C-H functionalization via transition metal electricatalysis

(1)

MSc Chemistry

Molecular Sciences

Literature Thesis

C-H functionalization via transition metal

electrocatalysis

by

Said Ortega Rosales

UvA ID: 11668660

July 2020

12 ECTS

September 2019 – July 2020

Supervisor/Examiner:

2

nd

_Examiner:

Dr. M.A. (Tati) Fernández Ibáñez

Dr. Ning Yan

Synthetic Organic Chemistry

(2)

(3)

(4)

4

C-H functionalization via transition metal electrocatalysis

ABSTRACT: Functionalization of inert C-H bonds is one of the more promising paths for the evolution of synthetic chemistry. It is hard to imagine other synthetic strategies that con surpass the level of atom efficiency that these transformations, inert bond functionalizations, can potentially offer. The field is rapidly developing, and solutions to reduce its limitations are continuously arising. In this regard, the implementation of electrocatalysis works synergistically towards the goal of a sustainable development of the field. The use of electricity as a source of electrons offers the possibility to leave behind the common practice of using stoichiometric oxidants, expensive and sometimes toxic. Additionally, the recurrence of undesired byproducts can be decreased, and milder, more selective conditions can be employed in the reaction systems, as it occurs with the use of redox mediators. In this work, the combined benefits of transition metal electrocatalyzed C-H functionalization strategies are explored, dedicating special attention to the different roles that electricity might play in the mechanism of the transformations presented, along its implications in the reactivity of the system. The first part of this review is focused on the background and a thorough description of the key concepts that coalesce within this field, aimed at facilitating the understanding and analysis of the methodologies presented to the reader. The next section discusses the electrochemical C-H phosphorylation reaction, achieved by different transition metal catalysts. For this transformation, bimetallic systems offered the highest catalytic activity, due to an increase in the turnover frequency when hetero metallic systems were employed.

The third section focuses on methodologies recently developed in which and Earth-abundant transition metal is used as the catalyst, namely, iron, cobalt, nickel and copper. Manganese is also included in this category, and thus it is also discussed in this section. The transformations involving these inexpensive metals, in contrast with more traditionally used noble metals, are discussed in detail, in order to maintain the overall focus of the work on the benefits on sustainability that these protocols offer. The different in the reactivities of these metals is highlighted, as manganese and iron tend to react through single-electron transfer, SET, mechanisms, reflecting on radical intermediates leading the reactivity. In the case of iron, redox-mediator strategies are prominent. In contrast, cobalt and nickel react almost exclusively through two-electron transfers mechanisms, translating into polar reactions, involving reductive elimination steps, and base-assisted internal electrophilic-type substitution (BIES). Interestingly, in reactions catalyzed by cobalt was stablished the first electro-removal strategy for the cleavage of a traceless directing group. Copper displays a characteristic chemical versatility, and its reactions can be radical or polar in character, with the possibility of involving a SET step, and a two-electron transfer within the same mechanism. Reactions involving Earth-abundant catalysts that have been discussed in recent reviews, are only presented schematically in order to present a summary that places the reader in context.

Finally, transformations reported recently involving noble metal catalysts are presented only schematically, without an in-depth discussion, partially because the focus is maintained on the cheaper, accessible metals, and partially because of complete reviews published recently that focus mostly on the methodologies of these metals. The reactions presented are classified according to the role that electricity plays in the mechanism proposed by the authors of each study. This proposed categorization is presented in page 17.

(5)

5

1. Introduction

What is the purpose of a scientist? What is the purpose of a chemist?

Science seems to arise as consequence of yearnings inherent to the human soul -or mind, whatever it might be called-. The need to understand the universe around us and our place in it, the yearn to create, and the aesthetic appreciation of these processes. In this sense, science is a mirror and complement of art, as they both are explorations; ways to appropriate the inner and outer world, and allies in the endless pursuit of the truth. Consequence of this quest, humanity is constantly provided with valuable tools to transform the world that surrounds us; making it adapt to us.

The tools provided by science cover a vast spectrum; and have been crucial for the survival and development of society towards its current state -as good or as bad as it might considered. Today, science is one of the pillars upon which humans rely to make the world suit their basic needs, increase the comforts of life and reduce the harshness of hard labour, at least for certain percentage of humanity; making our surroundings far less hostile and alleviating, in practical ways, the burdens of existence.

In this perpetual and collective endeavour, the role of chemistry has been essential; and it is now rooted in our everyday routine so deeply that we might not even see it sometimes. From the moment when we wake up, aided by our phone´s alarm, to the moment we cook our dinner prior to sleeping, we are under the wing of chemistry. Food, medicines, clothes, transportation, electronic devices, weapons, paper and ink; all different manifestations of our ability to guide the transformations of matter. Ideally, the shaping of our surroundings should be guided towards the common good, not only for humans, but in harmony with the world we exist within; which is as part of us as we are part of it.

The chemist has the opportunity, or duty, to strive for equilibrium; as sustainability is the only possible form of long-term development. Nowadays, humans have at their disposal a vast chemical toolbox that allows them to access almost any chemical architecture that might be required, but to do so in an efficient way, which is respectful with the resources and the environment, is one of the main tasks that should be addressed.

There are many approaches towards this goal, but it is hard to think of a more efficient and straightforward manipulation of molecules than activating and functionalizing what seems to be nature´s favourite atomic duo: the C-H bond. Just by looking at almost any molecule, most evidently in organic matter, this field shines with light of its own. The ubiquitous nature of C-H bonds makes their functionalization a shortcut to conventional organic transformations, opening the possibility of avoiding several synthetic steps towards a target molecule, having a direct positive impact on the atomic efficiency of the transformations.

The carbon-hydrogen bond is easy to find but not easy to manipulate synthetically, due to its low polarizability and relatively high bond strength. Thus, its functionalization uncharted territory until relatively recently. In the past 50 years this paradigm has been changing with the more systematic and conscious exploration of the enormous potential of this irresistible field, and strategies to simplify late stage functionalization in organic complex molecules are no longer considered exotic exceptions to the rules. The symmetries present in the orbitals of transition metals open the possibility of unique interactions to cleave existing bonds in the reacting substrates, and to facilitate asymmetric, enantioselective functionalization, which is a challenging but desirable task.1–12

To dive deeper into the topic, it might be important to clarify some subtle differences in terminology. Usually, the term “C-H activation” refers to the stage in which the C-H bond is replaced by a C-M bond by reacting with the metal catalyst. Then, the functionalization stage consists in converting this bond into a C-X bond that will appear in the final product, where X is any desired organic substituent. A different common functionalization steps consists in β-elimination of an intermediate alkylmetal, which leads to an unsaturated organic product.13_{Despite the}

differences, these terms are often used indistinctly, both for practical purposes and because the line that divides them is often blurry; and the same will apply in this work.

(8)

8

1.1. A brief voyage through time

Nature has always been source of inspiration and guidance for humans in many aspects, and C-H activation is not the exception; it has been practiced for several billion years by some of the most exceptional chemists we know of: bacteria.10,14,15_{More recently, humans made their first incursions into the field, aided by serendipity and ignorance,}

when in 1879, A. W. Hofmann was pursuing the elucidation of the structure of piperidine.16_{His studies, involving}

N-haloamides, resulted later into a synthetic route towards pyrrolidines.17,18_{Despite its utility, the mechanistic inner}

workings of the synthesis remained a mystery until the late 1950s, when studies from Wawzonek, Thelan, and later the group of the great E.J. Corey demonstrated that the reaction proceeded via a free-radical chain mechanism, involving intramolecular hydrogen transfer as one propagation step; making it the first example of C-H functionalization, to the best of our knowledge.19,20

These studies were extended by Löffler and Freytag, who used it for the elegant synthesis of no less than nicotine in 1909, becoming the earliest example of direct C-H functionalization in total synthesis! (Scheme 1).21

Scheme 1: Synthesis of nicotine via Hofmann-Löffler-Freytag reaction.

Then in 1898, another early example that paved the way into the field was documented, known as the Dimroth reaction, which can be considered an analogous of the Friedel-Crafts reaction, thought to proceed via an electrophilic attack on an arene π-system.22_{It was with the renaissance of inorganic/organometallic chemistry in}

the 1950s and early 1960s that another relevant step towards the emergence of the field was taken, relying on some important and clever observations in reactions involving the closest relative in polarity and bond strength to the C-H bond; the bond in C-H2. In the year 1954, the copper catalyzed cleavage of H2 was found by Halpern and coworkers;

which eventually led to Cu-catalyzed reduction of various metal ions by H2(eq. 1). 23,24

2𝐶𝑢𝐴𝑐2+ 𝐻2+ 𝐻2𝑂 → 𝐶𝑢2𝑂 + 4𝐻𝐴𝑐 (eq 1)

The discovery that other metal ions, such as mercury (II) and silver (I), could participate in the same type of reactions soon followed.24_{Later, the first well-defined complex oxidatively catalyzing hydrogen splitting reaction,}_trans

-[Ir(PPh3)2(CO)Cl], was reported by Vaska in 1962.25 These observations provided valuable insights that worked

synergistically to pave the way for the blossom of the field, and the relevance of transition metals was evident. Noticeably, these two described reactions, foreshadowed the two main approaches to polar C-H bond activation by transition metal complexes that are currently more widespread: electrophilic activations and activations through oxidative addition, which will be further discussed later.

Finally, in the early 1960s what is considered the first formal reported examples of C-H activation by transition metal complexes arrived, reported by Chatt and Dubeck (Scheme 2).26,27_{Activation of aromatic compounds was a}

(9)

9

remarkable achievement, due to their characteristic stability and reluctance to react. Hence, these methodologies were keys to access yet unexplored doors of synthetic chemistry.

Scheme 2: Early examples of C-H activation by well-defined complexes through oxidative addition.

Life is characterized by the alternance of periods of bustle and calmness; the same is true for science, and during this period organometallic chemistry was a pot of boiling water, with many pioneering works on the field seeing the light. Major contributions were made by Shilov´s seminal works in 1969 and 1972 on platinum aided olefin hydrogenation, which showed the C-H elimination from a Pt alkyl to be reversible, and that it could be used to promote the C-H activation of methane via an electrophilic C-H activation step. Additionally, Shilov´s group showed that the reaction could be turned into a catalytic cycle by introducing a stoichiometric amount of a two-electron oxidant into the system, [PtCl6]2-, to yield methanol, or methyl chloride when performed in the presence

of HCl (Scheme 3).13,28–30

Scheme 3: Methane oxidation via electrophilic C-H activation in Shilov´s system.

Shilov´s system is catalytic in Pt(II), but unfortunately is stoichiometric in Pt(IV). Despite this impracticality, this work and its relevance have not been eclipsed by time, as the efficient conversion of methane to valuable chemical commodities remains as one of the main challenges to be tackled in contemporary chemistry.

The complementary picture arrived in 1979, when Crabtreee made the discovery that the dehydrogenation of cyclopentane and cyclooctane was possible using iridium catalysts, which led to unsaturated compounds as products!31_{This C-H activation represented a shortcut to access many more functional groups. Around a year later}

(10)

10

it was first shown, but not last, that rhenium compounds could also catalyze this type of reactions, in both cases the process involved oxidative addition to the transition metal complexes (Scheme 4).32–34

Scheme 4: Dehydrogenation of cyclopentane and cyclooctane.

It was in 1982 when the first controlled oxidative addition of a completely saturated hydrocarbon was achieved in the group of Bergman using iridium compounds under photolytic conditions.35_{A surprising aspect of Bergman´s}

findings was the selectivity for C-H bonds encountered, as the Cp*Ir(PMe3) compound displayed a marked

preference for the stronger C-H bonds, showing a reactivity tendency of aryl>1°>2°>3°.36_{Further studies by other}

research groups soon made apparent that both kinetics and thermodynamics favoured the addition of less substituted, less hindered, alkyl groups.37_{Furthermore, it was demonstrated that π-coordination to metal centres is a factor}

favouring the increase in rate of activation or aromatic compounds.38,39_{The reactivity trend aroused interest by the}

community, as it is opposite to those commonly observed in other reagents, meaning a complementary approach, and a promising opportunity in the much-pursued functionalization of n-alkanes. Additionally, this trend can also account for greater selectivity towards the reactants with respect to the products of the reaction in most cases.40

Every piece of evidence contributed to stablishing the foundations of the field. Such bases have proven solid and yet flexible, to allow a rapid development, were concepts are expanded, and sometimes changed and improved. With all its potential, research in C-H activation still hides many mysteries, and with them, opportunities to move the field forward, and exploit more and more of its possibilities. Because looking into the past is only the first step for moving into the future, the narrative of this work will shift its focus towards the other key aspects. Namely, the following pages will be addressing the relevance of mechanism and energetics, selectivity, directing group, electrocatalytic implementations to the field. These will lead to a discussion of the latest developments in C-H functionalization via transition metal electrocatalysis, finalizing with a humble discussion of the present panorama and the envisioned directions that could be fruitful and interesting to explore.

2. Setting the ground: Important concepts

2.1. Mechanism and energetics

In all the examples discussed thus far, the reaction pathway involves the coordination of the C-H bond to be activated, directly to the metal centre, forming an organometallic complex. These processes are often called “inner sphere mechanisms”, because the C-H cleavage takes place within the inner-sphere of the coordination compound. On the other hand, so-called “outer sphere mechanisms” also exist, in which the insertion of the C-H bond, that triggers the activation, occurs into the ligand of the transition metal compound. Outer sphere mechanisms are common in enzymatic and bioinspired functionalization of inert bonds, where they usually proceed via rebound mechanism. In these events, an initial hydrogel radical abstraction by an oxo species is followed by a rapid rebound of the radical species into a metal hydroxo intermediate. Also belonging to this category are metal-carbenes and metal-nitrenes, in which the C-H inert bond reacts with either the carbon or nitrogen atom respectively, with concomitant metal dissociation, through a 3-centre transition state. Additionally, in these mechanisms, the formation of radical species that will lead to cleavage of the C-H bond of interest, might be induced by an electron transfer from the ligand in the catalyst. To achieve this, it is often necessary that the metal transfers itself an electron to the

(11)

11

ligand, this electron can be derived from electrochemical methods. The focus of this work will be on processes involving inner sphere mechanisms, which are traditionally understood as C-H activation reactions, as in those there is a formal M-C bond involved, although also radical mechanisms will be presented and discussed, in light of their significance for certain Earth-abundant metals, as the case of iron.4,11,41–43

Different mechanistic studies have been performed which have shed some light into the energetic factors at play in these reactions, and they have facilitated a systematic division depending on the preferred pathways for reaction of the substrates. For example, with more electron rich compounds, such as Ru(dmpe)2 in scheme 2, the reactions

tended to proceed through a nucleophilic mechanism, involving oxidative addition; while for the more electron-deficient species, such as Pt(II) in scheme 3, an electrophilic mechanism is often displayed.27,28_{In most cases, both}

type of reactions proceeded via a σ-bond complexed intermediate and often, a cyclometallated intermediate.44,45

Additionally, electrophilic activations in many cases involve complete oxidative addition, followed by deprotonation of the resulting metal hydride.46_{This classification facilitates the study of the reactions but, is not meant to be a rigid}

wall between the two, and despite being driven by entirely different forces, they still share many similarities, making their distinction an organizational tool, and thus a self-propagating categorization although their differentiation is made with a blurry line.

In inner sphere mechanisms, the different symmetries of the orbitals of transition metals allow a two-path charge-transfer interaction. On one hand, the forward charge transfer, from the filled σ(C-H) to an empty metal-based dσ orbital; and on the other, the occupied dπ orbitals of the metal interact with the empty σ* orbitals of the coordinated

C-H bond, what is known as reverse charge transfer or backdonation, which is essential for weakening and ultimately cleaving the C-H bond (Scheme 5).47

Scheme 5: Frontier orbital interactions for a) electrophilic and b) nucleophilic mechanisms, highlighting in blue the dominating direction of charge transfer.

The driving force for a preferred pathway of reaction will be mostly dependent on the electronic nature of the metal promoting the activation and, in general determines the dominant direction of electron transfer in the reaction intermediate.47_{Schematically, the mechanisms can be classified in a) oxidative addition (OA), b) electrophilic}

addition (EA) and c) sigma-bond metathesis (SBM).46–48

In electron deficient coordination compounds, including cationic complexes, the dσ and dπ orbitals are expected to have low energies, and the charge transfer (CT) goes preferentially in the forward direction. These transition metal complexes usually react via electrophilic addition mechanism. Likewise, electron rich compounds are expected to possess high energy dπ and dσ orbitals, which favours the reverse CT, and are defined as nucleophilic compounds and react usually via an oxidative addition mechanism.46

Electron deficient coordination compounds might be defined by a metal that is too electronegative and electron poor, such as late transition metals in high oxidation states (PdII_{, Pt}II_{, Rh}III_{, Ir}III_{…), and/or by the presence of}

(12)

12

population of the σ*(C-H) orbital, and heterolytic deprotonation by an external ion is often followed, via a cyclic, concerted mechanism, sometimes receiving the name of concerted metalation deprotonation (CMD). Often this deprotonation is done intramolecularly by a ligand in the compound. An open coordination site in the coordination compound facilitates this type of mechanism. 49,50

In the case of electrophilic early d0_{transition metals in high oxidation states, OA is not possible, and the most}

expected reactivity is via a sigma-bond metathesis mechanism, reminiscent of a four-centre [2σ+2σ] cycloaddition, with a kite-like transition state, occurring without change in the oxidation state of the metal. This type of mechanism is often common for d4_{to d}8_{as well, with formation of stable σ-adducts before and after metathesis.}46,51

In terms of concerted four-centre additions, it has been observed that the C-H bond can also add across a M=X bond via a 1,2-addition, without detachment of the X-H motif and involving the lone pairs of π-electrons of X. This mechanism is common for amido, alkylidene, alkoxy and alkylidyne complexes of early and middle transition metals.46

Scheme 6: Possible energetic profiles for C-H bond activations, excluding radical pathways.

The types of mechanism for C-H activation are distributed in a spectrum, with bountiful of subclasses lying in between each category. The most likely mechanism for each reaction will depend on the metal centre, its oxidation state, the steric and electronic properties of the ligands, as well as the substrate involved; ultimately influencing the energy of the participating orbitals. It should be beared in mind that due to the number of factors involved, is not rare to find that certain reaction deviates from the predicted mechanistic category, and surprises are found often. Scheme 6 summarizes the possible scenarios discussed above.41,48,52

When analyzing the mechanistic aspects of any chemical process is always important to keep in mind the general state of thermodynamic and kinetic aspects; and in this field in particular there is a critical point: the addition of C-H bonds to transition metals is generally a thermodynamically unfavourable process, being reductive elimination the rule. This is evident from comparing the strengths of C-H, M-H, and M-C bonds, being the C-H bond the strongest of the three, then the M-H bond, and the weakest is in general the M-C bond; although data for absolute bond strengths is limited for many complexes. Nevertheless, when trying to predict the outcome of these transformations, it should be take into account the delicate energetic balance due to the differences in M-C bond strength, as primary metal-carbon bond is stronger than the secondary metal-carbon bond by more than the difference in the C-H strengths that need to be broken. Likewise, arenes are easier to functionalize partially because of the higher M-C bond strength, and perhaps partially due to a facilitated coordination of the pi electrons, compared to those in sigma bonds. Additional to the differences in bond strength, the energetic cost of dissociating a ligand to open a coordination site, when applicable, is also an expensive factor for these reactions.53,54

(13)

13

To surmount the limits that thermodynamics impose to these methodologies, an approach is to generate, by applying much effort and energy, a highly reactive, coordinatively unsaturated, species. The energy to achieve such intermediates can come from photons, heat or electricity in some instances. There have been attempts to understand how C-H can be facilitated from a theoretical point of view. Hoffman provided a framework to rationalize the propensity of CpML d8_{compounds to activate these reactions. This seminal work showed that if 16-electron}

square-planar ML4 d8 compounds are “bent back”, they produce a fragment possessing a high-energy filled orbital of

π-symmetry, which can mix with the σ*(C-H) orbital, and a low energy empty σ-symmetry hybrid orbital. This fragment was shown to be isolobal to CpML, as well as to CH2, and calculations showed that both fragments could

smoothly add C-H bonds.53,55

The proper selection of ligands, such as Cp, can prevent the formation of a relaxed state, and make the transition state for addition more easily accessible. Coordination compounds with monodentate ligands can more easily form low energy arrangements rather than undergoing C-H addition. In this way, kinetics can be used to balance thermodynamics.55

Despite the differences in the mechanistic pathways presented above, some features have been found to be quite common in activations. Namely, a) the formation of a σ-complex is usually the first step in the mechanism, generally favoured in electronically unsaturated compounds, and b) the generation of a cyclometalated intermediate, commonly a five or six-membered ring, which has proven to be relevant both for aryl and for aliphatic activations, and their importance should not be overlook, as their formation is relevant to induce ortho activations in aryl compounds, but it also represents a catalyst deactivation pathway the cyclometalated compound is too stable. As an inorganic analogous to the Thorpe-Ingold effect, cyclometallation can be favoured by the presence of geminal substituents. Overall, the formation of such intermediates can be modulated by a balance of two factor: the conformational effect, which disfavours cyclometallation, and steric congestion, which favours it.13,56–58

Some highlights that might aid into the design of a catalyst can be extracted from these interrelations. The compound must be coordinatively unsaturated or have a potential for this, to allow interaction the filled σ(C-H), and to achieve the desirable precoordination. Likewise, second and third row metals are preferred due to their M-C strength, as well as early d-block and f-block elements. Additionally, steric congestion can be limited to prevent cyclometallation. Although not necessary if there is a base that can assist the mechanism, a filled orbital based on the complex is necessary to interact with the σ*(C-H). Furthermore, when the substrate to be funtionalized is too unreactive, such as alkanes, a high concentration can improve the results, in some cases even be used as a solvent.

2.2. Selectivity

To be able to exploit the full potential of a chemical transformation, understanding and controlling its selectivity is a key aspect. In the following section, a brief discussion on key parameters involved in the preferred reactivity is presented. Although some clues have been gathered, still many questions remain unanswered, as appealing challenges for the future, that difficult to develop an absolute guide to explain selectivity. What might be clear, is the distinction between “innate” and “guided” selectivity. Innate, or substrate controlled, reactivity, is dominated by the qualities of the substrate, such as steric factors and the energetics, meaning that it possesses inherently more reactive sites, preferred for the reaction if not other parameters are considered. Guided reactivity, on the other hand, might be though as arising not from the substrate or the catalyst alone, but from their interaction within the reaction conditions. This means that it can be influence by the presence of a directing group, or other additives, hence being easier to modulate. A rationalization of the factors involving and their interrelations is essential for the distillation of trends, that could make possible to envision a map to facilitate the navigation through the maze of selectivity in C-H functionalization.59

Focusing on the substrate-controlled selectivity, both the steric and energetic contribution have been partially described in the previous section. For steric factors, not much more will be said, as it might be understood more intuitively. Conversely, some aspects should be pointed on the other parameter, as the energetic balance acts as an

(14)

14

orchestra conductor, influencing the pathway and outcome, always embedded in selectivity observed in chemical transformations.

One interesting observation is that, as mentioned before, the regioselectivity for functionalization by a transition metal coordination compounds is often the inverse, and thus complementary, to that of classic organic functionalization, showing preference for activation of primary carbon atoms, and decreasing for tertiary ones, which can be explained by the above mentioned trend in M-C bond strength.36,40_{Is then evident that one essential}

determinant to understand these transformations is the bond energy. There is a very concise review written by Xue and coworkers which goes in depth to this complex topic,54_{thus we will have only a pragmatic approach, focusing}

on some of the more relevant highlights encountered so far.

In the case of activation of bonds by transition-metal complexes, which usually involve the formation of a new C-M bond, accompanied by the cleavage of a C-H bond, just comparing the energetics of these two bonds might represent a good approximation to understand the activity/selectivity relationships, ignoring any reorganizational energy factors for the sake of simplicity. Once the drawback of finding the proper bond energies to perform an analysis of the energetic balance is surpassed, the next challenge is choosing the most adequate definition for the mode of C-H bond scission. The three major modes are represented in equations 2-4 below; where the bond dissociation energy (BDE) corresponds to the enthalpy of homolytic dissociation (kcal mol-1_{), pK}

a to the acidity

equilibrium, and ΔHhydride to the hydride affinity of carbocation (kcal mol-1).54

(𝐵𝐷𝐸) 𝐶 − 𝐻 → 𝐶._{+ 𝐻}.₍_eq₂₎

(𝑝𝐾𝑎) 𝐶 − 𝐻 → 𝐶−+ 𝐻+ (eq 3)

(∆𝐻ℎ𝑦𝑑𝑟𝑖𝑑𝑒) 𝐶 − 𝐻 → 𝐶++ 𝐻− (eq 4)

Which elementary process describes better the energetic changes in C-H functionalization will strictly depend on the mechanism that controls each reaction. Unfortunately, often this information is not available, and certain degree of accuracy and specificity must be sacrificed to reach a useful generalization. In this regard, the strength of a target C-H bond to undergo homolysis is usually the first parameter to consider in a C-H functionalization study, often finding useful correlations with the observed results.54

In many cases, it is only possible to correlate the reaction rate with the BDEs of the interacting bonds, showing in general higher selectivity towards bonds with lower BDEs. For instance, it has been observed that bonds adjacent to electron-withdrawing groups are more prone to activation due to a decrease in their BDEs, as determined by competition experiments, that facilitates abstraction of a hydrogen atom. This effect would be more important in mechanisms involving a radical abstraction of some kind.60

An important limitation to keep in mind when applying BDEs, is that C-H heterolysis is dramatically affected in solution, in comparison to gas phase BDEs, which are not influenced by the medium. In cases where heterolytic cleavage takes place, for example when concerted metalation-deprotonation mechanism is operating, pKa has been found to be relevant, for example when Pd or Cu are employed. The use of pKa as parameters is downturned because these values are much less studied that those of BDEs. 54

Energetic analysis can also be helpful to rationalize cases of direct hydrogen atom transfer (HAT), which is favoured by thermodynamics and kinetics and is often in reliable agreement with the BDEs. HAT consists in the concerted movement of both a proton and an electron between two substrates in a single kinetic step (eq. 5). It can be considered a subclass of proton-coupled electron transfer (PCET) processes, in which both the proton and the electron move together, sharing the starting and final orbitals. 43,54

𝐴1− 𝐻 + 𝐴2. → 𝐴1.+ 𝐴2− 𝐻 (eq 5)

This elementary form of R-H bond scission seems to be a key process for these activations, particularly those in aliphatic compounds, as the results of mechanistic studies of different reactions, suggest HAT to often be the

(15)

rate-15

limiting step. An indication of typical single-step hydrogen atom transfer reactions is found when the kinetic analysis shows that the second-order rate constants are in correlation with the bond dissociation enthalpies of activation of the C-H bonds that is being cleaved. This mode of reaction is mostly present when high-valent transition metal-oxo complexes are used to promote reactions.60

Unfortunately, energetic analysis of inert bond activation is still in early stages of development and due to the many experimental and theoretical drawbacks encountered, the cases in which it can be applied to understand the reactivity observed are exceptional. Additionally, the different trends observed might indicate that the reactivity and selectivity of C-H activation in different systems might be dictated by different types of bond energies. Hence, it remains as a field of opportunity for development and innovation.

Because thermodynamics is not open to negotiations, a lot of attention has been paid to “guided selectivity” which can me thought to be more related to kinetics. This approach offers greater flexibility, and possibilities to be conveniently tuned. The interaction of substrate and catalyst can be directed with considerations related to the already mentioned cyclometallated intermediates. Other strategies focus on restricting the trajectories in which the substrate can approach the catalyst, such as a non-heme iron catalyst that incorporates minimal steric hindrance elements that reduce the freedom of coordination for certain C-H bonds to a non-heme Fe- oxo coordination compound.61

2.3. Directing Group

Likewise, is not a secret that proximity induces reactivity, both in humans and molecules, and the presence of atoms in the substrate that can coordinate to the catalyst greatly influences the site of reaction. In this way, a preorganization of the system can be induced, which conveniently brings close the reactive centres, thus favouring a desired (regio)selectivity.62

In many cases, the directing moiety is a σ-coordinating functional group, which sometimes can even improve the activity of the catalyst. In aromatic heterocycles, the heteroatoms usually determine the site of functionalization due to their coordinating properties to the metal centre, and the influence they excel on the electron density of the C-H positions and the catalyst itself. On the other hand, in arenes and aliphatic compounds these differences are not so determinant, and they require the assistance of a somehow “internal ligand” to direct the metal catalyst into proximity. These directing moieties can sometimes overrule internal reactivities of the substrates. The directing groups (DGs) often enhance the formation of the metalacyclic intermediate.62

DGs are usually small, readily available molecules, which more commonly direct the reactivity towards the ortho position, and with less frequency to the meta position, and rarely to the para position. It is possible to find DGs which are monodentate, bidentate, heterocyclic, and heteroatom-based, and the selection of the adequate moiety will depend on the specific transformation, the conditions in which is performed, and even the availability of resources; each DG having its own advantages and disadvantages, depending on the desired goal.62_{For example,}

bidentate directing groups seemed to be very effective compared to others, such as pyridine, in annulation reaction with ethylene and ethene in the presence of a copper as catalyst.

The carbonyl group in ketones has been one of the first functional groups used as DG, applied in reactions such as, arylation, alkylation, alkenylation, alkynylation, acylation, and C-N bond formation. Use of aldehydes has been far more limited as traditional directing groups, and carboxylic acids haven only been used recently, mostly as ortho directing groups in arylation, alkylation, allylation, alkenylation, alkynylation, aminocarbonylation, C-Heteroatom bond formation, and C(sp3_{)-H activation. Esters, which are quite ubiquitous, have also been employed, although}

not as widely due to their weak coordinating properties towards transition-metal catalysts, and few examples are found in alkylations, alkelynations and interestingly, in borylation, so relevant for cross-coupling reactions. Amides have proven to be very valuable directing groups and are among the most frequently used today. Bidentate N,N-ligands have shown to remarkably effective directing groups, although there are also many examples in which

(16)

16

monodentante amide base moiety are used for arylations, alkylation, allylation, alkenylation, C-C, C-O, and C-N bond formation, halogenation, and C(sp3_{)-H activation. Besides carbamates,urea derivatives, oximes, hydrazine and}

hydrazine derivatives, pyridine, pyrimidine, among others have been employed in similar ways, but a particularly interesting DG used in recent years are nitriles. Although only few ortho funtionalisations have been reported for this group, it is very well suited to achieve for aromatic meta funtionalisation, a quite uncommon feature, believed to arise from its linearity. In most examples, the the cyano group must be attached to a linker to perform this.

Scheme 7. General strategy for designing meta-directing moieties.

In the most convenient scenario, the DG constitutes a part of the target molecule or a precursor, but unfortunately this is not always the case, thus an important drawback is the installation and/or removal of the directing group. There are detailed protocols to achieve this, but the additional steps considerably decrease the overall efficiency of the methodologies. Naturally, some groups, such as carboxylic acids, are easier to introduce, and others, like bidentate N,S-groups, are easier to cleave, representing this an advantage for them.

It is then ideal that a DG is traceless, or easily converted into a useful functionality. Thus, turning ubiquitous moieties into DG is a point of high interest for research, and scaffolds possessing an heteroatom such as P, Si, O, N, etc. have been widely studied, and even more due to their potential advantage to possess a stereogenic character, facilitating diatereoselective C-H funtionalisations.

To open new possibilities, a strategy for the use of transient directing groups is emerging, which focuses in the reversible in situ installation and deconstruction of Lewis-basic entities, with the aid of co-catalytic additives. These additives transform a weakly coordinating functional group into a better σ-donor motif. Examples include formation of ketone derivatives, such as imines, oximes or hydrazones, or phosphites. Imine DGs, can be formed in situ by adding co-catalytic amounts of nitrogen-derivatives to aldehyde-containing molecules. Also, the use of multidentate additives can be used to further exploit chelation effects of the DGs.63

It is important that the transient formation of the DG occurs reversibly and with excellent chemoselectivity; that the transient ligand do not interfere with the attempted transformation, and the ligand should also be stable prior and during the target reaction. As the example mentioned, there are many others, with the list constantly expanding, and an extensive review on the topic already exists, as this approach bears great potential for more sustainable funtionalisation. Efforts are required in this area, remaining largely unexplored, and cheaper chemicals, like non-noble metals, need to find their way into this approach, that for sure will continue to attract the scientific curiosity and attention.62,63

2.4. Going electric

Electricity has always been a source of fascination for humans; from the shock of lightnings to the relieve of light bulbs. It has been a crucial tool for our development, and today impregnates practically all our activities.

We find the flow of electrons at the core of chemistry, hence is not surprise that the possibility to use electric potential to drive chemical reactions has been explored for almost 200 years, starting with the electrolysis of acetate to produce hydrocarbons, carried by Faraday, pioneering the field. But despite this example, and many others known throughout time, it is surprising that electrochemistry and synthesis have been traditionally conceived as

(17)

17

completely separated areas of chemistry. Fortunately, this paradigm seems to be changing in recent years, and the application of electrochemical methods to synthesis is gaining interest, as a novel and potentially cleaner driving force for chemical transformations.64–67_{Additionally, deprotonation reactions can be performed without}

H-acceptors, and be replaced by the concomitant production of H2.68

The use of electrochemical strategies enables, another implication is it unlocks a special reactivity. By adding electrons to electron-poor functional groups, new nucleophiles are regenerated. Likewise, by removing them from electron-rich groups to generate electrophiles, this enabling umpolung reactions, which can be capitalized to explore new synthetic routes.69_{. The later makes evident an important and unique advantage for electrochemistry: the flow}

of electric current can be used as an ON/OFF switch to trigger and stop reactions in the system.70

In the field of C-H activation, the exploration of electrochemical approaches is particularly appealing as an alternative to expensive stoichiometric oxidants as source of electrons. Besides being a more sustainable option, governing the energy of an electrochemical system with an applied oxidative potential, measured with a reference electrode, allows to overcome the limitation of relying on the intrinsic potential of a chemical oxidant. Thus, even when a reaction has a high activation barrier, it can be reached at ambient temperature, facilitating the generation of highly reactive species under mild conditions. These advantages have led to Numerous examples of electrocatalytic C-H activation, both using metals and metal free. 4,56,65,68–72_{The focus of this work will be on those}

including transition metals, to keep the discussion narrow and meaningful.

Palladium catalysts can be considered the pioneers in the use of electrocatalysis as novel methodologies, with the use of electro-oxidative Fujiwara-Moritani reactions, C-H oxygenations, methylations and halogenations.75,76_Since

then, examples using different metals. Such methodologies can be schematically classify as either direct or indirect electrocatalysis with respect to the metal centre. Furthermore, it’s also possible classify them depending on the prosed role of the applied potential on the reaction system.56,70,73,74

Among the direct electrolysis approaches, common roles that the transfer of electrons might have are: a) Electrochemical oxidation of the metal centre to generate the active catalyst.77,78

b) Anodic oxidation of the catalyst to a higher valent species to facilitate the reductive elimination step.79,80

c) Anodic oxidation after catalytic cycle to regenerate the active catalyst.81–83

d) Induce electron-transfer from metal-ligand complex to generate a radical, reactive species.84

In regards of the indirect electrolysis methodologies, among the common roles that the applied potential might have we find:

I. The use of a redox mediator as an electron shuttle, which trigger the reaction of interest after undergoing an electron transfer itself. These type of electrolysis can help to avoid issues such as overpotentials.85

II. The in-situ generation of electrophilic cations for further reaction with nucleophiles, sometimes known as “cation pool”.86,87

III. The electro-removal of the directing moiety in the molecule, with the assistance of an additive.88

IV. In-situ generation of a basic species, that participates in a base-assisted mechanism.

On this work, it will be indicated, when it is known, the role that the applied potential has in the catalytic cycle, by using the code corresponding to each case (a,b,c,d,I,II or III).

As it might be expected, in some cases more than one electrochemical step is involved, and more than one of the previously described scenarios are present.89,90_{To reduce ambiguity in identifying which processes are occurring, it}

is essential to carefully consider the redox potentials of the all the species involved, including the expected products. The applied potential of the working electrode, will determine which electron transfer process is thermodynamically feasible.91

(18)

18

Additionally, the selected design of the reactor, the electrochemical cell, will also impact the selectivity and efficiency of the transformation. The elements of the cell are represented in scheme 8, and they involve the substrate, an electrolyte, solvent, and at least two electrons: an anode and a cathode. Two different options for the set up are usually employed: undivided cells and divided cells (scheme 8). In undivided cells, the anode and the cathode are placed in the same compartment, thus oxidation and reduction occur within the same vessel. Although these cells are easier to build, they might not always be suitable since the substrate is exposed to all the species present in the reaction, which might lead to undesired side-reactions. On the other hand, divided cells reduce these interferences by making use of two compartments, one anodic and one cathodic, which are physically separated by a small, porous frit, that allows the transfer of charge while keeping the two half-reactions separated. This setup reduces the occurrence of undesired redox process of substrate or products, such as the cathodic reduction of high-valent metallic centres, but these benefits come with the price of a more complicated construction of the cell.72,91,92

The electrodes employed vary in different experiments, depending on the requirements of each experiment. In general, they are made from inert materials that allow electron transfer in solution. Some of these are; carbon-based materials, such as graphite, glassy carbon, reticulated vitreous carbon (RVC), or platinum. The RVC electrode exemplifies important qualities of electrodes, as it possesses a high surface area, high void volume, and chemical resistance. And advantage of using electrodes as components of the reaction system, is that they can simply be removed afterwards, meaning that there is no complicated purification procedure involved. Nevertheless, it is always important to consider the stability of the material of the electrodes under the reaction conditions.93_{The important}

characteristic of the reaction medium is that it must be able to conduct electricity. The supporting electrolyte is present to enable this behaviour. In regards of the solvents used, traditional organic solvents are common choice, but additionally, less conventional solvents, such as trifluoroacetic acid, trifluoroethanol, ionic liquids, and supercritical fluids have also received attention.94–96

Scheme 8: Representation of divided and undivided cells.70

Although not so common, another strategy reported to improve selectivity is the use of an electro-auxiliary species, which can be transiently bonded to the substrate to decrease its oxidation or reduction potential, thus minimizing secondary electrochemical reactions. In any case, it is informative to perform preparative electrolysis experiments, such as cyclic voltammetry, which enable the convenient assessment of the reaction system, and provide insights into the redox susceptible moieties present, and the magnitude of the free energy needed to promote certain electron transfer.70,72

Although it seems plausible that, in principle, most chemocatalytic reactions could be translated to their electrocatalytic equivalent, in doing so it should be considered that despite the relative rates of competing oxidation

(19)

19

reactions could prevail, a difference in the conditions, such as solvent, could affect considerably the catalyst performance and/or reaction outcomes, and some effort and experimentation will be required to adjust the reaction.72

Moving towards electrocatalytic systems represents a highly promising approach, with countless benefits, such as the easy separation by avoiding the formation of by products from chemical oxidants, or the possibilities of turning catalytic classic reactions such as the Shilov methane system. Nevertheless, many challenges remain to be tackled. To begging with, the difficulty to build the electrolysis apparatus and the outstanding number of variables might make the task seem overwhelming and unpredictable. We shall not let this perspective discourage us from digging deeper into this area; on the contrary, we can envision all the opportunities waiting to be explored, inviting us to embark with excitement into the construction and unveiling of this promising field.

3. Recent advancements in the field of C-H functionalization via transition metal

electrocatalysis: Metals and their reactivity.

In the following pages the current state of the field will be explored. Because many methodologies have already been discussed in previous reviews,56,70,73,74_{most of which are based on noble-metals, the main focus of this section}

will be in the detailed discussion of novel examples involving Earth-abundant metals iron, nickel, cobalt, copper and manganese. Furthermore, this focus is also coherent with the environmental benefits inherent to the implementation of an electrochemical approach towards C-H functionalization.

The first part of the following section will be devoted to the phosphorylation reaction, within the chemistry of manganese complexes, due to the abundance of reports evaluating this C-P bond formation, which allow to compare the performance of different transition metals for its catalysis, which will lead to more fruitful conclusions. Then, the remaining examples corresponding to the rest of the abundant metals, in order of increasing atomic number, will be reviewed in detail, while categorizing them depending on the role played by electricity within the specific reaction mechanism. In the case of cobalt, its methodologies that have been reviewed recently will only be presented schematically, in order to provide a broader context into its chemical behaviour in this type of transformations. To finalize the section, recently reported reactions, which have not been reviewed yet, catalyzed by noble metals, will only be presented schematically, without any in depth discussion, exclusively to provide a notion of the current general developments within the field.

In regards of the schemes presented, for most cases, the original numbering of the presented structures of each article will be kept the same as that presented by the authors of the paper. This implies that the numbering labels will be reinitiated in each subtopic. Additionally, some schemes are not modified and taken directly from their original source.

Navigating through the current research of the field will allow to contemplate it from higher ground, showing a clear, or clearer at least, panorama of the remaining challenges and interesting paths to be explored. A few words will be dedicated to this panorama.

3.1. Manganese and the phosphorylation reaction

Due to its redox dynamicity, manganese and its compounds have been part of the chemical toolbox as traditional oxidants for both or organic and inorganic substances. Their applications have been slowly covering more areas of chemistry, with special relevance in the field of free-radical cyclizations. In 1968 it was reported the use of Mn(III) acetate for the formation of 𝛾-lactones from olefins and acetic acid,97_{and more examples reported later, even}

showing viability for C-C bond formation,98_{and in recent years, its use for total synthesis.}99_{In most of these}

examples, the radical reactivity of manganese is made explicit and exploited. Some of the cyclizations even have used an electrochemical approach for the oxidation, which sometimes involved the radical activation of a C-H

(20)

20

bond usually by a single electron transfer, but being these examples more stoichiometric electrosynthesis than catalysis, we shall not discuss them in detail here.100–105

Manganese-based electrocatalytic methodologies for activation of inert bonds remains relatively unexplored, with only a few examples reported in recent years. Despite the scarcity of reports, the phosphorylation of aromatic compounds catalyzed by manganese has been devoted special attention by the group of Budnikova and coworkers.106,107_{The phosphorylation studies of the aforementioned research group have cover also other metals,}

namely cobalt,108_nickel,106,109–111_palladium,112,113_{and silver.}114,115_{Additionally, the some comparative studies}

including these metals in formation of C-P bonds , and using bimetallic catalytic systems have been reported.84,107,112,116,117_{In order to condense and extract essential ideas in C-P bond formation via C-H}

electrocatalytic functionalization, the different metals studied in this reaction will be discussed together and complementary in the following paragraphs.

3.1.1. Electrocatalytic C-P bond formation

The interest in developing improved, sustainable methodologies for the construction of aromatic phosphites arises from the valuable applications found for these compounds. Likely derived of their structural similarities to biomolecules, phosphorous containing scaffolds, such as aryl and heterocyclic phosphonates, have shown interesting medicinal and biological applications, beyond their use as pesticides, which might need to be reassess due to their neurotoxicity.118–120_{An important example a family with promising applications is coumarins, which have shown}

varied pharmacological activity, including cytotoxicity against human leukemia cells both in vitro and in vivo, exhibited by 3-phosphorylated coumarins.121_{Additionally, phosphorylated compounds have found applications in}

synthesis and catalysis; in reactions of addition to unsaturated compounds, or present as ligands in transition metal catalysts, for example.122

The most common methods for the synthesis of phosphorylated aromatic compounds include the phosphorylation of aryl halides;123_{radical approaches;}124,125_{nucleophilic additions;}126_{or the Arbuzov}127_{or Knoevenagel}128_reaction

for synthesis of coumarins. Up to recently, the most straightforward approach for the construction of these C(sp2

)-P bonds was the transition metal-catalyzed cross-coupling reactions, such as the methodology developed by T. Hiraro and coworkers in 1981.129_{Most of these methodologies present some drawbacks, such as; hard conditions,}

such as strong oxidants or elevated temperature; low yields and conversion of phosphite; many side products; or the need of prefunctionalization of substrates in cross-coupling reactions.123,126_{Some electrochemical methodologies,}

which are often milder and more atom efficient, have also been developed. Nevertheless, these synthetic approaches do not work when the aryl compounds are unsubstituted, and selectivity is often difficult to control in direct electrolysis.130

Most of the limitations mentioned above can be overcome employing transition metals for the electrocatalytic C-P bond formation, by activation of C-H bonds, as it was demonstrated by the studies of Budnikova and coworkers. The methodologies reported were performed under mild conditions: at room temperature and normal pressure, with catalytic amounts of the transition metal complexes, often in bimetallic systems, of around 1% with respect to the substrate, and in a 1:1 ratio of the aromatic and phosphite partners, under galvanostatic mode. The yields obtained are of up to 90%, with almost 100% conversion of the phosphite. Selectivity issues caused by the strong coordinating character of phosphorous can be partially addressed implementing divided electrochemical cells. Besides this, the implementation of indirect electrolysis contributes to higher selectivity, as the reaction can be performed at lower potentials than the oxidation potentials of the initial reactants, due to the presence of redox mediators. Additionally, the aromatic substrates employed do not require any special functionalization most of the time, neither as an activating functionality or as directing group. 84,117

Among the aromatic compounds that have been investigated, including some of their derivatives, are: benzene;131

courmarin;106,110_azole;113,132_{and pyridine.}109,111_{The phosphorylation of these compounds was performed using both}

monometallic catalysts, as well as bimetallic. The catalytic system often involves metal complexes and salts in the oxidation state II, M(II), which can further oxidize under electrochemical conditions, to yield M(III). Although

(21)

21

manganese is recurring in these activations, nickel and cobalt have also been relevant, and even palladium and silver have been included. Phosphorylation was carried using mainly diethyl phosphite as the phosphite source, with further extension of the scope to include alkyl groups such as isopropyl and butyl. The general reaction is presented in Scheme 9.

Scheme 9: Representation of metal-catalyzed electrochemical aromatic C-H phosphonation.

Because all the studied systems share important similarities, a detailed discussion of the reaction for the phosphonation of benzene will work as a model to understand the reactivity involve in these methodologies, allowing to later discuss the other cases more briefly, but still substantially. The common thread, that allows to assemble the results into a coherent story strongly relies on cyclic voltammetry experiments and the yields observed in each case. An important aspect, common to all substrates and catalysts employed, is the formation of metal-phosphites prior to the functionalization steps. This preceding chemical reaction is characterized by an increase in yields and selectivity after an induction period, of up to 1 day, and by an appearance of anodic peaks, which are not present in the isolated substrates, upon addition of the phosphites. In the system for oxidative phosphorylation of benzene (Scheme 10),131_{neither diethyl phosphite nor benzene exhibit anodic peaks under the reaction conditions,}

while unsaturated catalyst CoCl2bpy undergoes one-electron oxidation at 1.34 V. Addition of diethyl phosphite

induces interesting changes in the CV, such as the appearance of a new quasi-reversible anodic peak appears at 0.54 V. This is attributed to the metal-phosphite complex, and its appearance at lower potentials could be explained by the electron donor properties of the phosphite, which facilitate oxidation. When benzene is incorporated to the system, the quasi-reversible anodic peak is shifted to 0.70 V, and the oxidation current associated to this peak also increases upon further addition of both substrates (Scheme 11). The important changes observed in the CV curves, related to the redox properties of the complexes, might relate to the strong coordinating properties of the diethyl phosphite.

(22)

22

Scheme 11:CVA curves of complex CoCl2 (5x10-4 mol L.1) in MeCN; Et4NBF4, reference electrode Ag/AgCl. a) In the presence of

increasing amounts of HP(O)(OET)2 with the ratio CoCl2bpy : HP(O)(OEt)2 = 1:0 (1), 1:1 (2), 1:2 (3), 1:3 (4), 1:6 (5), 1:9 (6), and 1:27

(7). b) In presence of increasing amounts of HP(O)(OET)2 and benzene, with ratio CoCl2bpy : HP(O)(OEt)2 : C6H6 = 1:0:0 (1), 1:1:1 (2),

1:2:2 (3), 1:3:3 (4), 1:5:5 (6), and 1:6:6 (7).

The ratio of the oxidation current in the presence of benzene and phosphite, the catalytic current, icat, withrespect

to the current of the first oxidation peak in absence of benzene, idif, is used as a parameter to measure the catalytic

activity in the system. In the reaction catalyzed by CoCl2bpy, the maximum value of icat/idif is equal to 2.07, and is

reached when the proportion of CoCl2bpy:HP(O)(OEt)2:C6H6 is equal to 1:6:6. After this value, addition of

phosphite and benzene do not cause an increase of the current. This might indicate that the rate of regeneration of the catalyst becomes the rate limiting step of the catalytic cycle after the addition of the 6 equivalents of substrates. After this point, further addition of substrates, without an increase of the catalyst, might even be detrimental for the rate of the process, by interfering with diffusion processes.

The electrochemical behaviour of other catalytic systems, namely CoCl2dmphen and Ni(BF4)2 is similar to that of

CoCl2dmphen. Addition of benzene to the catalyst solution does not produce any changes in the redox properties,

but addition of the phosphite leads to the appearance of a new quasi-reversible anodic peak. The identity of the metal and the electronic properties of the ligands modulate the potential at which the oxidation process occurs, but in all cases the peak associated to the metal-phosphonate complex is at lower potentials than that of the isolated catalysts. The first oxidation peak for NiBF4dmphen shifts from 0.90 V to 0.75 V when adding the phosphite.

Mn compounds, such as MnCl2bpy, are special cases, as their CV waves display a different behavior from the rest

of the metals. Addition of phosphite to the solution of the metal leads to an increase in the catalytic current of its irreversible one-electron oxidation wave, which also shifts from 1.14 V to 0.99 V (vs Ag/AgCl), but addition of benzene does not cause any enhancement on the catalytic current, nor other changes in the redox properties, suggesting that no significant phosphonation takes place. Still, the addition of manganese as cocatalyst for other metals leads to an enhancement in their performance, terms of higher catalytic current and lower catalyst loading. Insights into this behaviour are given by spectroelectrochemical studies. EPR spectroelectrochemistry studies have shown that manganese compounds, presumably in the anionic form of [MnCl4]2- in acetonitrile, are poorly soluble

and remain largely in precipitate, and the registered EPR X-band is smaller than expected, which is explained in terms of the formation of Mn(II) dimers, or polymers. The addition of phosphite cause an increase in the solubility of the manganese compounds, as the addition of acetic acid does too, and the addition of Ni(BF4)2 to the [MnCl4]

2-solution causes its ESR to disappear, likely because of the decrease formation of dimers and polymers.

This might suggest that the role of manganese compounds in the bimetallic systems is focused in enhance the regeneration of the catalyst, thus improving the overall kinetics and the yields obtained. For this reason, most of the phosphonations of aromatic compounds studied employed bimetallic systems as the optimal conditions.132_{This is}

also reflected in a decrease in the formation of byproducts, probably because the reaction leading to the desired product takes place faster, translating the kinetic advantage into greater selectivity, as reactions with only one metal often yielded also di- and triphosphited products. Such is the case of reactions catalyzed with Ni(BF4)2, which

(23)

23

presented the highest activity on its own, but also led to polyphosphonation. The maximum yields obtained in the phosphonation of benzene with monometallic systems were of no more than 30%, while for bimetallic system such as MnCl2/CoCl2bpy the yield was of up to 90%. An increase in steric hindrance in the catalyst might prove benefical

to lower the recurrence of polyphosphonation, by favouring smaller substrates. Likewise,as increasing the electronic saturation of the metal centre, might prevent the formation of metalphosphite, although these two hypothesis would need to be tested.

Additionally, the mathematical benefits of electroanalytical techniques, such as cyclic voltammetry, to provide solid numerical data, which can be used to unravel details of the mechanistic and kinetic workings of the systems, was exploited. In this case, it allowed to study the turnover frequency, TOF, of the metallic systems, by increasing the amounts of the phosphite partner to the reaction mixture, until the ration of the icat/ipwas constant. This was done

for monometallic, as well as bimetallic systems. Although not in all cases the TOF could be calculated, sometimes due to the appearance of non-catalytic new waves, synergistic effect of bimetallic systems was evident, as can be appreciated in (Table 1).84

Table 1: Catalytic growth of the current at the first oxidation peak, and TOF of catalytic systems used in the phosphorylation reactions. [Mbpy]=5 x 10-3_{mol L}-1_{(1.7 mol L}-1 _{for [Mnbpy]).}

Catalytic System Ratio

[Complex:Diethyl phopshite] ic /ip (v=0.1 V/s) TOF (s-1₎ CoCl2 : HP(O)(OEt)2 1:144 - -

Ni(BF4)2bpy: HP(O)(OEt)2 1:72 1.3 160

MnCl2bpy: HP(O)(OEt)2 1:196 3.6 355

MnCl2bpy/Ni(BF4)2bpy: HP(O)(OEt)2 1:180 6.1 690

MnCl2bpy/CoCl2bpy: HP(O)(OEt)2 1:180 - -

CoCl2bpy/Ni(BF4)2bpy: HP(O)(OEt)2 1:24 2.7 721

The observations of the induction period in all transformations, together with the electrochemical studies, spectroscopic measurements and the optimal reaction conditions led to the proposal of a mechanism of reaction that could dominate these catalytic systems (Scheme 12). It is proposed that the first step is the formation of metalphosphites with both metals, which is followed by a two-electron anodic oxidation of the complex. It is likely that is kinetically favoured to extract one electron from each metallic centre, rather than two from the same one, and this might be linked to the faster regeneration of the catalyst in bimetallic systems. An intramolecular electron transfer from the phosphite units to the metallic centres yields then two radical phosphites. The phosphite is then eliminated into the solution, followed by its deprotonation, which might be aided by acetate in the solution. Then, the phosphorous-centred radical would attack the benzene ring, leading to the loss of a proton and an electron, to yield the final phosphited benzene product. Meanwhile, in the cathode the process of proton reduction to afford hydrogen takes place. Hydrogen evolution could be enhanced with the use of Pt cathodes, as they tend to facilitate this process. The proposed mechanism would belong to the d) category.