and 12b), we were eager to investigate their reactivity towards small unsaturated molecules that typically undergo insertion reactions, CO in particular (Scheme 20).[40] We found that both complexes react rapidly with bubbling CO at the room conditions, but only the Pd gave a straightforward result. The product was isolated and structurally characterized as the binuclear complex 25b, that contains two [(iPrPCP)Pd] units linked by a bridging -C(O)O- ligand (another way to see this ligand is as [CO2]2- a doubly reduced CO2 molecule). The formation of this product can be rationalized as the result of the insertion of CO into the M-OH bond to afford a hydroxycarbonyl (M-COOH) species, 26. The latter behaves as an acidic "metalacarboxylic" acid, and reacts with one equivalent of the starting hydroxide to yield the binuclear product. The nickel analog, 25a, was obtained by careful reaction of 12a with the precise stoichiometric amount of CO (i. e., 0.5 equiv) in hexane, from which it precipitates. It was found that a larger amount of CO causes it to decompose into a mixture of unidentified products. More recently, Lee[82] and Zargarian[83] have reported that similar reactions of pincer-Ni hydroxide or siloxide complexes with CO lead to mixtures. Coming back to the reaction of the Pd hydroxide 12b with CO, full carbonylation of the Pd-OH bonds is enforced under 3 bar of CO. When this is carried out in a thick-walled NMR tube, the hydroxycarbonyl complex 26b was formed and was spectroscopically characterized in situ. However, upon CO removal 26b reverts to the binuclear µ-CO2 species 25b, which implies the reversibility of CO insertion. Scheme 20. Carbonylation of hydroxide pincer complexes 12a and 12b. The hydroxycarbonyl chemistry supported by the iPrPCP pincer contrasts with the straightforward results described by Wendt for the carbonylation of the closely related hydroxide M P P OH iPr iPr iPriPr M P P CiPr iPr iPriPr HO O M P P O iPr iPr iPr iPr C O M P P iPr iPr iPr iPr M= Ni, 12a M= Pd, 12b + CO - CO Detected for M = Pd (26b) M= Ni, 25a M= Pd, 25b -H2O +H2O 32 tailored to harness their reactivity, in a way somehow reminiscent of enzymes. However, an important advantage of small rigid ligands is their relative simplicity, accessible to detailed mechanistic and theoretical investigation. Research in this direction has highlighted the similarities and differences between discrete alkoxides and alkyls. Perhaps one of the most salient developments is the elucidation of the roles of thermodynamic and kinetic control on the chemical reactivity of alkoxides. Contrary to some early ideas that invoked HSAB theory concepts or repulsive nonbonding interactions between p-d filled orbitals, neither the available experimental evidence and computational analysis support the concept that late transition metaloxygen bonds are weak. Much of the high reactivity that in practice characterizes late transition metal alkoxides has a kinetic origin. The presence of reactive lone electron pairs localized directly on the metalbonded oxygen atom provide a highly reactive basic/nucleophile site that can be readily attacked by acids and electrophiles. An example is the extremely facile exchange with weak acids like water, which is the cause of the high hydrolytic sensitivity of Ni and Pd alkoxides. DFT calculatons have also demonstrated that these lone electron pairs assist migratory insertion of CO, and promote the insertion of CO2 by directly donating into the empty p* orbital of the electrophilic molecule before it interacts with the metal center. It is remarkable that quantitative measurements show that Ni complexes are thermodynamically more prone to hydrolysis than their Pd congeners, despite the nucleophilicity and lower thermal stability of the latter. This difference is a consequence of the highly polar nature of the M-O bonds. The ionic stabilization forces are stronger for Ni but, at the same time, this implies a larger bias in favor of hydroxides over alkoxides. Such polarity effects also control the selectivity other reactions under thermodynamic control, like CO2 insertion, appreciably more favorable for the M-OR than for the M-OH bond. An interesting consequence of this is that, whereas the M-OR complexes are very prone to hydrolysis, the trend is reverted in the presence of CO2 and alkylcarbonates M-OCOOR become preferred over bicarbonates M-OCOOH. Despite the progress in the understanding of the reactivity of late transition metal alkoxides and related complexes, there is much to be said still on the reactivity of the organometallic-like reactivity of metal-element bond. Thus, it is not clear yet to which extent the active role of the heteroatom lone electron pairs could be enhanced by their repulsive, nonbonding interactions with the filled d orbitals in the metal. Although such interactions play a quantitatively minor role on the stability of the M-O bonds, they certainly exist and manifest in spectroscopic and structural effects. Destabilization of the oxygen lone pairs may have a role increasing their nucleophilicity. Furthermore, it is uncertain whether the lower M-O bond strength predicted by DFT calculations for Pd as compared to Ni, has to do only with the decreased contribution of 33 ionic bonding, or there is also a reduction in the covalent contribution, as predicted by the HSAB theory. As mentioned in the Introduction, the combination of "organometallic-like" reactivity with exceptional basicity and nucleophilicty makes the chemistry of well-defined, discrete alkoxides of the late transition metals a fascinating subject. To this it should be added that the alcohols, phenols and their derivatives compounds are one of the most abundant substances in natural resources and their transformation is one of the keys in Organic Synthesis. There are many applications of palladium catalysts for the transformation of functional groups carbonheteroatom coupling (e. g., Buchwald-Hartwig amination, and its etherification counterpart) or alcohol oxidation that rely on the properties of Pd-heteroatom bonds. Extending some of with applications to Ni catalysts, or other first-row, non-precious metals will augment the versatility of the available reactions and contribute to reduce the costs associated with the use of scarce and essentially non-renewable metals. Acknowledgement. This work was supported by the Ministry of Science, Education and Universities of Spain (MCIU), the Spanish Research Agency (AEI), the European Union and the Polytechnic University of Valencia (UPV) through the Projects PGC2018-095768-B-100 (MCIUAEI-EU, granted to IIQ) and PAID-06-18 (UPV, granted to ITQ). LMMP gratefully thanks a “Juan de la Cierva” fellowship (IJCI-2016-27966). References [1] A. D. McNaught, A. Wilkinson, Compendium of Chemical Terminology, 2nd ed. (“the “Gold Book”), IUPAC, 1997. [2] A. Yamamoto, Organotransition Metal Chemistry, Fundamental Concepts and Applications, John Wiley & Sons, 1986. [3] For relevant reviews relative to complexes with polar covalent late transition metal-heteroatom bonds, see: (a) H. E. Bryndza, W. Tamm, Chem. Rev. 1988, 86, 1163 - 1188. (b) M. D. Fryzuk, C. G. Montgomery, Coord. Chem. Rev. 1989, 95, 1 - 40. (c) H. W. Roesky, S. Singh, K. K. M. Yisuff, J. A. Maguire, N. S. Hosmane, Chem. Rev. 2006, 106, 3813- 3846. (d) J. Webb, S. A. Burgess, T. R. Cundari, T. B. Gunnoe, Dalton Trans. 2013, 42, 16646-16665. (e) T. Corona, A. Company, Chem. Eur. J. 2016, 22, 13422-13429. [4] C. N. Cornell, M. S. Sigman, Inorg. Chem. 2007, 46, 1903-1909. [5] (a) P. W. N. M. van Leeuwen, M. A. Zuiveld, B. H. H. Swennenhuis, Z. Freixa, P. J. C. Kamer, K. Goubitz, J. Fraanje, M. Lutz, A. L. Spek, J. Am. Chem . Soc. 2003, 125, 5523-5539. (b) A. Brennführer, H. Neumann, M. Beller, ChemCatChem 2009, 1, 28-41. [6] J. E. Huheey, in Inorganic Chemistry, Principles of Structure and Reactivity, Harper & Row, 1975, pp. 225 - 235. [7] K. G. Caulton, New J. Chem. 1994, 18, 25-41. [8] J. T. Poulton, M. P. Sigalas, K. Folting, W. E. Streib, O. Eisenstein, K. G. Caulton, Inorg. Chem. 1994, 33, 1476 - 1485. [9] (a) V. V. Grushin, Acc. Chem. Res. 2010, 43, 160-171. (b) V. V. Grushin, Chem. Eur J. 2002, 8, 1007 - 1014. (c) J. M. Mayer, Comments Inorg. Chem. 1988, 8, 125-135. [10] (a) R. J. Fulton, A. W. Holland, D. J. Fox, D. G. Bergman, Acc. Chem. Res. 2002, 35, 44 - 56. (b) P. L. Holland, A. R. A., R. G. Bergman, Comm. Inorg. Chem. 1999, 21, 115-129. [11] Chapter 4 of Hartwig's t textbook Organotransition Metal Chemistry summarizes main aspects of transition metal alkoxides, amides and other metal complexes with direct metal-heteroelement s bonds, with a special focus on the late transition elements, and the final chapters collect many examples of their involvement in catalytic processes. J. F. Hartwig, Organotransition Metal Chemistry: From Bonding to Catalysis, University Science Books, Sausalito, California, USA, 2010. [12] P. Forero-Cortés, A. M. Haydl, Org. Process Res. Dev. 2019, 23, 1478-1483. [13] (a) S. Ostrowska, J. Lorkowski, C. Kubicki, ChemCatChem 2016, 8, 3580-3583. (b) C. Amatore, A. Jutand, G. Le Duc, Chem. Eur J. 2011, 17, 2492 - 2503. [14] (a) P.-A. Payard, l. A. Parego, I. Cirolini, L. Grimaud, ACS Catal. 2018, 8, 4812 - 4823. (b) A. H. Christian, P. Müller, S. Monfette, Organometallics 2014, 33, 2134-2137. 34 [15] (a) M. Holmes, L. A. Schwartz, M. J. Kirsche, Chem. Rev. 2018, 118, 6026-6052. (b) T. Kurahashi, S. Matsubara, Acc. Chem. Res. 2015, 48, 1703 - 1716. (c) Y. Hoshimoto, M. Ohashi, S. Ogoshi, Acc. Chem. Res. 2015, 48, 1746-1755. [16] S. Chakraborty, P. Bhattacarya, H. Dai, H. Guan, Acc. Chem. Res. 2015, 48, 1995-2003. [17] (a) D. Wang, A. B. Weinstein, P. B. White, S. S. Stahl, Chem. Rev. 2018, 118, 2636-2679. (b) C. Parmeggiani, F. Cardona, Green Chem. 2012, 14, 547-564. [18] R. H. Crabtree, Chem. Rev. 2017, 9228-9246. [19] B. G. Reed-Berendt, K. Polidano, L. C. Morrill, Org. Biomol. Chem. 2019, 17, 1595-1607. [20] (a) N. A. Eberhardt, N. P. N. Wellala, Y. Li, J. A. Krause, H. Guan, Organometallics 2019, 38, 1468-1478. (b) C. Gunanathan, D. Milstein, Science 2013, 341, 1229712-1229723. [21] A. F. Holleman, E. Wiberg, N. Wiberg, Holleman-Wiberg Inorganic Chemistry, 1st English Edition., Academic Press and de Gruyter, 2001. [22] (a) D. C. Bradley, R. C. Mehrotra, I. P. Rothwell, A. Singh, Alkoxo and Aryloxo Derivatives of Metals, Academic Press, 2001. (b) B. P. Baranwal, R. C. Mehrotra, Auts. J. Chem. 1980, 33, 37-43. (c) D. C. Bradely, J. Inorg. Nucl. Chem. 1972, 15, 259-232. [23] J. Cámpora, M. L. Reyes, K. Mereiter, Organometallics 2002, 21, 1014-1016. [24] The crystal structure of one such THF solvates was communicated later to the Cambridge Structural Database CSD Reference Code: HAYHEV; CSD deposition code 272530. [25] (a) M. L. Lipschutz, T. D. Tilley, Chem. Commun. 2012, 48, 7146-7148. (b) A. Bryan, W. A. Merrill, W. M. Reiff, J. C. Fettinger, P. P. Power, Inorg. Chem. 2012, 51, 3366-3373. (c) C. I. Lin, J. D. Guo, J. C. Fettinger, S. Nagase, F. Grandjean, G. L. Long, N. F. Chilton, P. P. Power, Inorg. Chem. 2013, 52, 13585-13593. [26] J. Cámpora, M. L. Reyes, T. Hackl, A. Monge, C. Ruiz, Organometallics 2000, 19, 2950-2952. [27] A. Klein, A. Dogan, Organometallics 2002, 21, 2340-2342. [28] (a) H. F. Klein, T. Wiemer, Inorg. Chim. Acta 1991, 189, 267-273. (b) H. F. Klein, H. H. Karsch, Chem. Ber. 1973, 106, 267-273. [29] (a) E. Carmona, J. M. Marín, P. Palma, M. Paneque, M. L. Poveda, Organometallics 1985, 4, 2053-2055. (b) E. Carmona, J. M. Marín, P. Palma, M. Paneque, M. L. Poveda, Inorg. Chem. 1989, 8, 1895-1900 [30] V. V. Grushin, H. Alper, Organometallics 1993, 12, 1890-1901. [31] D. M. Pearson, R. M. Waymouth, Organometallics 2009, 28, 3896-3900. [32] (a) J. P. Stambuli, Z. Weng, C. D. Incarvito, J. F. Hartwig, Angew. Chem. Int. Ed. 2007, 46, 2674-2677. (b) G. Mann, C. Incarvito, A. Rheingold, J. F. Hartwig, J. Am. Chem. Soc. 1999, 121, 3224-3225. [33] A few examples: (a) CO insertion: W. E. Lindsell, D. D. Palmer, P. N. Preston, G. M. Rosair, Organometallics 2005, 24, 1119-1133. (b) ethylene insertion: U. Klabunde, R. Mulhaupt, T. Herskowitz, A. H. Janowicz, J. Calabrese, S. D. Ittel, J. Polym. Sci. Ser. A. 1987, 25, 1989-2003. [34] a) J. Cámpora, C. M. Maya, P. Palma, E. Carmona, E. Gutiérrez-Puebla, C. Ruiz, J. Am. Chem. Soc. 2003, 125, 1482- 1483. b) J. Cámpora, C. M. Maya, P. Palma, E. Carmona, E. Gutiérrez, E. Ruiz, C. Graiff, A. Tiripicchio, Chem. Eur. J. 2004, 6889-6904. [35] J. Cámpora, I. Matas, C. M. Maya, P. Palma, E. Álvarez, Organometallics 2006, 25, 3124-3129. [36] In our original report we suggested that one of the enolate moieties would isomerize to C coordination in order to undergo attack by the second. In Scheme 5 we draw a simpler mechanism without such isomerization. [37] (a) D. Zargarian, A. Castonguay, D. M. Spasyuk, in Organometallic Pincer Chemistry (Eds.: G. van Koten, D. Milstein), Springer, Berlin-Heidelberg, 2013. (b) G. van Koten, in Organometallic Pincer Chemistry (Eds.: G. van Koten, D. Milstein), Springer, Berlin-Heidelberg, 2013. [38] P. Castan, F. Dahan, S. Wimmer, F. L. Wimmer, J. Chem. Soc. Dalton Trans. 1990, 2679-2683. [39] a) M. Akita, T. Miyaji, S. Hikichi, Y. Moro-Oka, Chem. Commun. 1998, 1005-1006. b) M. Akita, T. Miyaji, N. Muroga, C. Mock-Knoblauch, W. Adam, S. Hikichi, Y. Moro-Oka, Inorg. Chem. 2000, 39, 2096-2102. [40] J. Cámpora, P. Palma, D. del Rio, D. Álvarez, Organometallics 2004, 23, 1652-1655. [41] (a) I. Morales-Becerril, M. Flores-Álamo, A. Tlauhuext-Aca, A. Arévalo, J. J. García, Organometallics 2014, 33, 6796- 6802. (b) A. Sacco, P. Mastorilli, J. Chem. Soc. Dalton Trans. 1994, 2761-2764. [42] For alcohol-induced b-H elimination in alkoxides, see: (a) N. A. Smythe, K. A. Grice, B. S. Williams, K. I. Goldberg, Organometallics 2009, 28, 277-288. (b) C. M. Fafard, O. V. Ozerov, Inorg. Chimica Acta 2007, 286-292. (c) O. Blum, D. Milstein, J. Organomet. Chem. 2000, 593-594, 479-484. [43] S. Oh, S. Kim, D. Lee, J. Gwak, Y. Lee, Inorg. Chem., 2016, 55, 12863 - 12871. [44] B. Tran, M. Pink, D. J. Mindiola, Organometallics 2009, 28, 2234-2243. [45] O. Vechorkin, Z. Csok, R. Scopelliti, X. Hu, Chem. Eur. J. 2009, 15, 3889-3899. [46] D. del Rio, Ph. D. thesis, University of Seville 2003, page 279. [47] G. Mann, J. F. Hartwig, J. Am. Chem. Soc. 1996, 118, 13109-13110. [48] (a) L. Postigo, B. Royo, Adv. Synth. Catal. 2012, 2613-2618. (b) D. Powell-Jia, J. W. Ziller, A. G. DiPasquale, A. Rheingold, A. S. Borovik, Dalton Trans, 2009, 2986-2992. (c) C. M. Macaulay, T. Ogawa, R. McDonald, O. L. Sydora, M. Stradiotto, L. Turculet, Dalton Trans. 2019, 48, 9581-9587. [49] L. M. Martínez-Prieto, P. Palma, E. Álvarez, J. Cámpora, Inorg. Chem. 2017, 56, 13086-13099. [50] J. Cámpora, P. Palma, D. del Rio, M. M. Conejo, E. Álvarez, Organometallics 2004, 23, 5653-5655. [51] (a) C. Melero, L. M. Martínez-Prieto, D. del Rio, E. Álvarez., J. Cámpora, Chem. Commun. 2010, 46, 8851-8853. (b) L. M. Martínez-Prieto, E. Ávila, P. Palma, E. Álvarez, J. Cámpora, Chem. Eur. J., 2015, 21, 9833-9849. [52] J. Cámpora, J. A. López, C. Maya, P. Palma, E. Carmona, P. Valerga, J. Organomet. Chem. 2002, 643, 331-341. [53] (a) J. Cámpora, I. Matas, P. Palma, C. Graiff, A. Tiripicchio, Organometallics 2005, 24, 2827-2830. (b) J. Cámpora, I. Matas, P. Palma, E. Álvarez., C. Graiff, A. Tiripicchio, Organometallics 2007, 26, 3840-3849. (c) J. Cámpora, I. Matas, P. Palma, A. Tiripicchio, Organometallics 2007, 26, 5712-5721. [54] L. M. Martínez-Prieto, C. Melero, D. del Río, P. Palma, E. Álvarez, Organometallics 2012, 31, 1425-1438. [55] J. F. Hartwig, Inorg. Chem. 2007, 46, 1936-1947. [56] a) A. Behnia, P. D. Boyle, J. M. Blacquiere, R. J. Puddephatt, Organometallics 2016, 35, 2645 - 2654. a) N. M. Camasso, M. Pérez-Temprano, M. N. Sanford, J. Am. Chem . Soc. 2014, 136. c) K. Koo, G. L. Hillhouse, A. L. Rheingold, Organometallics 1995, 14, 456-460. [57] a) G. Mann, Q. Shelby, A. H. Roy, J. F. Hartwig, Organometallics 2003, 22, 2775-2789. b) R. A. Widenhoefer, S. L. Buchwald, J. Am. Chem . Soc. 1998, 120, 6504-6511. [58] For a leading reference, see: S. L. Buchwald, P. Ruiz-Castillo, H. Zhang, Org. Lett. 2018, 20, 1580-1583. [59] K. A. Bernard, W. M. Rees, J. D. Atwood, Organometallics 1986, 5, 390-391. 35 [60] J. Zhao, H. H., J. F. Hartwig, J. Am. Chem . Soc. 2001, 123, 7220-7227. [61] O. Blum, D. Milstein, J. Am. Chem. Soc. 1995, 117, 4582-4594. [62] J. C. M. Ritter, R. G. Bergman, J. Am. Chem . Soc. 1998, 6826-6827. [63] H. E. Bryndza, J. Calabrese, M. Marsi, C. Roe, W. Tam, J. E. Bercaw, J. Am. Chem . Soc. 1986, 108, 4805-4813. [64] F. Ozawa, T. Ito, A. Yamamoto, J. Am. Chem . Soc. 1980, 102, 6547-6463. [65] J. F. Hartwig, J. Am. Chem . Soc. 1996, 118, 7010-7011. [66] I. Matas, J. Cámpora, P. Palma, E. Álvarez, Organometallics 2009, 28, 6515-6523. [67] Ligand dissociation and formation of three-coordinate amido complexes is a known mechanistic feature in Pd(II)- catalyzed Buchwald-Hartwig amination reaction. In some cases these coordinatively unsaturated species are stable enough to allow isolation. See: a) M. S. Driver, J. F. Hartwig, J. Am. Chem. Soc. 1999, 119, 8232-8245. b) M. Yamashita, J. F. Hartwig, J. Am. Chem. Soc. 2004, 126, 5344-5345. [68] S. A. Macgregor, P. Vadivelu, Organometallics 2007, 26, 3651-3659. [69] A. C. Zimmerman, M. D. Fyzuk, Organometallics 2018, 37, 2305-2318. [70] H. E. Bryndza, L. K. Fong, R. A. Paciello, W. Tam, J. E. Bercaw, J. Am. Chem . Soc. 1987, 109, 1444-1456. [71] (a) P. L. Holland, R. A. Andersen, R. G. Bergman, J. Huang, S. P. Nolan, J. Am. Chem . Soc. 1997, 119, 12800-12814. (b) P. L. Holland, R. A. Andersen, R. G. Bergman, Comments Inorg. Chem. 1999, 1999, 115-129. [72] C. Bergquist, H. Storrie, L. Koutcher, B. M. Bridgewater, R. A. Friesner, G. Parkin, J. Am. Chem . Soc. 2000, 122, 12651-12658. [73] (a) J. Cámpora, C. M. Maya, P. Palma, E. Carmona, E. Gutiérrez-Puebla, C. Ruiz, C. Graiff, A. Tiripicchio, Chem. Commun. 2003, 1742-1743. (b) J. Cámpora, C. M. Maya, P. Palma, E. Carmona, E. Gutiérrez-Puebla, A. Monge, Chem. Eur J. 2007, 13, 3675-3687. [74] S. Huang, B. Yang, S. Wang, X. Ma, Chem. Soc. Rev. 2015, 44, 3079-3116. [75] (a) H. Yasuda, J.-C. Choi, S.-C. Lee, T. Sakakura, Organometallics 2002, 21, 1216-1220. (b) P. Giannocaro, N. Ravasio, M. Aresta, J. Organomet. Chem. 1993, 451, 243-248. (c) F. Rivetti, U. Romano, J. Organomet. Chem. 1979, 174, 221-226. (d) F. Rivetti, U. Romano, J. Organomet. Chem. 1978, 154, 323-326. [76] (a) M. A. Bennett, T. Yoshida, J. Am. Chem . Soc. 1979, 100, 1750-1759. (b) M. A. Bennett, A. Rokicki, Organometallics 1985, 4, 180-187. [77] R. A. Michelin, M. Napoli, R. Ros, J. Organomet. Chem. 1978, 175, 239-255. [78] H. E. Bryndza, Organometallics 1985, 4, 1686-1687. [79] J. Ruiz, M. T. Martínez, F. Florenciano, V. Rodríguez, G. López, J. Pérez, P. A. Chaloner, P. B. Hitchcock, Inorg. Chem. 2003, 42, 3650-3651. [80] S. A. Macgregor, G. W. Neave, Organometallics 2003, 22, 4547-4556. [81] S. A. Macgregor, G. W. Neave, Organometallics 2004, 23, 891-899. [82] C. Yoo, Y. Lee, Organometallics 2013, 32, 7195-7203. [83] J. Hao, B. Vabre, B. Mougang, D. Zargarian, Organometallics 2014, 30, 12544-12552. [84] a) R. Johannson, O. F. Wendt, Organometallics 2007, 26, 2426-2430. b) K. J. Jonasson, A. H. Mousa, O. F. Wendt, Polyhedron 2018, 143, 132-137. [85] A. Ambrosi, S. E. Denmark, Angew. Chem. Int. Ed. 2016, 55, 12164-12189. [86] C. Yoo, Y. Lee, Angew. Chem. Int. Ed. 2017, 56, 9502-9506. [87] Y. A. Daza, J. N. Kuhn, RSC Advances 2016, 6, 49675-49691. [88] M. Can, F. A. Armstrong, S. W. Ragsdale, Chem. Rev. 2014, 114, 4149-4174. [89] D. Huang, R. H. Holm, J. Am. Chem . Soc. 2010, 132, 4693-4701. [90] (a) J.-W. Wang, W. J. Liu, D.-C. Zhong, T.-B. Lu, Coord. Chem. Rev. 2019, 378, 237-261. (b) M. Rakowski Dubois, D. L. Dubois, Acc. Chem. Res. 2009, 42, 1974-1982. [91] (a) N. Hazari, J. E. Heiman, Inorg. Chem. 2018, 56, 13655-13678. (b) J. Darensbourg, Inorg. Chem. 2010, 49, 10765- 10780. [92] (a) A. Aresta, A. Dibenedetto, A. Angelini, I. Papái, Top. Catal. 2015, 58, 2-14. (b) M. Aresta, A. Dibenedetto, Dalton Trans. 2007, 2975-2992. [93] G. Parkin, Chem. Rev. 2004, 104, 699-767. (b) C. Bergquist, T. Fillebeen, M. M. Morlok, G. Parkin, J. Am. Chem . Soc. 2003, 125, 6189-6199. (c) H. Vahrenkamp, Dalton Trans 2007, 4751-4782. [94] (a) D. Huang, O. V. Makhynlets, L. L. Tan, S.-C. Lee, E. V. Rybak-Akimova, R. H. Holm, Proc. Nat. Academ. Sci. 2011, 108, 1222-1227. (b) D. Huang, O. V. Makhynlets, L. L. Tan, S.-C. Lee, E. V. Rybak-Akimova, R. H. Holm, Inorg. Chem. 2011, 50, 10070-10081. [95] T. J. Schmeier, A. Nova, N. Hazari, F. Maseras, Chem. Eur. J. 2012, 18, 6915-6927. [96] L. M. Martínez-Prieto, C. Real, E. Ávila, P. Palma, J. Cámpora, Eur J. Inorg. Chem. 2013, 5555-5566. [97] G. R. Fulmert, W. Kaninsky, R. A. Kemp, K. I. Goldberg, Organometallics 2011, 30, 1627-1636. [98] L. M. Martínez-Prieto, P. Palma, J. Cámpora, Dalton Trans 2019, 48, 1351-1356. [99] D. J. Darensbourg, W.-Z. Lee, A. L. Phelps, E. Guidry, Organometallics 2003, 22, 5585-5588. [100] Lee has reported the crystal structures of two nickel alkylcarbonates stabilized with phosphino-phosphido (PPP) pincers, with no mention of a similar water-alcohol exchange. (a) Y. E. Kim, S. Oh, S. Kim, O. Kim, J. Kim, S. W. Han, Y. Lee, J. Am. Chem. Soc. 2015, 137, 4280-4283. (b) S. Oh, S. Kim, D. Lee, J. Gwak, Y. Lee, Inorg. Chem. 2016, 55, 12863-12871. [101] (a) J. Ruiz, M. T. Martínez, F. Florenciano, V. Rodríguez, G. López, J. Pérez, P. A. Chaloner, P. B. Hitchcock, Dalton Trans. 2004, 929-932. (b) C. A. Ghilardi, S. Midollini, L. Sacconi, Inorg. Chem. 1977, 16, 2377-2380. [102] P. S. Hanley, J. F. Hartwig, Angew. Chem. Int. Ed. 2013, 52, 8510-8525. [103] H. Brombacher, H. Vahrenkamp, Inorg. Chem. 2004, 43, 6042-6049. [104] S. Chakraborty, J. A. Krause, H. Guan, Organometallics 2009, 28, 582-586. [105] J. Hao, B. Vabre, D. Zargarian, J. Am. Chem . Soc. 2015, 137, 15287-15298.