Small molecule activation with bimetallic systems: a landscape of cooperative reactivity

11220 | Chem. Commun., 2022, 58, 11220–11235 This journal is © The Royal Society of Chemistry 2022 Cite this: Chem. Commun., 2022, 58, 11220 Small molecule activation with bimetallic systems: a landscape of cooperative reactivity Miquel Navarro, Juan Jose ´Moreno, Marina Pe ´rez-Jime ´nez and Jesu ´s Campos * There is growing interest in the design of bimetallic cooperative complexes, which have emerged due to their potential for bond activation and catalysis, a feature widely exploited by nature in metalloenzymes, and also in the field of heterogeneous catalysis. Herein, we discuss the widespread opportunities derived from combining two metals in close proximity, ranging from systems containing multiple M–M bonds to others in which bimetallic cooperation occurs even in the absence of MM interactions. The choice of metal pairs is crucial for the reactivity of the resulting complexes. In this context, we describe the prospects of combining not only transition metals but also those of the main group series, which offer additional avenues for cooperative pathways and reaction discovery. Emphasis is given to mechanisms by which bond activation occurs across bimetallic structures, which is ascribed to the precise synergy between the two metal atoms. The results discussed herein indicate a future landscape full of possibilities within our reach, where we anticipate that bimetallic synergism will have an important impact in the design of more efficient catalytic processes and the discovery of new catalytic transformations. Introduction The field of organometallic chemistry, especially its application in homogeneous catalysis, has been largely dominated by mononuclear transition metal complexes. However, a vivid interest in bimetallic and polymetallic compounds has recently emerged. 1 Nonetheless, the historical development of organometallic chemistry contradicts the potentially misleading notion of Instituto de Investigaciones Quı ´micas (IIQ), Departamento de Quı ´mica Inorga ´nica and Centro de Innovacio ´n en Quı ´mica Avanzada (ORFEO-CINQA), Consejo Superior de Investigaciones Cientı ´ficas (CSIC) and University of Sevilla, Avenida Ame ´rico Vespucio 49, 41092 Sevilla, Spain. E-mail: jesus.cam[email protected]; Web: https:// jcamposgroup.iiq.us-csic.es/ Miquel Navarro Dr Miquel Navarro studied Chemistry at Universitat de Barcelona. He obtained his PhD in organometallic chemistry from Universita ¨tBern under the supervision of Prof. Martin Albrecht in 2017. Then, he joined the group of Dr Didier Bourissou at Laboratoire He ´te ´rochemie Fondamentale et Applique ´e (CNRS – Universite ´Paul Sabatier, Toulouse) as a Postdoctoral Fellow thanks to an Early Postdoc Mobility Fellowship from the Swiss National Science Foundation. In 2020, he joined the group of Dr Jesus Campos (CSIC – Universidad de Sevilla) as a Juan de la Cierva Fellow. His current research is focused on the design of bulky phosphine ligands for metal-containing cooperative systems. Juan Jose ´Moreno Juan Jose ´Moreno completed his BSc and MSc in Chemistry at the University of Sevilla in 2014, after which he joined the group of Professor Ernesto Carmona to pursue a PhD in organometallic chemistry. In 2019, he moved to the University of Virginia to perform electrocatalytic CO 2 reduction under the guidance of Professor Charles Machan. In 2021, he secured a Junta de Andalucı ´aPostdoctoralFellowship to join the Campos group, where he is currently exploring the reactivity of metalloradical species in bond activation processes. Received 1st August 2022, Accepted 6th September 2022 DOI: 10.1039/d2cc04296g rsc.li/chemcomm ChemComm FEATURE ARTICLE Open Access Article. Published on 06 September 2022. Downloaded on 11/3/2022 3:44:16 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online View Journal | View Issue This journal is © The Royal Society of Chemistry 2022 Chem. Commun., 2022, 58, 11220–11235 | 11221 genuine novelty based on the concept bimetallic design. In fact, the study of M–M bonds has sparked interest from organometallic chemists since the early beginnings of this discipline. Seminal discoveries at that time included, for instance, the first unambiguous recognition of a M–M bond in Mn 2 (CO) 10 , 2 or shortly after the disclosure of multiple bonding in [Re 2 Cl 8 ] 2 , 3 which shattered the assumption of a maximum bond order of three. However, the evolution of bimetallic complexes has been discontinuous since then, 4 where it was only in the last decade that renewed efforts have been devoted to the design and use of bimetallic organometallic architectures in a plethora of applications. This re-emergence is substantiated by several facts. Firstly, in the search for more efficient catalysts, chemists have gained inspiration from nature. In metalloenzymes, it is common to find active sites that rely on the cooperative action of two or more metals to effect catalysis with unparalleled efficiency, 5 especially in transformations that remain among the most challenging for modern synthetic chemistry. 6 Thus far, bioinspired and biomimetic bimetallic approaches have shown great prospects for future developments. 7 The area of heterogeneous catalysis also serves as inspiration for molecular chemists seeking innovative bimetallic designs. 8 This is not surprising given that metal–metal cooperation is crucial in many mixed-metal heterogeneous catalysts and nanoparticles. 9 Moreover, the advancement of operando techniques and more sophisticated computational approaches evidence increasing cases where catalytic transformations that were believed to be mediated by genuine monometallic species involve the key participation of bimetallic intermediates and transition states. 10 Continuous research on bimetallic complexes demonstrates that the presence of a second metal provides multiple tunable features that are unattainable for mononuclear species (Fig. 1). In the latter case, organometallic chemists have traditionally focused on tuning the stereoelectronic properties at the metal center by ligand modification. Beyond this approach, the presence of a second metal enables the polarity, distance and order of the M–M bonds to be tuned, which are features exhibiting a strong impact on their reactivity (activity and selectivity) and other intrinsic properties such as photoluminescence, stability, solubility, magnetism and structural conformation. For instance, the polarity of the M–M bond is associated with the mechanism by which small molecules are activated, usually following homolytic pathways in non-polarized bonds and heterolytic pathways for polar M–M bonds. Besides, bond activation can proceed either by single-site or multi-site pathways, thus offering mechanistic possibilities that are absent in mononuclear complexes, which may lead to the discovery of new transformations. The above-mentioned features have been elegantly exploited by many research groups for a variety of purposes (Fig. 2). For instance, the pseudo-C 3 -symmetric Zr/Co complex developed by Thomas is versatile, exhibiting either single-site or multi-site activation depending on the added substrate. 11 An additional advantage of bimetallic species is the ability to circumvent the often undesirable one-electron pathways, thus providing nobility to first-row transition metals. This concept has been exploited by Mankad in highly polarized group 8/group 11 complexes, which exhibit rich bond activation reactivity that can be implemented in, for instance, borylation catalysis. 12 Alternatively, the non-polarized naphthyridine-diimine dinickel Fig. 1 Tunable features of bimetallic complexes and some examples of representative pathways for single vs. multi-site bond activation. Marina Pe ´rez-Jime ´nez Marina Perez-Jimenez obtained her PhD in Organometallic Chemistry in 2021 at the University of Sevilla under the supervision of Professor E. Carmona and Dr J. Campos. Her PhD work focused on the synthesis and reactivity of dimolybdenum complexes containing quadruple metal–metal bonds. During her PhD, in 2019, she had a Predoctoral stay at the University of Berkeley, California, with the group of Professor T. Don Tilley. In 2022, she was awarded a Postdoctoral Margarita Salas Fellowship to join the group of Professor Mark R. Crimmin at Imperial College London, where she is developing her research at the moment. Jesu ´s Campos Jesu ´s Campos obtained his PhD (2012) in Organometallic Chemistry at the University of Sevilla (E. Carmona). He developed his Postdoctoral research career at Yale (R. Crabtree) and Oxford University (S. Aldridge). In 2017, he became a CSIC tenured researcher at the Institute for Chemical Research and was awarded an ERC Starting Grant on molecular cooperative systems. In 2020, he was appointed Fellow of the Spanish Young Academy. His interests include all aspects of organometallic chemistry, particularly on the study of cooperative mechanisms for bond activation and catalysis. Feature Article ChemComm Open Access Article. Published on 06 September 2022. Downloaded on 11/3/2022 3:44:16 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online 11222 | Chem. Commun., 2022, 58, 11220–11235 This journal is © The Royal Society of Chemistry 2022 complex developed by the Uyeda group is a highly active catalyst for unusual vinylidene transfer reactions. 13 Moreover, the M–M bond polarity can be tuned in such a way that the intrinsic polarity of a certain atom can be even reversed, as demonstrated by Aldridge and Goicoechea in a Au d /Al d+ complex, where the gold center behaves as a nucleophile towards small molecules. 14 This example demonstrates that besides transition metals, main group-based combinations also hold great potential. In a related example, Camp reported a strongly polarized Ir d /Al d+ system, which promotes the unusual reductive cleavage of carbon dioxide, relying on the synergistic cooperation between the two metals. 15 Whittlesey and Macgregor examined the effects of incorporating several main group metal fragments in the [Ru(PPh 3 ) 3 HCl] compound, revealing the remarkable acceleration of dihydrogen activation due to the formation of the very uncommon Ru/Zn bimetallic pair. 16 The partnership between transition and main group metals, although in its infancy, offers a landscape of possibilities, even for the apparently simple s-block metals,. A classic example is the formation of the inverse crown complex [Na 4 Mg 2 (TMP) 6 ( n Bu) 2 ] (TMP = 2,2,6,6-tetramethylpiperidine) reported by Mulvey and O’Hara, which promotes dimetalation reactions with meta–meta’ regioselectivity. 17 In fact, the combination of alkali metals with less polar organometallics has proven to be a highly prolific strategy for altering selectivity during C–H bond activation. 18 Take intoaccountthatthisisjustasmallcollectionofselected examples that demonstrate the unlimited possibilities within our grasp. In this work, we describe our efforts to better understand the synergy and cooperative mechanisms evolving from bimetallic complexes of different nature. In the last few years, our group focused on the study of bimetallic species that contain not only multiple and single M–M bonds, but also the extreme case of those that cooperate without exhibiting a direct bimetallic interaction. Our designs include systems in which the two active sites are transition metals and also those incorporating a metal of the main group series. Herein, we discuss our results pertaining to their synthesis and especially their reactivity towards small molecules, focusing on key mechanistic aspects. The discussion is organized in the following sections: – Frustration versus interaction in bimetallic systems. JFrustrated Lewis pairs based on transition metals. JBimetallic frustrated Lewis pairs. JM–M-polarized single bonds. – Multiple metal–metal bonds in X–Y bond activation. – Transition metal–main group bimetallic cooperativity. Frustration versus interaction in bimetallic systems Frustrated Lewis pairs based on transition metals It was long believed that despite the well-known ability of Lewis acids and bases to promote chemical transformations, their cooperative involvement in bond activation and catalysis was inhibited by the irreversible formation of Lewis adducts. In 1942, Brown first recognized that steric effects precluded the formation of a Lewis adduct between BMe 3 and 2,6-lutidine. 19 In the following years, various combinations of sterically demanding Lewis acids and bases were found to promote unforeseen stoichiometric reactions, 20 butitwasnotuntil2006thatStephan and coworkers reported the reversible cleavage of H 2 by a phosphino–borane (Fig. 3(a)). 21 This reactivity, considered at the time exclusive to transition and f-block metals, revolutionized main group chemistry and laid the foundation for the development of frustrated Lewis pairs (FLPs). 22 Although performing metal-free bond activation and catalysis is one of the preeminent features of FLPs, the reluctance of main group elements to partake in elemental reactions, such as oxidative addition, reductive elimination and migratory insertion encouraged the incorporation of Lewis acidic and basic transition metal sites as FLP constituents, aiming to expand the limited scope of traditional main-group FLPs by exploiting the rich reactivity of the d orbital shell. 27 The pioneering work by the group of Wass leveraged the Lewis acidity of Zr(IV) to perform stoichiometric 23 and catalytic 24 FLP reactivity in conjunction with a pending phosphane base (Fig. 3(b)). Soon after, Fig. 2 Selected examples of bimetallic systems that exhibit remarkable cooperative reactivity. Fig. 3 (a) Seminal discovery of reversible dihydrogen activation at a phosphino–borane pair. 21 (b) Representative examples of the first transition metal FLPs developed based on Zr(IV) 23–25 and first example of a bimetallic FLP based on the Pt(0)/Al(III) pair. 26 ChemComm Feature Article Open Access Article. Published on 06 September 2022. Downloaded on 11/3/2022 3:44:16 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online This journal is © The Royal Society of Chemistry 2022 Chem. Commun., 2022, 58, 11220–11235 | 11223 Erker and coworkers expanded the scope of Zr/P systems by employing a geminal pair in FLP addition reactions and hydrogenation catalysis (Fig. 3(b)). 25 Recently, Bourissou et al. developed a bimetallic Pt(0)/Al(III) complex, which displayed thermally induced FLP behaviour (Fig. 3(b)). 26 Bimetallic frustrated Lewis pairs Attempts to synthesize FLPs solely constructed around transition metal centres remained unfruitful until 2017, 28 when the use of bulky phosphine ligands allowed us to report the first example of a transition metal-only frustrated Lewis pair (TMOFLP). 29 The combination of a sufficiently congested Lewis acidic Au(I) fragment and a Lewis basic Pt(0) centre enabled the activation of dihydrogen and acetylene in a way that resembled main group FLP systems (Scheme 1), whereas the individual fragments were unreactive. In the case of dihydrogen, the heterolytic bond cleavage was followed by the formation of a heterobimetallic Au(I)/Pt(II) complex featuring a bridging and a terminal hydride. In turn, two competitive reaction pathways existed in the activation of acetylene, i.e., alkyne deprotonation, giving rise to a bridged acetylide p-bonded to the cationic Au(I)centreands-bonded to a formal Pt(II) hydride, and the formation of two metal–carbon s bonds (1,2-addition), establishing a bridging vinylene linker between the two metallic fragments. These two isomers, which were obtained in a 4:1 ratio, highly resemble the products typically obtained during alkyne activation by traditional phosphine/borane FLPs. 30 Following this seminal report, our group studied a family of Au(I)/Pt(0) systems to gain insight into the mechanism of the reaction between dihydrogen and the Au(I)/Pt(0) TMOFLP by experimental and computational means. 31 By only varying the steric profile of the terphenyl phosphine of the Au fragment, it became evident that, similar to conventional FLPs, the degree of frustration had a strong impact on the reaction rate. Fig. 4 presents a summary of our findings in this regard. Employing the smaller terphenyl phosphine PMe 2 Ar Xyl2 (Ar Xyl2 =C 6 H 3 -2,6- (C 6 H 3 -2,6-Me 2 ) 2 ) led to the observation of a heterobimetallic Lewis adduct, whereas the bulkier one, PCyp 2 Ar Xyl2 (Cyp = cyclopentyl), ensured complete frustration even in more polar solvents, which favoured the formation of the ionic Lewis adduct. The intermediate phosphine, PMe 2 Ar Dipp2 (Ar Dipp2 = C 6 H 3 -2,6-(C 6 H 3 -2,6i Pr 2 ) 2 ), as presented in Scheme 1, displayed solvent-dependent behaviour, where in apolar solvents such as benzene, frustration was retained. In dichloromethane, the broadening of the NMR resonances was consistent with the dynamic equilibrium between the independent fragments and the bimetallic Lewis adduct, which could be shifted to the latter by increasing the polarity of the medium by adding methanol. Density functional theory (DFT) calculations displayed excellent agreement with the thermodynamics of these solution equilibria. Importantly, while bulkier systems effected fast dihydrogen activation at room temperature, that based on the PMe 2 A Xyl2 phosphine, for which the resting state was the heterobimetallic adduct, displayed sluggish reaction kinetics. This was consistent with an FLP-type mechanism, which requires prior dissociation of the bimetallic adduct, and thus an additional barrier to overcome. Furthermore, the addition of either excess Au or Pt precursor to the latter system had a detrimental effect on the reaction rate, supporting that access to both independent metallic fragments was paramount to the observed reactivity. The system constructed around PMe 2 Ar Dipp2 displayed greater activity than that based on PCyp 2 Ar Xyl2 , where similar to traditional FLPs, there is an optimal degree of bulkiness, above which reactions become slower or even do not occur. The product speciation was also affected by the steric profile of the phosphine ligands bound to the Au centre, where in the case of PMe 2 Ar Xyl2 , the only discernible product was the bridged heterobimetallic dihydride, whereas for PMe 2 Ar Dipp2 , the rapid formation of cationic digold and platinum hydride complexes eventually evolved to the thermodynamic product, i.e., the heterobimetallic Scheme 1 Lewis adduct formation equilibrium and FLP-type bimetallic activation of dihydrogen and acetylene by Au(I)/Pt(0) system. Fig. 4 Solvent-dependent thermodynamics of the FLP equilibrium. Feature Article ChemComm Open Access Article. Published on 06 September 2022. Downloaded on 11/3/2022 3:44:16 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online 11224 | Chem. Commun., 2022, 58, 11220–11235 This journal is © The Royal Society of Chemistry 2022 dihydride. However, while for PCyp 2 Ar Xyl2 the digold hydride was observed initially, the formation of the bridged heterobimetallic dihydride did not occur, which allowed us to disclose the goldcatalyzed formation of the trans Pt(II) dihydride. One of our main goals is the precise determination of the mechanism for dihydrogen splitting. This transformation has served as the benchmark to gauge FLP reactivity, although it still remains under debate for main-group FLPs. 32 For this, we examined all the potential relevant pathways by computational means (see representation of key transition states in Fig. 5), including activation of the H–H bond at the Pt centre followed by Au-assisted cis–trans isomerization (orthogonal activation), across the Pt–Au bond in the Lewis adduct (bimetallic activation), and at the Au centre followed by deprotonation by the NTf 2 anion, which were all found to be unfeasible. In turn, the only energetically accessible route encompassed a true FLP-type transition state. The intermediate preceding this TS is an encounter complex in which the H 2 molecule interacts with both metal centres in an end-on fashion. Interestingly, despite its weak coordinating ability, a stabilizing interaction between the NTf 2 anion and the Au centre was found to be essential, shifting the Au–H–H angle from 79.31to 136.51. A strong inverse kinetic isotope effect (KIE) was found for the systems containing PMe 2 Ar Xyl2 (0.46) and PMe 2 Ar Dipp2 (0.50, carried out at 20 1C), suggesting that similar reaction pathways were operative for both. The calculated zero-point energy differences (DDZPE) for the PMe 2 Ar Xyl2 system gave an inverse primary KIE k H /k D = 0.40, in excellent agreement with the experimentally determined values, thus further supporting our genuine FLPtype mechanism for a bimetallic system. We also found the same feature in a more recent study based on related Au(I)/Pt(0) pairs, in which the platinum species also contains a terphenyl phosphine. 33 In contrast to the majority of examples of inverse KIEs which involve H–H bond cleavage, we attributed its origin to the collective isotopically sensitive vibrational modes at the rate-determining transition state and not due to a preequilibrium involving an inverse equilibrium isotope effect (EIE). Another aspect that has attracted our attention in this area is controlling the selectivity. In the case of acetylene activation, the ratio between the isomers described in Fig. 6 was highly sensitive to the steric profile of the phosphines employed. 34 As mentioned above, the system based on PMe 2 Ar Dipp2 gave an 80:20 mixture of a s,p-acetylide and a vinylene, respectively. For the least sterically demanding Au complex, (PMe 2 Ar Xyl2 ) Au(NTf 2 ), the regioselectivity shifted towards the formation of the bridging acetylide isomer (95:5 ratio), whereas for the bulkier complex, (PCyp 2 Ar Xyl2 )Au(NTf 2 ), a drastic change was observed, as the heterobimetallic vinylene was quantitatively formed (Fig. 6). Low-temperature NMR spectroscopy and DFT calculations indicated alkyne coordination to the electrophilic Au centre takes place first, where the attack of the Pt atom towards the C or H atoms of the activated alkyne determines the selectivity of the process. Calculations indicated that the transmetalation of the acetylide ligand following deprotonation is a facile process. Interestingly, the divergence in the reaction outcome was achieved by modifying the steric hindrance around the Lewis acid and not the strength of the Lewis base alike in main-group FLPs. This is particularly advantageous considering that modifying the structure of the Lewis acid in traditional FLPs (typically fluorinated boranes) may be synthetically challenging, 35 while subtle changes on the ligand attached to the electrophilic gold site is straightforward. 36 Heavier tetrylenes, which are species containing divalent group 14 elements (from Si to Pb), display ambiphilic character, given that they present both a vacant p orbital and lone pair with scharacter on the same atom, which can cause them to behave as Lewis acids and/or bases. This single-site ambiphilicity resembles the concept of frustration, given that the aforesaid empty and filled orbitals cannot form a dative bond. However, this aspect has been exploited by some authors to stabilize highly reactive tetrylene fragments by push–pull interactions in the presence of a donor and an acceptor. 37 In fact, Rivard used metallic Lewis acidic W(CO) 5 and basic (C 5 H 5 )Rh(PMe 2 Ph) 2 and Pt(PCy 3 ) 2 moieties to trap simple forms of divalent group 14 compounds. 38 These results prompted us to study the reactivity of tetrel dihalides (GeCl 2 and SnCl 2 ) towards the Au/Pt TMOFLP presented above. 39 The addition of GeCl 2 dioxane to a solution of the frustrated Lewis pair based on PMe 2 Ar Dipp2 shifted the aforesaid equilibrium towards the formation of the Lewis adduct, likely by sequestration of the triflimide anion (Scheme 2(a)). In turn, SnCl 2 promoted the transfer of a phosphine ligand from the Fig. 5 Representation of the key transition states for the most relevant modelled mechanistic pathways and their associated overall energy barriers for H 2 splitting. Fig. 6 Selectivity during acetylene activation by Au(I)/Pt(0) TMOFLPs. ChemComm Feature Article Open Access Article. Published on 06 September 2022. Downloaded on 11/3/2022 3:44:16 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online This journal is © The Royal Society of Chemistry 2022 Chem. Commun., 2022, 58, 11220–11235 | 11225 Pt complex to the Au centre (Scheme 2(a)). In fact, the major Pt-containing species in this process is a Pt/Sn cluster with a 1 : 1 phosphine : Pt ratio, for which computational studies indicate that all but one metal atom display ambiphilic donor–acceptor character (Scheme 2(b)). In the presence of added phosphine ligands, SnCl 2 was also found to be an effective promoter of phosphine exchange reactions between Pt(0) compounds of formula Pt(PR 3 ) 2 and free phosphine, offering a route towards unusual heteroleptic Pt(PR 3 )(PR’ 3 ) complexes. Subtle changes on the steric profile of the electrophilic Au(I) fragment proved impactful when combined with the Rh(I) Lewis base (C 5 Me 5 )Rh(PMe 3 ) 2 , where intriguing divergent reactivity was observed (Scheme 3). 40 The system bearing PMe 2 Ar Xyl2 ,the least sterically demanding ligand, quantitatively formed the corresponding bimetallic Lewis adduct, which is a Rh–Au heterobimetallic complex. In turn, the use of the bulkiest PCyp 2 Ar Xyl2 promoted an unforeseen C–H activation event at the pentamethylcyclopentadienyl (C 5 Me 5 ) ligand, which led to the formation of new Au–C and Rh–H bonds. The direct migration of a hydride from a methyl group of the C 5 Me 5 unit to the metal is well known for early transition metals, 41 but unprecedented for late ones. In the case of the intermediate PMe 2 Ar Tripp2 (Ar Tripp2 =C 6 H 3 -2,6- (C 6 H 2 -2,4,6i Pr 3 ) 2 ), the product distribution was a 70 :30 ratio between the Lewis adduct and the C–H activation product. The latter fully converted into the Lewis adduct after 24 h at room temperature, showcasing the reversibility of the hydride migration process and indicating that the Lewis adduct was the thermodynamic product. Interestingly, while the Lewis adducts displayed no reactivity towards small molecule activation, the Rh hydrides derived from C–H activation rapidly carried out X–H bond activation reactions (X = N and O). Due to their coordinative saturation and the reversibility of their formation, we proposed that these complexes behave as unusual thermally induced TMOFLPs in the activation of ammonia, water and methanol. The cleavage of N–H bonds in ammonia is a reaction of particular interest, given that it is often challenging to achieve with monometallic transition metal complexes due to the undesirable formation of unreactive Werner-type adducts. The treatment of independently prepared [(PCyp 2 Ar Xyl2 ) Au(NH 3 )](NTf 2 ) with the Rh base gave the gold amide and rhodium hydride species stemming from N–H bond activation, supporting our hypothesis. Isotopic labelling further confirmed the reversibility of hydride migration and the heterolytic FLPtype mechanism by which the X–H bonds are cleaved by the cooperative action of gold and rhodium. M–M polarized single bonds Heterobimetallic complexes, in contrast to their homobimetallic counterparts, present intrinsic polarization at the M–M bond, and therefore they often exhibit cooperative reactivity towards different substrates in a way that may resemble FLPs. In this regard, bimetallic complexes without bridging ligand frameworks that are only stabilized through a M–M dative bond between a Lewis basic and a Lewis acidic metal are known as metal-only Lewis pairs (MOLPs). 42 In fact, some of the complexes discussed in the previous section fall in this category, in particular those in which the steric constrains are reduced to the extent that the formation of a M–M dative bond is allowed. This type of complex has gathered growing attention in recent years, in no little part due to the key role of metal–metal interactions in a broad variety of catalytic transformations. 43 Braunschweig and co-workers explored the formation of a wide variety of unsupported dative bonds between the transition metal base Pt(PCy 3 ) 2 and sand p-block metal acidic fragments (Scheme 4(a)). 44 Exchange reactions served as a useful tool for experimentally studying the basicity of different metallic fragments and showed the lability and dynamic behaviour of the M–M bond, forecasting the great potential of this type of system to act as thermally induced FLPs, 45 as shortly after demonstrated for the Au(I)/Pt(0) pairs described in the prior section. Despite their great potential, the ability of MOLPs to activate small molecules has been rarely investigated. In a pioneering work, Cutler described the activation of carbon dioxide through a M - Zr bond (M = Ru or Fe), leading to metallocarboxylates stabilized by push–pull interactions (Scheme 4(b)). 46 Recently, Mankad Scheme 2 Reactivity of Au(I)/Pt(0) TMOFLP towards tetrylene dihalides. Scheme 3 Bimetallic FLPs based on Au(I)/Rh(I) combinations. Intramolecular reactivity and N–H bond activation. Feature Article ChemComm Open Access Article. Published on 06 September 2022. Downloaded on 11/3/2022 3:44:16 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online 11226 | Chem. Commun., 2022, 58, 11220–11235 This journal is © The Royal Society of Chemistry 2022 investigated the capability of unbridged polarized heterobimetallic systems based on earth–abundant transition metals to activate different small molecules such as carbon disulfide, iodomethane, benzyl chloride and dihydrogen (Scheme 4(c)). 47 In addition, these heterobimetallic complexes have shown high activity in a variety of applications in catalysis. 48 Our group also investigated the ability of Pt(P t Bu 3 ) 2 as a basic metal fragment to form different MOLP systems. The reaction of Pt(P t Bu 3 ) 2 with the acidic fragment AgX (X = NTf 2 or OTf) in either benzene or dichloromethane in the absence of light readily formed the [(P t Bu 3 ) 2 Pt -AgX] adducts (Scheme 5). 49 In addition to crystallographic characterization, 31 P{ 1 H} NMR spectroscopy revealed a slight change in the chemical shift together with a pronounced decrease in the 1 J PPt coupling constant, where 1 J PPt = 3244 Hz vs. 1 J PPt = 4410 Hz in the Pt(P t Bu 3 ) 2 precursor. This decrease in the 1 J PPt value indicates the reduced scharacter of the P–Pt bond as a consequence of the new Pt -Ag interaction. The use of coordinating anions such as NTf 2  or OTf  was shown to be crucial for the stability of these Pt(0)/Ag(I) adducts, given that the analogous MOLPs presenting less coordinating counteranions (i.e.,BF 4  and PF 6  ) display limited stability. The reactivity of the [(P t Bu 3 ) 2 Pt -AgNTf 2 ] adduct towards the activation of small molecules was examined. For instance, the MOLP readily reacted with dihydrogen and phenylacetylene under mild conditions to form the corresponding heterobimetallic dihydride and an uncommon trimetallic dibridged bisacetylide, respectively (Scheme 5). It is worth noting that none of the metal precursors exhibited any reactivity towards dihydrogen or phenylacetylene even under harsher conditions, demonstrating the cooperative cleavage of the H–H and C–H bonds. The [(P t Bu 3 ) 2 Pt -AgNTf 2 ] MOLP could also activate X–H bonds in water and ammonia, generating the oxidized platinum(II) hydrides and the corresponding precipitation of the silver hydroxide or amide salts. As discussed earlier, the activation of the N–H bond in ammonia, although less efficient than that for the above-mentioned Au(I)/ Rh(I) pair, constitutes a relevant result given its typically challenging nature. We also investigated the formation and reactivity of zinccontaining MOLPs based on the basic metal fragment Pt(P t Bu 3 ) 2 . 50 Treatment of Pt(P t Bu 3 ) 2 with zinc halides and pseudohalides in several solvents did not result in the expected adduct formation, in contrast to the readily accessible [(PCy 3 ) 2 Pt -ZnBr 2 ]. 51 We attributed this reluctance towards adduct formation to steric reasons. However, the reaction of Pt(P t Bu 3 ) 2 with 1 equivalent of Zn(C 6 F 5 ) 2 in benzene afforded the bimetallic adduct [(P t Bu 3 ) 2 Pt -Zn(C 6 F 5 ) 2 ], which represents the first example of a Pt(0)/organozinc MOLP (Scheme 6(a)). Similarly to the Pt(0)/Ag(I) adduct, a pronounced decrease in the 1 J PPt coupling constant to 3328 Hz ( 1 J PPt = 4410 Hz in Pt(P t Bu 3 ) 2 ) was detected, indicating the formation of an adduct. In contrast, the less acidic ZnR 2 (R = Me, Et, Ph and Z 5 -C 5 Me 5 ) did not react with Pt(P t Bu 3 ) 2 even under harsher conditions, highlighting the necessity for a highly electrophilic zinc centre to overcome the distortion of the linear Pt(0) precursor to accommodate the bimetallic dative bond. The reaction of Pt(P t Bu 3 ) 2 with the more exotic Zn(I)dimerZn 2 (C 5 Me 5 ) 2 resulted in the release of two equivalents of free phosphine and the precipitation of bright orange crystals of the Pt(ZnC 5 Me 5 ) 6 complex (Scheme 6(a)). This adduct is formed through the insertion of the Pt centre into the Zn–Zn bonds of three molecules of Zn 2 (C 5 Me 5 ) 2 .Itrepresentsan Scheme 4 Representative examples of MOLPs for the activation of small molecules. Scheme 5 Activation of dihydrogen, alkynes, water and ammonia by a metal-only Lewis pair based on Pt and Ag. ChemComm Feature Article Open Access Article. Published on 06 September 2022. Downloaded on 11/3/2022 3:44:16 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online This journal is © The Royal Society of Chemistry 2022 Chem. Commun., 2022, 58, 11220–11235 | 11227 unusual 16-electron octahedral complex in which each vertex is occupied by a neutral 1-electron Zn(C 5 Me 5 )ligandproviding steric crowding, which stabilizes the encapsulated electron-rich platinum centre. This is different from all the prior Zn-rich polymetallic compounds of late transition metals, which consistently fulfil the 18-valence electron rule. Computational exploration of the topology of the Pt(ZnH) 6 model system showed bond critical points (BCPs) together with their associated bond paths (BPs) between the zinc and platinum centres. In addition, no BCPs or BPs were located between the zinc atoms, indicating the absence of ZnZn interactions in the Pt(ZnC 5 Me 5 ) 6 complex. The [(P t Bu 3 ) 2 Pt -Zn(C 6 F 5 ) 2 ] and Pt(ZnC 5 Me 5 ) 6 complexes, which are inactive towards H 2 , rapidly hydrolyse in the presence of water. Similarly, a mixture of Pt(P t Bu 3 ) 2 and ZnX 2 (X = Cl, Br, I and OTf) readily activated the O–H bond of H 2 O, affording platinum hydride complexes of the formula [PtHX(P t Bu 3 ) 2 ], accompanied by the precipitation of zinc hydroxide salts (Scheme 6(b)). This reactivity is analogous to the activation of polar bonds by the combination of Pt(P t Bu 3 ) 2 with the transition metal Lewis acid [Cu(CH 3 CN) 4 ]PF 6 previously reported by Jamali and co-workers, 52 as well as the reactivity with AgNTf 2 described in Scheme 5. 49 The Lewis basic Pt(P t Bu 3 ) 2 complex could also activate dihydrogen in the presence of catalytic amounts of Zn(OTf) 2 , generating the trans Pt(II) dihydride under mild conditions. This process was found to be reversible and presented a strong inverse kinetic isotopic effect (KIE), which suggests that an FLP-type dihydrogen activation similar to the one previously detailed for the Au(I)/Pt(0) pair may be operative. As with the Lewis base Pt(P t Bu 3 ) 2 , we also capitalized on the ability of [(Z 5 -C 5 Me 5 )Rh(PMe 3 ) 2 ] as a transition metal Lewis base to design MOLPs, where the highly constrained Au(I)electrophiles described in Scheme 3 were substituted by smaller Lewis acidic fragments (Scheme 7). 53 Treatmentof[(Z 5 -C 5 Me 5 )Rh(PMe 3 ) 2 ]with lithium and sodium salts of the weakly-coordinating tetrakis (3,5-bis(trifluoromethyl)phenyl)borate anion (BAr F ) in the noncoordinating solvent bromobenzene generated the corresponding Rh -M (M = Li and Na) MOLPs. Isolation of these Lewis adducts was not possible due to rapid hydrolysis, which generated the corresponding [(Z 5 -C 5 Me 5 )Rh(PMe 3 ) 2 H][BAr F ] and alkali hydroxides, illustrating the weakness of the Rh -Li/Na interaction. However, reaction of [(Z 5 -C 5 Me 5 )Rh(PMe 3 ) 2 ] with the Grignard reagent MgMeBr affords the dimeric MOLP [(Z 5 -C 5 Me 5 )Rh - Mg(Me x Br 1x )(m-Br)] 2 , in which the methyl group bound to magnesium is mostly exchanged byabromide(Me:Brwith15:85 occupancies). The Rh -Mg bond length in complex [(Z 5 - C 5 Me 5 )Rh -Mg(Me x Br 1x )(m-Br)] 2 accounts for 2.651(3) Å, which is shorter than the sum of the covalent radii of these metals and represents the first unambiguous example of an unsupported Rh -Mg bond. The [(Z 5 -C 5 Me 5 )Rh(PMe 3 ) 2 ] Lewis base is also capable of forming MOLPs with p-block acids. The reaction of the Scheme 6 (a) Reaction of Pt(P t Bu 3 ) 2 with Lewis acidic organozinc compounds. (b) Activation of polar O–H bonds by Pt(0)/Zn(II) and Pt(0)/Cu(I) cooperative bimetallic species. Scheme 7 Synthesis of Rh(I) MOLPs with s-, pand d-block Lewis acids. Feature Article ChemComm Open Access Article. Published on 06 September 2022. Downloaded on 11/3/2022 3:44:16 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online 11228 | Chem. Commun., 2022, 58, 11220–11235 This journal is © The Royal Society of Chemistry 2022 heavier tetrylenes, GeCl 2 and SnCl 2 ,andAlMe 3 with the Rh(I)base afforded the corresponding Rh(I)-E Lewis adducts (E = Ge, Sn and Al). The formation of the bimetallic adducts was illustrated by a marked decrease in the 1 J PRh coupling constant of ca. 40 Hz and by the corresponding upfield 103 Rh{ 1 H} NMR resonances, shifted by B400 ppm compared to the Rh(I) precursor. The solid-state structure of these MOLPs showed unsupported Rh -Mbonds, withdistancesshorterthanthesumoftherespectivecovalent radii. Interestingly, the germanium and tin MOLPs represent the first examples of rhodium-bound germylene and stannylene not stabilized by the coordination of a base. Finally, [(Z 5 - C 5 Me 5 )Rh(PMe 3 ) 2 ]alsoreactswiththed-blockLewisacidsZnMe 2 , Zn(C 6 F 5 ) 2 andCuCl,formingthecorrespondingRh(I)MOLPs. Overall, the [(Z 5 -C 5 Me 5 )Rh(PMe 3 ) 2 ] transition metal Lewis base hasproventobeusefultoconstructafamilyofMOLPswitha wide range of s-, pand d-block Lewis acids with unsupported Rh -M bonds. Curiously, DFT calculations demonstrated that thedativeRh-M bond is dominated by electron donation from the Rh–P s-bonds to the electrophilic metal, rather than from a filled Rh d-orbital to the acidic site, as we originally anticipated. This finding is consistent with the decrease in the M–P coupling constants upon the formation of an adduct. Multiple metal–metal bonds in X–Y bond activation Highly polarized M–M bonds hold intrinsic reactivity, as evidenced by the selected examples discussed in the section above. However, the electron density contained in multiply bonded homobinuclear complexes also offers fertile ground for bond activation and catalysis. Since the discovery of the first quadruple metal–metal bond between Re atoms (Scheme 8(a)), 3 the vast majority of research in the field of multiple metal– metal bonding has focused on the isolation of complexes exhibiting high bond orders between different metal centres. This has led to important landmarks such as the synthesis of the first compound with a Cr–Cr quintuple bond reported by Power’s group in 2005 (Scheme 8(a)). 54 After this discovery, diverse homobimetallic 55 and some heterobimetallic 56 complexes were synthesised, including those based on first-row transition metals. 57 However, the exploitation of the close proximity of the two metals (ca. 2Å) 58 for the activation of small molecules has not been explored to a great extent and its true potential still remains mostly undisclosed. Nonetheless, promising reactivity studies regarding compounds containing multiple metal–metal bonds with high bond orders (more than three) have been accomplished by the groups of Tsai and Kempe, among others. Some instances include the carboalumination reactions towards Cr–Cr quintuply bonded units reported by Kempe’s group, 59 as well as the activation of small molecules such as CO 2 or SO 2 , 60 white phosphorous, yellow arsenic, 61 alkynes, ketones and allenes. 62 The reactivity of quintuply bonded Cr 2 species towards terminal alkynes revealed the reversible cleavage of a Cr–Cr quintuple bond to yield an inverted arene sandwich dichromium complex. 63 These species, as well as similar quintuply bonded Mo 2 complexes, 64 can act as catalysts in [2+2+2] cyclotrimerization reactions of terminal alkynes, forming 1,3,5trisubstituted benzenes (Scheme 8(b)). The catalytic cycle for this transformation was studied theoretically by Sakaki and co-workers. 65 However, the development of useful catalytic transformations mediated by metal–metal multiple bonds is still in its infancy. 66 Although paddlewheel compounds with metal–metal bonds, mainly single bonds, have been reported as active species in catalytic processes, where dirhodium(II) carboxylates are the classic example, 67 only a limited number of multiply bonded bimetallic species behaves as effective catalysts. 68 Remarkably, heterobimetallic Zr/Co complexes containing triple metal–metal bonds exhibit good catalytic activity for the hydrogenation of unsaturated hydrocarbons, which is partially attributed to the presence of a metal–metal bond, contributing to lowering the energy barrier required for the H 2 cleavage. 69 In the last few years, our group contributed to this field by investigating the reactivity of quadruple Mo–Mo bonds. We studied the activation of unsaturated molecules such as alkenes, alkynes and heterocumulenes (CO 2 and CS 2 )via a quadruply bonded dimolybdenum system, stabilized by the coordination of two amidinate Ad Dipp2 (Ad Dipp2 =HC(NDipp) 2 ;Dipp= 2,6i Pr 2 C 6 H 3 )ligandsinatrans position to each other, featuring, in addition, two trans hydride donor groups. 70 Structure Ain Scheme 9 illustrates the ambivalent Lewis acid/Lewis base character of this unique molecule, which presents polarized Mo d+ –H d bondsnexttoemptyorreadilyavailablecoordinationsites.The short Mo–Mo distance in these complexes of around 2.10 Å, a typical value for quadruple dimolybdenum bonds, 58 allows cooperative effects between the two Mo–H units. The reactivity towards alkenes demonstrated the participation of the trans-[H–MoS Mo–H] central unit in elementary organometallic reactions as reversible migratory insertion, b-H elimination and alkane reductive elimination, which are well-known transformations for mononuclear compounds but have been less studied in bimetallic complexes. 71 The sequential and reversible bimetallic migratory insertion of ethylene formed an unsaturated bis(hydrocarbyl) species. The latter evolved under an ethylene atmosphere to an ethyl-vinyl Scheme 8 (a) Representative examples of quadruple (Cotton) and quintuple (Power) bonds in bimetallic complexes. (b) [2+2+2] alkyne cyclotrimerization mediated by a quintuply bonded Cr 2 complex (Tsai). ChemComm Feature Article Open Access Article. Published on 06 September 2022. Downloaded on 11/3/2022 3:44:16 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online This journal is © The Royal Society of Chemistry 2022 Chem. Commun., 2022, 58, 11220–11235 | 11235 83 R. Taube and N. Stransky, Z. Chem., 1979, 19, 412. 84 M. Garçon, C. Bakewell, G. A. Sackman, A. J. P. White, R. I. Cooper, A. J. Edwards and M. R. Crimmin, Nature, 2019, 574, 390. 85 R. J. Somerville, A. M. Borys, M. Perez-Jimenez, A. Nova, D. Balcells, L. A. Malaspina, S. Grabowsky, E. Carmona, E. Hevia and J. Campos, Chem. Sci., 2022, 13, 5268. 86 P. P. Power, Nature, 2010, 463, 171. 87 (a) G. Bouhadir and D. Bourissou, Chem. Soc. Rev., 2016, 45, 1065; (b) D. You and F. P. Gabbaı ¨,Trends Chem., 2019, 1, 485. 88 R. J. Somerville and J. Campos, Eur. J. Inorg. Chem., 2021, 3488. 89 S.Bajo,M.M.Alcaide,J.Lo ´pez-Serrano and J. Campos, Chem. – Eur. J., 2021, 27, 16422. 90 S. Bajo, C. A. Theulier and J. Campos, ChemCatChem, 2022, e202200157. 91 (a) V. S. V. S. N. Swamy, S. Pal, S. Khan and S. S. Sen, Dalton Trans., 2015, 44, 12903; (b) T. A. Engesser, M. A. Lichtenthaler, M. Schleep and I. Krossing, Chem. Soc. Rev., 2016, 45, 789; (c) H. Fang, Z. Wang and X. Fu, Coord. Chem. Rev., 2017, 344, 214. 92 S.Bajo,M.M.Alcaide,J.Lo ´pez-Serrano and J. Campos, Chem. – Eur. J., 2020, 26, 15519. Feature Article ChemComm Open Access Article. Published on 06 September 2022. Downloaded on 11/3/2022 3:44:16 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. View Article Online