Design Principles in Frustrated Lewis Pair

Catalysis for the Functionalization of Carbon

Dioxide and Heterocycles

Frédéric-Georges Fontaine*a, Marc-André Courtemanchea, Marc-André Légaréa ,

Étienne Rochettea

a Département de Chimie, Centre de Catalyse et Chimie Verte (C3V), Université Laval,

1045 Avenue de la Médecine, Québec (Québec), Canada, G1V 0A6

Email : frederic.fontaine@chm.ulaval.ca

This is the peer reviewed version of the following article: [Design Principles in Frustrated Lewis Pair

Catalysis for the Functionalization of Carbon Dioxide and Heterocycles, Coord. Chem. Rev. 2017,

334, 124 - 135.], which has been published in final form at [10.1016/j.ccr.2016.05.005].

Keywords

Frustrated Lewis pairs, Ambiphilic molecules, Boron, Aluminum, Carbon dioxide,

Methanol, C-H activation.

Highlights

1. Phosphine-aluminum frustrated Lewis pairs (FLPs) can activate CO2 but show

significant stability issues.

2. The Lewis acidity and basicity need to be fine-tuned to the desired reactivity for

efficient FLP catalysis.

3. FLP catalysts need to generate strong hydrides to reduce catalytically carbon

dioxide.

4. Transition metal catalysis can serve as an inspiration for the design of metal-free

catalysts for many transformations, including the borylation of heteroarenes.

Graphical abstract

Abstract

This account describes our work on the use of ambiphilic molecules as catalysts for the

reduction of carbon dioxide. Starting with the discovery that aluminum ambiphilic

species (Me2PCH2AlMe2)2 (1) and Al(C6H4-PPh3)3 (5) could coordinate CO2 but showed

significant instability, we found that phosphinoborane Ph2P-2-Bcat-C6H4 (7) was an

exceptional catalyst for the hydroboration of CO2 to methoxyboranes, a molecule that

can be readily hydrolyzed to methanol. A mechanistic investigation allowed us to outline

the similarities between the catalytic activity of frustrated Lewis pairs (FLPs) and that of

transition metal complexes. We were able to extrapolate four important guidelines,

described in this account, to synthesize efficient metal-free catalysts. We demonstrate

that using these concepts it was possible to rationally design FLP catalysts for the C-H

borylation of heteroarenes.

Table of Content

1. Introduction

2. Ambiphilic molecules: cooperative Lewis pairs

3. Ambiphilic molecules as catalysts for the reduction of carbon dioxide

4. Investigating the role of phosphino-boranes in the catalytic reduction of CO2

5. Rationalizing the catalytic activity of ambiphilic molecules

6. Developing a system for the metal-free hydrogenation of CO2

7. Guidelines for the design of FLP catalysts

8. C-H borylation of heteroarenes: a proof of concept

9. Conclusion

10. References

1. Introduction

Decades of investigation in organometallic chemistry has allowed the development of

powerful catalysts that are now indispensable tools in all areas of synthetic chemistry.

Most catalytic processes employing transition metals rely on fundamental steps involving

electron transfer reactions between the metal center, the ligands in the coordination

sphere and the reagents. It is well accepted that noble metals are amongst the most

efficient elements for the design of highly efficient catalysts since they easily perform 2-

electron redox transformations such as oxidative addition and reductive elimination

reactions. Although quite efficient in many processes, these metals are scarce and

expensive. In addition, their known toxicity causes concerns and as such, the metals often

need to be separated from the end-products, resulting in higher production costs. In the

past decade, much efforts have been invested in finding base-metal catalysts to replace

second- and third-row transition metal complexes. However, the propensity of first-row

transition metals to effect one-electron redox reactions makes them less likely to undergo

2-electron transformations typical of their heavier analogues. In order to circumvent this

problem, several elegant strategies have been developed, notably using cooperativity

between the metal and the redox-active ligands [1-5].

Cooperativity between two active sites is also a dominant aspect of frustrated Lewis pair

(FLP) chemistry, a concept first introduced by Douglas W. Stephan in order to explain

the splitting of the dihydrogen molecule by a sterically congested Lewis acid and base

pair [6]. Although some previous observations already suggested that cooperativity

between Lewis pairs could promote metal-free transformations, notably for

hydrosilylation [7], it is the use of FLPs for molecular hydrogen activation that led to a

surge in research in this field [8-9]. Since, FLPs have been shown to activate many small

molecules in addition to being active catalysts in reductive transformations [10-12]. The

activation of substrates like H2 by transition metal complexes occurs generally by the

transfer of electrons from the bond to be cleaved to the LUMO of the metal complex

while electrons from the HOMO, either associated with the d orbitals or a ligand in the

coordination sphere, are transferred to the anti-bonding orbital (Scheme 1A). Similarly, in

FLP chemistry, the high-lying HOMO of the Lewis base can transfer electrons to an anti-

bonding orbital of the E-E’ bond while the electrons in the bonding orbital are

simultaneously transferred to the low-lying LUMO of the Lewis acid (Scheme 1B). In

that sense, frustrated Lewis pairs can be seen as main group counterparts of noble metals

in which 2-electron transformations can take place.

Scheme 1. Representation of the similarity between A) the cleavage of substrate E’-E by

a transition metal complex and B) the activation of substrate E’-E by a frustrated Lewis

pair.

This account of our work will describe our pursuit of a better understanding of frustrated

Lewis pairs in the context of CO2 reduction catalysis. Using instinctive concepts from the

organometallic chemist toolbox, we were able to discover metal-free systems effecting

the hydroboration, hydrosilylation and hydrogenation of CO2. We also demonstrate that

the developed concepts can be further extended to other important catalytic

transformations, such as the functionalization of C-H bonds.

2. Ambiphilic molecules: cooperative Lewis pairs

Before the advent of FLP chemistry, our interest in the use of ambiphilic molecules,

which can be described as intramolecular Lewis pairs, has first been motivated by their

use as ligands for transition metals. While investigating the role of methylaluminoxanes

(MAO) in the homologation of hydrosilanes with Zargarian, we were able to demonstrate

that species (Me2AlCH2PMe2)2 (1) could react with nickel(II) precursors to generate an

extremely active catalyst for the oligomerization of phenylsilane (Scheme 2A) [13].

Later, it was found that 1 coordinates to Cp*Rh(III) species and that the Lewis acidic

alkylalane moiety could reversibly ionize the metal center and lead to interesting

reactivity, including the stabilization of putative ethylene-AlR3 interactions (Scheme 2B)

[14-15] In these systems, the coordination of the Me2PCH2AlMe2 moiety was possible

even if 1 is known to be a dimeric species both in solution and in the solid state. This

behavior suggested that some FLP-type chemistry is possible even in the absence of any

evident “frustrated” character.

Scheme 2. Reactivity of ambiphilic molecule (Me2AlCH2PMe2)2 (1) with A)

(Indenyl)nickel(II)(PPh3)(Me) and B) Cp*Rh(III)(DMSO)(Me)2.

Scheme 3. Trapping of interrupted Nazarov cyclization using (Me2AlCH2PMe2)2 (1)

Entering the field of FLP chemistry in 2011, we reported that species 1 could trap

intermediates in the Nazarov cyclization [16] (Scheme 3) and could sequester carbon

dioxide to generate adduct 2 (Scheme 4) [17]. The CO2 adduct thus generated is

structurally similar to other reported FLP adducts, with the basic phosphine interacting

with the electrophilic carbon and the Lewis acid bound to oxygen [18]. It was

demonstrated by Uhl, not long after our original report, that bulkier ambiphilic species

(tBu2PCH2AlMe2)2 exhibits similar reactivity [19]. In contrast with the bulkier tBu2P-

and Ph2P- systems, the Me2P- moiety in species 1 was small enough to allow a second

addition of the ambiphilic ligand to CO2, generating an unusual spirocyclic species (3,

Scheme 4). A similar spirocyclic CO2 adduct was later reported by Stephan [20]. Species

2 was found to be thermally unstable since the cleavage of the CH2-Al bond rapidly

occurs, leading to stable aluminum carboxylate species 4, as observed in Scheme 4.

Another important observation was that the CO2 activation products 1 to 4 could be

observed by MAS 31P{1H} solid state NMR spectroscopy when a solid sample of 1 was

exposed to an atmosphere of CO2 for a period of 24 hours.

Scheme 4. Trapping of CO2 by (Me2AlCH2PMe2)2 (1)

Although the exact mechanism for this transformation remains unknown, the ability of

the aluminum center to access hyper-coordination seems to be key for the activation of

CO2. Since aluminum species are kinetically labile, the reactivity is dictated by the

thermodynamic equilibrium, favoring the formation of strong Al-O interactions.

3. Ambiphilic molecules as catalysts for the reduction of carbon dioxide

Since the first report of a FLP CO2 adduct by Stephan and Erker in 2009 [21], several

other FLPs have been reported to sequester CO2 [22-34]. Although the reduction of CO2

has been shown to be possible in some instances [22, 35-36], no FLP system was able to

perform efficient and selective catalytic transformations until our original report that

demonstrated the catalytic activity of 1-Ph2P-2-Bcat-C6H4 (7, cat = catechol) for the

hydroboration of CO2 to methoxyboranes, species that can be readily hydrolysed to

methanol [37].

The reason for the lack of efficient catalytic transformations using FLPs was rationalized

by the fact that only strong acids and bases were employed in FLP systems, leading to the

formation of thermodynamically stable adducts of CO2. As such, the binding of a strong

Lewis acid to the oxygen atom of CO2 is thought to limit the release of reaction

intermediates. As a matter of fact, the reduction of CO2 to methanol derivatives is a 6-

electron reduction process that involves the successive formation of formate and acetal

intermediates (Scheme 5). The binding of these species by strong and oxophilic Lewis

acids would lead to thermodynamic sinks that would prevent rapid catalysis.

Simultaneously, the stabilization of the empty orbital on carbon by a Lewis basic lone

pair would result in a kinetically stable system in which no orbital is available on CO2 for

a nucleophilic attack from a hydride. We hypothesized that an efficient way to promote

catalysis was to activate the CO2 molecule without preventing a nucleophilic attack on

the carbon atom while increasing simultaneously the nucleophilicity of the reducing

agent. Instead of focusing solely on the binding of CO2 itself, we targeted the

simultaneous activation of both CO2 and the reductant.

Scheme 5. Typical chemical reduction of CO2 using hydroboranes (HE = H-BR2),

hydrosilanes (HE = H-SiR3) or molecular hydrogen (HE = H-H).

Building on our previously reported aluminum-phosphorus based systems, we targeted

tripodal ambiphilic species Al(C6H4-PPh3)3 (Scheme 6, 5) [38]. In addition to the

robustness of the aryl framework compared to the reactive methylene, we were hoping

that weakly Lewis basic arylphosphines would enable a reversible binding of CO2, thus

enabling catalysis. Another key element in the design was the presence of multiple Lewis

basic groups, granting the ability to simultaneously bind a reductant and increase its

nucleophilicity.

Scheme 6. Reactivity of Al(C6H4-PPh2)3 (5) with CO2 and HBcat to generate species 6

and 7.

As expected, species 5 reversibly coordinates CO2 (Scheme 6). While it enabled the

catalytic reduction of CO2 to CH3OBcat in the presence of HBcat as a reducing agent, the

aluminum compound was found to decompose to unusual aluminum boronate species 6

during the process (Scheme 6), underlining the limitations of reactive and Lewis acidic

aluminum in CO2 reduction chemistry. It was found, however, that phosphine-borane 1-

Ph2P-2-Bcat-C6H4 (7) was also generated in the reaction conditions (Scheme 6). This

family of phosphinoboranes is well known and several derivatives were previously

reported by Bourissou and coworkers [39-40]. The independent synthesis of 7 did show

that this species does not react with CO2 nor with catecholborane (Scheme 7). Indeed,

since the Lewis acidity and the Lewis basicity of this FLP is relatively weak, the

formation of a CO2 adduct was calculated to be endergonic by 9.9 kcal.mol-1. As such, in

contrast with previously reported systems, addition of CO2 to this mild FLP system does

not result in any thermodynamic stabilization. Yet, mixing CO2, the hydroborane and 5

together results in the rapid and catalytic formation of methoxyboranes. It was also found

that the hydrogen rich BH3.SMe2 could be used as a reducing agent with record-breaking

TOF values of 853 h-1 at 70 °C (Scheme 8) [37].

Scheme 7. Synthesis of phosphinoborane (7) and the absence of reactivity observed in

presence of CO2 and HBcat.

Scheme 8. Catalytic hydroboration of CO2 to methoxyboranes catalyzed by

phosphinoborane 7 (HBR2 = HBcat, HBpin and BH3.SMe2).

4. Investigating the role of phosphinoboranes in the catalytic reduction

of CO2

A mechanistic investigation was carried out in order to better understand the efficiency of

species 7 as a catalyst for the hydroboration of CO2 [41]. Using a computational study

based on DFT, we were able to demonstrate that the hydroboration of the CO2 adduct of

7 following a classical pathway would have a very high barrier of activation of 55.7

kcal.mol-1 (Figure 1A). As explained above, the coordination of the Lewis base of the

FLP to the electrophilic carbon of CO2 hinders the reduction process. While several

possible transition states were located for this crucial step, the most favorable was

calculated to have a much lower barrier of 38.3 kcal.mol-1, involving the activation of the

hydroborane by the phosphine with simultaneous activation of one oxygen of the CO2 by

the weak Lewis acid (Figure 1B). In this arrangement, no stabilization of the electrophilic

carbon was observed and this site was therefore available for an attack from the hydride

of the activated-hydroborane. Interestingly, although the formylborane and formaldehyde

produced in the reaction can be reduced in the absence of the catalyst, species 7 was

found to be an active catalyst for the reduction of the reaction intermediates to ultimately

generate MeOBR2, species which can readily be hydrolysed to methanol.

Figure 1. Calculated transition states for the hydroboration of CO2 using catalyst 7 using

a A) classical hydroboration pathway and B) assisted by the activation of the reducing

agent. [B] = Bcat.

Although the reduction of CO2 using 7 is thought to be the initial step in the catalytic

process, the energy barrier that was calculated for the ambiphilic mechanism was too

high to be consistent with the exceptional catalytic activity that is experimentally

measured for this system. This led us to conduct an in-depth experimental investigation

on the role of the Lewis acid and the Lewis base components of these ambiphilic

molecules, in collaboration with the Bourissou group [42]. We discovered that the nature

of the Lewis base had only a minor influence on catalytic activity and that the original

Bcat as the Lewis acid component of the FLP catalyst was the most efficient in addition

of being the most thermodynamically stable. It was indeed shown that in presence of

HBcat, various phosphinoborane ambiphilic catalysts would rearrange to 7 (Scheme 9).

The most important finding however, was that the aldehyde adduct 8, which was initially

thought to be the resting state of the catalyst, was rather the catalyst itself. Indeed, a

labelling experiment on 8 demonstrated that the aldehyde that is first generated by the

reduction of CO2 by 7 remains on the Lewis pair throughout catalysis (Scheme 10).

Consequently, no 13C incorporation into methoxy products was observed when using the

H213C=O adduct of the catalyst 7 (8*). Similarly, 8* is not generated from the unlabelled

catalyst 8 during the catalytic reduction of 13CO2.

Scheme 9. Reaction of phosphinoboranes 1-PPh2-2-BR2-C6H4 (R2 = (OMe)2, -OC(Me)2-

C(Me)2-O-, -OCH2C(Me)2CH2O-) with HBcat leading to 7.

Scheme 10. Reduction of CO2 with HBcat catalyzed by aldehyde adduct 8.

A DFT investigation of the reduction of CO2 using 8 allowed us to find three possible

transition states that are all within 1 kcal.mol-1 of each other (Figure 2). These three

transition states are operating using a similar pattern of activating simultaneously both the

CO2 and the reducing agent, underlining the validity of this approach. More precisely,

one of the –OR group on the borate moiety, either from the aldehyde bridge or the

catechol group, dissociates freeing a coordination site for the formation of an adduct with

the reducing agent present in the system. This activation of the reducing agent

significantly increases the reducing power of the hydride. Simultaneously, a weak Lewis

acid, either the borane or the acidic C-H bond of the aldehyde bridge stabilizes the bent-

CO2 conformation.

Figure 2. Calculated transition states for the hydroboration of CO2 using catalyst 8.

5. Rationalizing the catalytic activity of ambiphilic molecules

The transition states that are proposed to explain the catalytic activity of species 8

involves two important components: the generation of a highly nucleophilic hydride

while simultaneously stabilizing the bent form of CO2 by weak Lewis acidic moieties.

Such interactions have been proposed to be crucial for some of the most active

organometallic catalysts for CO2 reduction. The most notable examples are the iridium

pincer complex A reported by Hazari [43] for the hydrogenation of CO2 to sodium

formate and the nickel cyclam electrocatalyst B reported by Sauvage [44-46] for the

reduction of CO2 to CO (Chart 1). One can speculate that direct comparisons can be made

between the most efficient FLP and transition metal catalysts. Therefore, one can also be

inspired by organometallic chemistry when designing metal-free catalytic systems. As

expressed by Guan and co-authors [47-48] on the high activity of pincer nickel species C

(Chart 1) for the hydroboration of CO2, “Insertion of CO2 into a metal-hydrogen bond

constitutes the critical step in many transition-metal-catalyzed reductions of CO2”. This

statement translates to the need of generating a strongly nucleophilic hydride in order to

attack the electrophilic carbon and effect the reduction. This was further illustrated by

Linehan and coworkers [49-50] who reported a highly active cobalt hydride catalyst D

(Chart 2) by tuning the thermodynamics of the system.

Chart 1. Some active transition metal catalysts for the reduction of carbon dioxide.

All these systems most likely operate by generating a strongly reducing hydride that is in

turn capable of CO2 reduction. As it was recently underlined by Cummins, BH4- salts

have been known to readily effect the reduction of CO2 to formatoborate [51]. However,

the possibility of catalytically generating hydrides and drive the reduction all the way to

methoxides engenders a resurgence of interest for these concepts as they could potentially

be extrapolated to CO2 hydrogenation strategies. The first example of metal-free catalytic

reduction of carbon dioxide was reported by Ying and coworkers in 2009 when they

demonstrated that N-heterocyclic carbenes can catalyze the hydrosilylation of CO2 to

methoxysilanes [52]. While the mechanism is debated [53-54], recently published

computations suggest that the NHC-CO2 adduct serves as a Lewis base in order to

activate silanes by hyper-coordination, thus favoring hydride transfer [55]. In that regard,

we also reported that phosphazenes could act as efficient catalysts for CO2

hydrosilylation, selectively forming either formatosilanes or methoxysilanes depending

on the reaction conditions (Scheme 11) [56]. Surprisingly, we found that in DMF

solution, the reduction of CO2 to methoxysilanes occurred readily in the absence of any

catalyst. Indeed, DMF and other strongly donating solvents were shown to promote the

formation of hyper-coordinate silanes [57-58]. It is rather surprising that such reactivity

was not reported before, since NHC catalyzed reductions were exclusively carried out in

DMF.

Scheme 11. Hydrosilylation of CO2 catalyzed by phosphazene.

Using the same principle of generating a strongly reducing hydride, we reported that

proton sponge 9 could react with BH3.SMe2, leading to a disproportionation reaction

forming the ion pair [9.BH2]+[BH4]

- (Scheme 12) [59]. Interestingly, the catalytically

generated tetrahydroborate anion can reduce CO2 to the formatoborate intermediate

whereas larger boron clusters would not. The subsequent reduction steps to generate

methoxyboranes are readily achieved by BH3.SMe2. In the past few years, a number of

Lewis base were shown to promote CO2 reduction based on the same principles. The

known catalysts include phosphines [60], guanidines [61], diazafluorenides [62],

carbenoids [63], and even BH4-[64]. However, there is only limited industrial viability in

using hydrosilanes or hydroboranes for transforming CO2 into methanol and the most

promising reductant for commercialization of a CO2 to methanol process is molecular

hydrogen [65].

Scheme 12. Hydroboration of CO2 catalyzed by proton sponge 9.

6. Developing a system for the metal-free hydrogenation of CO2

The hydrogenation of CO2 using metal-free catalysts is an important challenge. Although

many hydrogenation catalysts based on FLP frameworks have been reported, the efficient

reduction of ketones has only been recently achieved using very acidic B(C6F5)3 and

ethers as Lewis bases [66-67]. Simply put, the splitting of hydrogen by FLPs can be best

described as the heterolytic cleavage of the H2 molecule into a proton and a hydride

fragment. However, a fact that is often overlooked is that the protic/hydridic characters

are going to be directly related to the strength of the Lewis acid and base used in this

system. For instance, the splitting of H2 by classical FLP systems using B(C6F5)3

generates the highly stable [H-B(C6F5)3]-. With weaker bases, the net result is that a

reactive proton and a stable hydride are generated. Such a scenario is ideal for the

hydrogenation of polar basic substrates such as imines that are readily protonated.

However, in order to protonate the more weakly basic ketones, more acidic protons are

needed, thereby explaining the necessity of having very weak Lewis bases like ethers for

such processes to occur. Of course, the often rate-limiting hydride delivery step can be

promoted thermally, but often at the cost of catalyst stability.

From this point of view, it is clear that strongly acidic FLPs are not particularly well

suited for CO2 reduction since this molecule is known to be a Lewis acid and an

electrophile. Therefore, for better catalysis, the FLP should be able to generate more

reactive hydrides. This has been elegantly demonstrated in a theoretical study by

Musgrave and coworkers in which they modeled hydrogen transfer to CO2 from NH3BH3

and demonstrated that the hydride transfer was key for the reaction to take place [68].

Yet, very few reports have focused on generating strong hydrides using FLPs [69-70].

We thus hypothesized that an ideal CO2 hydrogenation system should a) favor H2

splitting over CO2 coordination to avoid thermodynamic stabilization; b) generate a

reactive hydride by using a weak Lewis acid and c) consist of an intramolecular FLP for

obvious entropic considerations.

In order to find the right ambiphilic system for the hydrogenation of CO2, calculations

were carried out using the more common components of FLPs (P, N, Al and B) on an

aromatic framework. As shown in Figure 3, it can be seen that in all cases, except for the

aminoborane 10NB, the formation of CO2 is favoured over the hydrogen activation.

Indeed, in this system the activation of H2 was found to require 12.2 kcal.mol-1, whereas

the CO2 adduct lies higher in energy, at 21.1 kcal.mol-1. For the three other FLP

examined, the CO2 adducts are more stable in energy than the activation of H2. These

results suggest that aminoboranes are the best suited for the catalytic hydrogenation of

CO2 [71].

Figure 3. Thermodynamics of H2 splitting and CO2 binding by a variety of aryl bridged FLP

systems. Energies are reported in kcal.mol-1 (DFT study, WB97XD/6-31++G**, solvent=benzene,

SMD).

Whereas many intramolecular aminoboranes have been reported and that some of them

have been used to catalyze hydrogenation reactions [72-80], none were reported to effect

the hydrogenation of carbon dioxide. It should be noted that an important side reaction in

the activation of H2 using 1-NR2-2-B(C6F5)2-C6H4 is the protodeborylation reaction

where the acidic proton on nitrogen once H2 is cleaved is transferred to one of the

pentafluoroaryl groups to generate C6F5H (Scheme 13). Repo and coworkers cleverly

exploited this transformation in the selective cis-hydrogenation of alkynes [78].

Scheme 13. Typical protodeborylation reaction.

In order to generate more reactive hydrides, we investigated, in collaboration with the

Stephan group, the reactivity of aminoborane species 1-NMe2-2-BMes2-C6H4 (Mes =

2,4,6-Me3-C6H4 (11a) or 2,4,5-Me3-C6H4 (11b)) for the reduction of CO2 [81]. It was

demonstrated that in the presence of CO2 and H2, the hydrogenation of CO2 occurred,

generating formyl, acetal and methoxyboranes (Scheme 14). Although no catalytic

activity was observed, up to 1.8 equivalents of hydrogen were consumed by species 11a

when carrying the reaction at 80 °C for 9 days. With less bulky 11b, the reaction could be

carried over 24 hours at ambient temperature, or at 80 °C over a period of 3 hours.

Interesting comparisons can be made with the only other known metal-free CO2

hydrogenation system reported by Ashley and O’Hare who reported that TMP/B(C6F5)3

generated 25% of methanol after 6 days of heating at 160 °C [35].

Scheme 14. Hydrogenation of CO2 using ambiphilic aminoboranes 11a and 11b.

When looking at the transition state of the hydrogenation reaction of CO2, a typical

mechanism related to Noyori-type reduction system was located (Figure 4) [82-83].

Indeed, a transfer of the hydride on the carbon occurs in concert to an interaction between

the acidic proton and the oxygen atom, leading to a significantly lower transition state,

once again demonstrating that for outer-sphere transformations, FLP catalysts could

potentially be used to replace transition metal systems. While an efficient catalyst for the

metal-free hydrogenation of CO2 remains elusive, these results suggest that careful design

could lead to the discovery of such a catalyst.

Figure 4. Calculated transition states for the hydrogenation of CO2 using catalyst 11b.

7. Guidelines for the design of FLP catalysts

It has often been demonstrated that ligand design can drastically alter the activity of

organometallic catalysts, mainly by tuning the steric and electronic properties of the

metal centers or by enabling outer-sphere transformations. As we have demonstrated in

our study of the reduction of CO2 with FLPs, a similar state-of-mind can be used in order

to develop metal-free catalysts, but by adjusting the steric and electronic properties of the

Lewis acid and base pair. The important key points to keep in mind are as follow.

1- “Frustration” in Lewis pairs gives a kinetic advantage

It is well known that Lewis acids and bases can form very strong adducts. However, in

some cases the formation of adducts can be prevented by steric or geometric constraints

around the Lewis base and the Lewis acid. The respective HOMO and LUMO orbitals

therefore remain readily available for bond activation. Figure 5 schematically illustrates

the thermodynamic profile of the reactions of associated and ‘frustrated’ Lewis pairs with

a substrate. When looking to overcome the energy needed to reach a desired transition

state, a FLP will have a significant advantage over a classical Lewis pair because it does

not form a low energy and thermodynamically stable adduct. In some instances, like in

the activation of H2, the advantage of the frustrated character is important, since it brings

a significant kinetic advantage by allowing the substrates to approach the Lewis pair.

However, in other instances like for the activation of CO2 where the cleavage of the π

bond is relatively easy and the E-O bonds generated are strong, no frustrated character is

needed if hyper-coordination or equilibrium to an opened form can make the activation of

the substrate kinetically accessible. A direct analogy can be made to transition metal

chemistry where the bond activation of a substrate takes place easily on a coordinatively

unsaturated complex, but can still take place for a saturated complex if an associative

substitution of a weak ligand can occur.

Figure 5. Representation of a thermodynamic profile in order for a Lewis pair (L

+Z) to reach a transition state (TS). A) In the case of a FLP, the Lewis pair does not

associate and is close by ΔGLZ to the TS compared to a classical Lewis pair (B).

2- The Lewis acidity and basicity of the catalyst should be fine-tuned for the intended

reaction

Pioneering early work on the activation of substrates with frustrated Lewis pairs was

concentrated on using strong acid and bases with a focus on characterizing bond cleavage

and substrate binding products. While the use of such molecules is particularly well

suited for the stoichiometric binding of small molecules, their applicability to catalytic

transformations is not evident. Indeed, strong Lewis acids are often incompatible with a

number of functional groups and, more fundamentally, can even inhibit the desired

reaction. As a matter of fact, while small molecule binding by FLPs is often referred to

as ‘‘activation’’, the formation of overly strong FLP-substrate complexes can result in

significant thermodynamic stabilization and as such, the small molecules are actually

‘‘de-activated’’. As represented in Figure 6, the transition state of a given transformation

can presumably be more difficult to reach when using strong Lewis pairs to bind a

substrate because of the presence of a thermodynamic well. No such inhibition of

reactivity should be observed in the case of mild, perfectly tuned acid and base

combinations.

Figure 6. Representation of a thermodynamic profile of a substrate (S) activated by a

strong (s) or weak (w) Lewis pair (L + Z) to reach a desired transition state (TS).

3- The polarity of the activated E-H bond can be modified by tuning the Lewis acidity and

basicity.

As demonstrated above, the presence of strong Lewis acids and weak Lewis bases will

lead to an acidic proton by stabilization of the hydride, which has been most of the focus

in the first decade of FLP design and reactivity. However, strong Lewis bases in the

presence of weak Lewis acids will lead to a significant hydride character of the FLP

which will open a whole new set of reactivity yet unexplored that could be of use in

many catalytic transformations. Such specific design can be extrapolated to the activation

of many bonds, such as CO2 as demonstrated in earlier sections. Therefore, the Lewis

pair to be used should be adapted to the reaction of choice. However, in many stances,

the use of intramolecular FLP will bring in a significant entropic advantage over

bimolecular FLPs.

4- The reactivity of transition metal complexes can inspire metal-free catalysts

Many transition metal catalysts operate by heterolytic cleavage reactions where ligands

play a crucial role in the transfer of electrons. Obviously, in concerted reactions, such as

in Noyori-type transformations, the nature of the transition state is dictated by the LUMO

and HOMO in the complex, which often involves the metal center and one non-innocent

ligand. These transformations operate in the exact same way as FLP catalysts do,

suggesting that it is possible to extend such transition-metal chemistry to metal-free

catalysis with the right design [84]. From that standpoint, FLP chemistry can be seen as a

“metal-free organometallic’ chemistry where the 2-electron transfer occurs over more

than one main group atoms rather than being centered on a transition-metal.

8. C-H borylation of heteroarenes: a proof of concept

With these four main criteria for the design of efficient metal-free catalysts in mind, we

looked the possibility of using FLPs for an important catalytic transformation: the

functionalization of C-H bonds [85]. While many noble metals have proven to be quite

effective to cleave inactivated Csp2-H and Csp3-H bonds, one of the most recognized

catalytic system for the functionalization of C-H bonds is the direct arylation pioneered

by Fagnou and coworkers [86-89]. The key feature of this catalytic transformation is the

C-H activation of an arene by a palladium pivalate complex [86]. The activation takes

place by the simultaneous interaction of the Pd atom with the carbon of the substrate and

abstraction of the H by the carboxylate ligand (Figure 7A). Such a process is similar to

FLP-type bond cleavage. Using that rationale, the Pd acts as a Lewis acid and the O as

the Lewis base of a FLP.

Figure 7. Transition state of the A) activation of a C-H bond of an heteroarene by a

pivalate palladium(II) species demonstrating the ambiphilic nature of the activation and

B) the protodeborylation reaction in amino-borane catalysts.

Intermolecular frustrated Lewis pairs had previously been shown by Stephan and

coworkers to be capable of cleaving aliphatic C-H bonds on the α-position of aliphatic

ethers [90]. Unfortunately, this reaction required the use of strongly acidic carbocations

which did not lead to any catalytic transformations. The same group also showed that

some alkenes could undergo C-H bond cleavage using the combination of aluminum

compounds and Lewis bases [91]. Terminal alkynes can also be activated by suitable

FLPs in a process that is somewhat reminiscent of the copper-assisted deprotonation of

alkynes that is at the core of the Sonogashira cross-coupling [25, 92-93]. However, none

of these C-H bond cleavage processes could be used for catalytic transformations. In our

study of the many processes involving ambiphilic molecules, we found that the closest

analogue of this type of transition state was reported by Repo and coworkers for

describing the protodeborylation process that is at the core of his mechanism for the

hydrogenation of alkynes (Figure 7B) [78]. The protodeborylation reaction is, formally,

the reverse reaction of C-H activation and is driven in this direction by thermodynamic

factors. Since such a transition state is accessible for the elimination of an arene

fragment, we hypothesized that with the right electronic and steric considerations, the

cleavage of such a C-H bond should be possible with a proper FLP.

After investigating several possible frameworks for the activation of Csp2-H bonds,

species 1-TMP-2-BH2-C6H4 (12; TMP = 2,2,6,6-tetramethylpiperidine) was found to be a

good candidate for the activation of C-H bonds. The BH2 group is significantly Lewis

acidic and should thus be suitable to activate arenes. Also, the bulky TMP base was

chosen in order to provide the constraints that are necessary for the FLP character to be

present by limiting dimerization and intramolecular self-quenching of the active site.

While we were carrying out this work, this species was also found to cleave molecular

hydrogen [94].

Scheme 15. C-H activation of N-Me-pyrrole using aminoborane 12 to generate species

13 and reactivity of 13 with HBpin generating species 12 (TMP = 2,2,6,6-

tetramethylpiperidine and pin = pinacol).

When 12 was added to electron-rich heteroarenes, such as N-Me-pyrrole, the cleavage of

the C-H bond occurred, releasing H2 in the process and generating the respective

organoborane 13 (Scheme 15) [95-96]. This step can be seen as the reverse of the

protodeborylation reaction after hydrogen activation from an amino-borane. It is

noteworthy that C-H bond cleavage is a difficult reaction, typically requiring reactive

transition-metal catalysts [97-98]. This transformation is shown to be accessible at 80 °C,

with a TS of 24.4 kcal.mol-1 and was found to be complete within 5 hours at that

temperature. While transition metals may employ various pathways to functionalize

activated molecules, organoboranes have more limited possibilities. On the other hand,

trivalent organoboranes HBR2 can undergo redistribution through σ-bond metathesis. By

this process, in the presence of HBpin, species 13 reacts to regenerate 12 with the release

of the organoborane N-Me-pyrolle-Bpin (Scheme 15). The C-H activation – H2

elimination – σ-bond metathesis are the three key steps of a catalytic cycle for the

dehydrogenative borylation of electron-rich heteroarenes. Following this mechanism,

pyrroles, furans and thiophenes can be borylated in excellent yields with catalyst loading

as low as 1%. The mechanistic pathway is described in Figure 8 [95]. We recently

demonstrated that bench-stable aminofluoroborate salts could also operate in such

catalytic transformation [99].

Figure 8. Mechanism for the borylation of heteroarenes using ambiphilic catalyst 12.

According to the DFT studies, the C-H activation step is the limiting step of this

transformation and is quite similar to that calculated to occur in direct arylation

processes. The kinetic isotope effect of 1.8 is very similar to what was observed by

Fagnou (2.1), suggesting a similar mechanism [100]. However, a key difference is the

carbophilic nature of the borane compared to the palladium center, which makes the B-C

interaction the key component of the activation process. By contrast, the deprotonation of

the reactive site is equally or more important in transition metal chemistry, with iridium-

boryl catalysts that favor activation at the most acidic site [101-102] and palladium-

carboxylate complexes for which the selectivity is determined by both the nucleophilicity

and the acidity of the position [89]. One of the limitations in the FLP mediated process is

the dimeric nature of the catalyst that prevents the reaction to occur at ambient

temperature. Indeed, it has been demonstrated that 12 is actually a dimeric species with a

3-center-2-electron interaction of the hydroboranes. The energy to breakdown the dimer

in the monomer is calculated to be around 7 kcal.mol-1, which is easily accessible in

solution, but which hinders the activation of less electron rich species such as arenes and

thiophenes. Nonetheless, it has recently been demonstrated that under neat conditions,

thiophenes could be activated by substrate 1 [103].

9- Conclusion

It has been a decade since the initial report of a metal-free activation of hydrogen by

Stephan and coworkers. Since then, many important advances have been made to

demonstrate that frustrated Lewis pairs could activate a long list of common unreactive

substrates. The combination of reactive and readily available HOMO and LUMO orbitals

that do not mutually quench allows for such reactivity to occur and is very similar to

many transition-metal mediated processes involving 2-electron transfer. Therefore, the

use of FLPs as catalysts should not been regarded differently than transition metal

catalysts, especially for the heterolytic activation of substrates and such chemistry can be

described as a “metal-free organometallic” chemistry, highlighting the importance of the

organometallic concepts underlined in FLP chemistry.

We demonstrated in the past few years that FLP molecules could be important

challengers to transition metal catalysts for important catalytic transformations. Species 7

has been shown to be an exceptionally active catalyst for the hydroboration of CO2 to

methoxyboranes, surpassing in turn-over frequency most transition metals catalysts

reported for such reactions whereas species 12 has been shown to be a potent catalyst for

the borylation of heteroarenes, challenging the best iridium based catalysts reported to

date. Indeed, with the right design in mind, a knowledge of the nature of the bonds to be

activated and by a careful control of the Lewis basicity and acidity of the FLP

components, one can design efficient metal-free catalysts. Although metal-free catalysis

using FLPs is still in its infancy, we can expect many developments in the future that will

extend significantly its reach and potential.

Acknowledgements

This work was supported by the National Sciences and Engineering Research Council

(NSERC) of Canada and the Centre de Catalyse et Chimie Verte (Quebec). M.-A. C., M.-

A. L and E. R. acknowledge NSERC and FRQNT for scholarships.

10. References

[1] A. M. Tondreau, C. C. H. Atienza, J. J. Weller, S. A. Nye, K. M. Lewis, J. G. P.

Delis, P. J. Chirik, Science 335 (2012) 567-570.

[2] T. J. Mazzacano, N. P. Mankad, J. Am. Chem. Soc. 135 (2013) 17258-17261.

[3] W. Zuo, A. J. Lough, Y. F. Li, R. H. Morris, Science 342 (2013) 1080-1083.

[4] T.-P. Lin, J. C. Peters, J. Am. Chem. Soc. 136 (2014) 13762-13683.

[5] T. Zell, D. Milstein, Acc. Chem. Res. 48 (2015) 1979-1994.

[6] G.C. Welch, R.R.S. Juan, J.D. Masuda, D.W. Stephan, Science 314 (2006) 1124–

1126.

[7] D.J. Parks, W.E. Piers, J. Am. Chem. Soc. 118 (1996) 9440–9441.

[8] P. A. Chase, G. C. Welch, T. Jurca, D. W. Stephan, Angew. Chem. Int. Ed. 46 (2007)

8050-8053.

[9] L. J. Hounjet, D. W. Stephan, Org. Process. Res. Dev. 18 (2014) 385-391.

[10] D. W. Stephan, G. Erker, Angew. Chem. Int. Ed. 54 (2015) 6400-6441.

[11] D. W. Stephan, J. Am. Chem. Soc. 137 (2015) 10018-10032.

[12] D. W. Stephan, G. Erker, Chem. Sci. 5 (2014) 2625-2641.

[13] F.-G. Fontaine, D. Zargarian, J. Am. Chem. Soc. 126 (2004) 8786-8794.

[14] M.-H. Thibault, J. Boudreau, S. Mathiotte, F. Drouin, O. Sigouin, A. Michaud, F.-G.

Fontaine, Organometallics 26 (2007) 3807-3815.

[15] J. Boudreau, F.-G. Fontaine, Organometallics 30 (2011) 511-519.

[16] J. Boudreau, M.-A. Courtemanche, V. M. Marx, D. J. Burnell, F.-G. Fontaine,

Chem. Commun. 48 (2012) 11250-11251.

[17] J. Boudreau, M.-A. Courtemanche, F.-G. Fontaine, Chem. Commun. 47 (2011)

11131-11133.

[18] F.-G. Fontaine, M.-A. Courtemanche, M.-A. Légaré, Chem. Eur. J. 20 (2014) 2990-

2996.

[19] F. Bertini, F. Hoffmann, C. Appelt, W. Uhl, A.W. Ehlers, J. C. Slootweg, K.

Lammertsma, Organometallics 32 (2013) 6764–6769.

[20] M. J. Sgro, D. W. Stephan, Chem. Commun. 49 (2013) 2610-1612.

[21] C.M. Mömming, E. Otten, G. Kehr, R. Fröhlich, S. Grimme, D.W. Stephan, G.

Erker, Angew. Chem. Int. Ed. 48 (2009) 6643–6646.

[22] G. Ménard, D.W. Stephan, J. Am. Chem. Soc. 132 (2010) 1796–1797.

[23] M.A. Dureen, D.W. Stephan, J. Am. Chem. Soc. 132 (2010) 13559–13568.

[24] X. Zhao, D.W. Stephan, Chem. Commun. 47 (2011) 1833–1835.

[25] C. Appelt, H. Westenberg, F. Bertini, A.W. Ehlers, J. C. Slootweg, K. Lammertsma,

W. Uhl, Angew. Chem. Int. Ed. 50 (2011) 3925–3928.

[26] E. Theuergarten, J. Schlösser, D. Schlüns, M. Freytag, C.G. Daniliuc, P.G. Jones, M.

Tamm, Dalton Trans. 41 (2012) 9101-9110.

[27] M. Harhausen, R. Fröhlich, G. Kehr, G. Erker, Organometallics 31 (2012) 2801–

2809.

[28] L. Houjnet, C. B. Caputo, D. W. Stephan, Angew. Chem. Int. Ed. 51 (2012) 4714-

4717.

[29] Z. Lu, Y. Wang, J. Liu, Y. Lin, Z.H. Li, H. Wang, Organometallics 32 (2013) 6753-

6758.

[30] C. Jiang, D.W. Stephan, Dalton Trans. 42 (2013) 630–637.

[31] M. Reißmann, A. Schäfer, S. Jung, T. Müller, Organometallics 32 (2013) 6736–

6744.

[32] B.M. Barry, D. A. Dickie, L.J. Murphy, J. A. C. Clyburne, R. A. Kemp, Inorg.

Chem. 52 (2013) 8312–8314.

[33] M. Sajid, G. Kehr, T. Wiegand, H. Eckert, C. Schwickert, R. Pöttgen, A. J. P.

Cardenas, T. H. Warren, R. Fröhlich, C. G. Daniliuc, G. Erker, J. Am. Chem. Soc. 135

(2013) 8882–8895.

[34] S. A. Weicker, D. W. Stephan, Chem. Eur. J. 21 (2015) 13027-13034.

[35] A. E. Ashley, A. L. Thompson, D. O’Hare, Angew. Chem. Int. Ed., 48 (2009) 9839-

9843.

[36] A. Berkefeld, W. E. Piers, M. Parvez, J. Am. Chem. Soc., 132 (2010) 10660-10661.

[37] M.-A. Courtemanche, M.-A. Légaré, L. Maron, F.-G. Fontaine, J. Am. Chem. Soc.

135 (2013) 9326–9329.

[38] M.-A. Courtemanche, J. Larouche, M.-A. Légaré, W. Bi, L. Maron, F.-G. Fontaine,

Organometallics 32 (2013) 6804–6811.

[39] S. Porcel, G. Bouhadir, N. Saffon, L. Maron, D. Bourissou, Angew. Chem. Int. Ed.

49 (2010) 6186-6189.

[40] O. Baslé, S. Porcel, S. Ladeira, G. Bouhadir, D. Bourissou, Chem. Commun., 48

(2012) 4495-4497.

[41] M.-A. Courtemanche, M.-A. Légaré, F.-G. Fontaine, L. Maron, J. Am. Chem. Soc.

136 (2014) 10708–10717.

[42] R. Declercq, G. Bouhadir, D. Bourissou, M.-A. Légaré, M.-A. Courtemanche, K. S.

Nahi, N. Bouchard, F.-G. Fontaine, L. Maron, ACS Catal. 5 (2015) 2513–2520.

[43] T. J. Schmeier, G.E. Dobereiner, R. H. Crabtree, N. Hazari, J. Am. Chem. Soc. 133

(2011) 9274–9277.

[44] M. Beley, J.-P. Collin, R. Ruppert, J.-P. Sauvage, Chem. Commun. 2 (1984) 1315-

1316.

[45] M. Beley, J.-P. Collin, R. Ruppert, J.-P. Sauvage, J. Am. Chem. Soc. 108 (1986)

7461–7467.

[46] J.P. Collin, A. Jouaiti, J.-P. Sauvage, Inorg. Chem. 27 (1988) 1986–1990.

[47] S. Chakraborty, J. Zhang, J. A. Krause, H. Guan, J. Am. Chem. Soc. 132 (2010) 132

8872-8873.

[48] F. Huang, C. Zhang, J. Jiang, Z. X. Wang, H. Guan, Inorg. Chem. 50 (2011) 3816–

3825.

[49] M. S. Jeletic, M. T. Mock, A. M. Appel, J. C. Linehan, J. Am. Chem. Soc. 135

(2013) 11533–11536.

[50] M. S. Jeletic, M. L. Helm, E. B. Hulley, M. T. Mock, A. M. Appel, J. C. Linehan,

ACS Catal. 4 (2014) 3755–3762.

[51] I. Knopf, C. C. Cummins, Organometallics 34 (2015) 1601–1603.

[52] S. N. Riduan, Y. Zhang, J. Y. Ying, Angew. Chem. Int. Ed. 48 (2009) 3322–3325.

[53] F. Huang, G. Lu, L. Zhao, H. Li, Z. X. Wang, J. Am. Chem. Soc. 132 (2010) 12388–

12396.

[54] S. N. Riduan, J. Y. Ying, Y. Zhang, ChemCatChem 5 (2013) 1490–1496.

[55] Q. Zhou, Y. Li, J. Am. Chem. Soc. 137 (2015) 10182-10189.

[56] M.-A. Courtemanche, M.-A. Légaré, E. Rochette, F.-G. Fontaine, Chem. Commun.

51 (2015) 6858–6861.

[57] S. Kobayashi, K. Nishio, Tetrahedron Lett. 34 (1993) 3453-3456.

[58] S. Kobayashi, K. Nishio, J. Org. Chem. 59 (1994) 6620-6628.

[59] M.-A. Légaré, M.-A. Courtemanche, F.-G. Fontaine, Chem. Commun. 50 (2014)

11362–11365.

[60] T. Wang, D.W. Stephan, Chem. Commun. 50 (2014) 7007–7010.

[61] C. Das Neves Gomes, E. Blondiaux, P. Thuéry, T. Cantat, Chem. Eur. J. 20 (2014)

7098–7106.

[62] Y. Yang, M. Xu, D. Song, Chem. Commun. 51 (2015) 11293-11296.

[63] S. Y.-F. Ho, C.-W. So, N Saffon-Merceron, N. Mézailles, Chem. Commun. 51

(2015) 2107-2110.

[64] K. Fujiwara, S. Yasuda, T. Mizuta, Organometallics 33 (2014) 6692-6695.

[65] G. A. Olah, A. Goeppert, G. K. S. Prakash, J. Org. Chem. 74 (2009) 487-498.

[66] T. Mahdi, D. W. Stephan, J. Am. Chem. Soc. 136 (2014) 15809-15812.

[67] D. J. Scott, M. J. Fuchter, A. E. Ashley, J. Am. Chem. Soc. 136 (2014) 15813-

15816.

[68] P.M. Zimmerman, Z. Zhang, C.B. Musgrave, Inorg. Chem. 49 (2010) 8724–8728.

[69] H. Li, A. J. A. Aquino, D. B. Cordes, F. Hung-Low, W. L. Hase, C. Krempner, J.

Am. Chem. Soc. 135 (2013) 16066-16069.

[70] S. Mummadi, D. K. Unruh, J. Zhao, S. Li, C. Krempner, J. Am. Chem. Soc. 138

(2016) 3286-3289.

[71] M.-A. Courtemanche, Transition-Metal Free Reduction of Carbon Dioxide, Ph.D.

Thesis, Université Laval, 2015.

[72] K. Albrecht, V. Kaiser, R. Boese, J. Adams, D.E. Kaufmann, J. Chem. Soc. Perkin

Trans. 2 (2000) 2153–2157.

[73] R. Roesler, W. E. Piers, M. Parvez, J. Organomet. Chem. 680 (2003) 218–222.

[74] V. Sumerin, F. Schulz, M. Atsumi, C. Wang, M. Nieger, M. Leskelä, T. Repo, P.

Pyykkö, B. Rieger, J. Am. Chem. Soc. 130 (2008) 14117–14119.

[75] S. Schwendemann, R. Fröhlich, G. Kehr, G. Erker, Chem. Sci. 2 (2011) 1842–1849.

[76] V. Sumerin, K. Chernichenko, M. Nieger, M. Leskelä, B. Rieger, T. Repo, Adv.

Synth. Catal. 353 (2011) 2093–2110.

[77] K. Chernichenko, M. Nieger, M. Leskelä, T. Repo, Dalton Trans. 41 (2012) 9029–

9032.

[78] K. Chernichenko, A. Madarász, I. Pápai, M. Nieger, M. Leskelä, T. Repo, Nat.

Chem. 5 (2013) 718–723.

[79] É. Rochette, M.-A. Courtemanche, A. Pulis, W. Bi, F.-G. Fontaine, Molecules 20

(2015) 11902–11914.

[80] M.-A. Courtemanche, É. Rochette, M.-A. Légaré, W. Bi, F.-G. Fontaine, Dalton

Trans. 45 (2016), 6129-6135.

[81] M.-A. Courtemanche, A. P. Pulis, É. Rochette, M.-A. Légaré, D. W. Stephan, F.-G.

Fontaine, Chem. Commun. 51 (2015) 9797–9800.

[82] R. Noyori, Chem. Rev. 95 (1995) 259-272.

[83] R. Noyori, Angew. Chem. Int. Ed. 41 (2002) 2008-2022.

[84] P. P. Power, Nature 463 (2010) 171-177.

[85] J.-Q. Yu, Z. Shi, Eds. (2010) Topics in Current Chemistry, Vol. 292 : C-H

Activation. Berlin, Springer-Verlag.

[86] M. Lafrance, K. Fagnou, J. Am. Chem. Soc. 128 (2006) 16496-16497.

[87] L.-C. Campeau, K. Fagnou, Chem. Commun. (2006) 1253-1264.

[88] D. Alberico, M. E. Scott, M. Lautens, Chem. Rev. 107 (2007) 174-238.

[89] S. I. Gorelsky, Coord. Chem. Rev. 257 (2013) 153-164.

[90] M. H. Holthausen, T. Mahdi, C. Schlepphorst, L. J. Hounjet, J. J. Weigand, D.W .

Stephan, Chem. Commun. 50 (2014) 10038-10040.

[91] G. Ménard, D.W. Stephan, Angew. Chem. Int. Ed. 51 (2013) 4409-4412.

[92] M. A. Dureen, D. W. Stephan, J. Am. Chem. Soc. 131 (2009) 8396-8397.

[93] C. Jiang, O. Blacque, H. Berke, Organometallics 29 (2010) 125-133.

[94] K. Chernichenko, B. Kótai, I. Pápai, V. Zhivonitko, M. Nieger, M. Leskelä, T. Repo,

Angew. Chem. Int. Ed., 54 (2015) 1749-1753.

[95] M.-A. Légaré, M.-A. Courtemanche, É. Rochette, F.-G. Fontaine, Science 349

(2015) 513–516.

[96] S. K. Bose, T. B. Marder, Science 349 (2015) 473-474.

[97] I. A. Mkhalide, J. H. Barnard, T. B. Marder, J. M. Murphy, J. F. Hartwig, Chem.

Rev. 110 (2010) 890-931.

[98] J.-Y. Cho, M. K. Tse, D. Holmes, R. E. Maleczka Jr., M. R. Smith 3rd, Science 295

(2002) 305-308.

[99] M.-A. Légaré, É. Rochette, J. Légaré-Lavergne, N. Bouchard, Chem. Commun. 52

(2016) 5387-5390.

[100] S. I. Gorelsky, D. Lapointe, K. Fagnou, J. Am. Chem. Soc. 130 (2008) 10848-

10849.

[101] B. A. Vanchura II, S. M. Preshlock, P. C. Roosen, V. A. Kallepalli, R. J. Staples, R.

E. Maleczka Jr, D. A. Singleton, M. R. Smith III, Chem. Commun. 46 (2010) 7724-7726.

[102] H. Tajuddin, P. Harrisson, B. Bitterlich, J. C. Collings, N. Sim, A. S. Batsanov, M.

S. Cheung, S. Kawamorita, A. C. Maxwell, L. Shukla, J. Morris, Z. Lin, T. B. Marder, P.

G. Steel, Chem. Sci. 3 (2012) 3505-3515.

[103] K. Chernichenko, M. Lindqvist, B. Kotai, M. Nieger, K. Sorochkina, I. Papai, T.

Repo, J. Am. Chem. Soc. 138 (2016) 4860-4868.