Transition-Metal-Free Reduction of Carbon Dioxide

Transition-Metal-Free Reduction of Carbon
Dioxide
Thèse
Marc-André Courtemanche
Doctorat en chimie
Philosophiae Doctor (Ph.D.)
Québec, Canada
© Marc-André Courtemanche 2015
Résumé
Seulement neuf années se sont écoulées depuis la découverte que les ‘’Paires de Lewis
Frustrées’’ (PLF) peuvent promouvoir le clivage de l’hydrogène, mais plus d’un millier
d’articles scientifiques ont déjà été publiés sur le sujet. Au début des travaux décrits dans
cette thèse, les catalyseurs pour la réduction du CO2 étaient excessivement rares et peu
efficaces. La présente thèse porte donc sur le développement de nouveaux systèmes sans
métal de transition pour la réduction catalytique du CO2 en molécules riches en énergie et
plus précisément, en méthanol.
D’abord, la préparation d’un nouveau système basé sur les PLF et sa capacité à activer le
CO2 de façon réversible est présenté. En présence de catécholborane, le CO2 est
catalytiquement réduit en méthoxyboranes, espèces facilement hydrolysables en méthanol.
Surprenamment, un produit de décomposition est identifié comme étant responsable de
l’activité catalytique.
En effet, l’espèce ambiphile 1-Bcat-2-PPh2-C6H4 constitue le premier exemple d’un
catalyseur sans métal de transition pour l’hydroboration du CO2. L’activité de ce catalyseur
excessivement simple surpasse celle des meilleurs systèmes basés sur des métaux.
Des études mécanistiques détaillées révèlent que l’activation simultanée du borane et du CO2
est d’une importance critique. Une investigation poussée révèle que la formation d’un adduit
entre le catalyseur et le formaldéhyde résulte en un organocatalyseur d’autant plus actif.
Il est aussi démontré que les phosphazènes, super bases organiques, sont des
organocatalyseurs très actifs pour la transformation du CO2 en dérivés de formate ou de
méthanol. De façon intéressante, le DMF (N,N-diméthylformamide) peut promouvoir
l’hydrosilylation réductive du CO2 en absence de catalyseur.
Une nouvelle stratégie d’hydrogénation a été développée en étudiant les aspects
fondamentaux de l’hydrogénation par les PLFs, permettant ainsi la conception d’un système
pour l’hydrogénation du CO2 en conditions ambiantes. Même si une voie de décomposition
inattendue rend le processus stoéchiométrique, une optimisation du catalyseur pourrait
générer le premier catalyseur sans métal pour l’hydrogénation du CO2.
iii Abstract
Only nine years have passed since the seminal discovery that Frustrated Lewis Pairs (FLPs)
could split dihydrogen and yet, more than a thousand research papers have already been
published on the subject. As the work presented herein commenced, metal-free systems
capable of catalytically transforming CO2 could be counted on a single hand while transitionmetal based systems were almost as scarce. As such, the present thesis deals with the
development of novel transition-metal-free systems for the catalytic reduction of CO2 to
energy rich materials, most notably methanol.
Firstly, the preparation of a new FLP system bearing three pendant phosphine groups
Al(C6H4(o-PPh2))3 and its ability to activate carbon dioxide in a reversible fashion are
presented. In the presence of catecholborane, CO2 is catalytically reduced to
methoxyboranes, species which are readily hydrolyzed to methanol. Interestingly, a
decomposition product is shown to be responsible for the catalytic activity
Indeed, species 1-Bcat-2-PPh2-C6H4 is the first report of a catalyst for the metal-free
hydroboration of carbon dioxide. The activity of this excessively simple catalyst surpasses
that of the best transition metal systems while using the cheap and high hydrogen content
borane BH3.SMe2.
In-depth mechanistic studies reveals that simultaneous activation of both the borane and CO2
molecules is of critical importance. Further investigation reveals that the formation of an
adduct between the catalyst and formaldehyde affords an even more potent organocatalyst.
It is also shown that phosphazene superbases are very active organocatalysts for the
transformation of CO2 to either formate or methanol derivatives. Unexpectedly, N,Ndimethylformamide (DMF) can promote the reductive hydrosilylation of CO2 in the absence
of any catalyst.
Finally, the challenging task of developing a metal-free system for the hydrogenation of CO2
was undertaken. A novel strategy was developed by studying the fundamental aspects of FLP
mediated hydrogenations, allowing us to achieve CO2 reduction under ambient conditions.
While an unexpected decomposition pathway hampered catalysis, optimisation of the
catalyst design is expected to yield the first metal-free catalyst for the hydrogenation of CO2.
v Table of Contents
Résumé ............................................................................................................................................... iii Abstract ............................................................................................................................................... v Table of Contents .............................................................................................................................. vii List of Tables ................................................................................................................................... xiii List of Schemes ............................................................................................................................... xvii Acknowledgements ....................................................................................................................... xxvii Inserted research articles and author contributions: ....................................................................... xxix 1 2 Introduction ................................................................................................................................. 1 1.1 A note on chronology:......................................................................................................... 1 1.2 A note on compound identification: .................................................................................... 1 1.3 General introduction............................................................................................................ 2 1.4 Carbon dioxide and the greenhouse effect .......................................................................... 3 1.5 Using CO2 as an energy vector: the methanol economy ..................................................... 5 1.6 Challenges in the transformation of carbon dioxide ........................................................... 7 1.7 The carbon dioxide molecule .............................................................................................. 8 1.8 Organometallic binding of carbon dioxide........................................................................ 10 1.9 Carbon dioxide for energy storage: general aspects.......................................................... 13 1.10 Metal enzymes for CO2 reduction ..................................................................................... 14 1.11 Transition metal catalyzed reduction of CO2 to energy rich materials ............................. 15 1.12 Photochemical reduction of carbon dioxide ...................................................................... 17 1.13 Transition-metal free transformations of carbon dioxide .................................................. 18 1.14 A short history of Frustrated Lewis Pair chemistry .......................................................... 20 1.15 Carbon dioxide activation and reduction using Frustrated Lewis Pairs ............................ 23 1.16 Original hypotheses and motivation.................................................................................. 24 1.17 Non-frustrated aluminum based ambiphilic molecules for CO2 capture........................... 27 1.18 Scope of thesis .................................................................................................................. 30 Experimental methods............................................................................................................... 31 2.1 Inert atmosphere chemistry ............................................................................................... 31 2.1.1 Glovebox ................................................................................................................... 31 2.1.2 Schlenk line ............................................................................................................... 32 2.2 Working with gases. .......................................................................................................... 33 2.3 Nuclear Magnetic Resonance (NMR) spectroscopy. ........................................................ 34 2.4 X-ray crystallography........................................................................................................ 35 vii 2.5 DFT calculations ............................................................................................................... 36 2.6 Gas chromatography.......................................................................................................... 38 3 Aluminum based ambiphilic molecules fort the capture of carbon dioxide. ............................. 39 3.1 Advances in the transition-metal mediated reduction of CO2............................................ 39 3.2 Advances in the binding of CO2 by FLPs.......................................................................... 41 3.3 Fundamental aspects of CO2 chemistry: the general concept. ........................................... 42 3.4 Overview of the project ..................................................................................................... 44 3.5 Research article: A Tris(triphenylphosphine)aluminum Ambiphilic Precatalyst for the
Reduction of Carbon Dioxide with Catecholborane.......................................................... 47 3.5.1 Résumé ...................................................................................................................... 47 3.5.2 Abstract ..................................................................................................................... 47 3.5.3 Introduction ............................................................................................................... 48 3.5.4 Results and discussion ............................................................................................... 50 3.5.5 Conclusion ................................................................................................................. 61 3.6 Experimental Section ........................................................................................................ 62 3.6.1 General procedure ..................................................................................................... 62 3.6.2 Synthesis of compounds ............................................................................................ 63 3.6.3 CO2 reduction catalytic tests ..................................................................................... 64 3.6.4 Crystallographic studies ............................................................................................ 65 3.6.5 Computational details ................................................................................................ 67 Metal-Free catalytic Reduction of CO2 to methanol ................................................................. 69 4 4.1 4.1.1 Résumé ...................................................................................................................... 69 4.1.2 Abstract ..................................................................................................................... 69 4.1.3 Introduction ............................................................................................................... 70 4.1.4 Results and discussion ............................................................................................... 72 4.1.5 Conclusions ............................................................................................................... 81 4.2 viii Research article: A Highly Active Phosphine-Borane Organocatalyst for the Reduction of
CO2 to Methanol using Hydroboranes............................................................................... 69 Experimental ..................................................................................................................... 82 4.2.1 General experimental ................................................................................................ 82 4.2.2 Synthesis of 1 ............................................................................................................ 83 4.2.3 General procedure for catalytic reduction of carbon dioxide .................................... 84 4.2.4 General procedure for big scale catalytic reduction of carbon dioxide ..................... 85 4.2.5 Reactivity with methylformate .................................................................................. 85 4.2.6 In-situ preparation of IM2 ......................................................................................... 86 4.2.7 Crystallographic information .................................................................................... 86 4.2.8 Computational Details:.............................................................................................. 88 4.3 5 Mechanistic investigations ........................................................................................................ 91 5.1 Metal mediated catalytic reductions of CO2...................................................................... 91 5.2 Metal-free activation of CO2 ............................................................................................. 93 5.3 Metal-free reduction of CO2 .............................................................................................. 94 5.4 Research article: Reducing CO2 to Methanol using Frustrated Lewis Pairs: On the
Mechanism of Phosphine-Borane Mediated Hydroboration of CO2 ................................. 95 5.4.1 Résumé ...................................................................................................................... 95 5.4.2 Abstract ..................................................................................................................... 95 5.4.3 Introduction ............................................................................................................... 96 5.4.4 Computational details................................................................................................ 98 5.4.5 General remarks ........................................................................................................ 99 5.4.6 First reduction step: CO2 to HCOOBcat ................................................................. 100 5.4.7 Second reduction step: from HCOOBcat to CH2O and derivatives ........................ 106 5.4.8 Third reduction step: reducing CH2O and derivatives to CH3OBcat ...................... 111 5.4.9 Discussion ............................................................................................................... 115 5.4.10 Conclusion .............................................................................................................. 117 5.5 Experimental Data:.......................................................................................................... 118 5.5.1 General experimental .............................................................................................. 118 5.5.2 Synthetic methodology............................................................................................ 118 5.5.3 Additional computational information regarding the catalyzed reduction of CO2 to
HCOOBcat .............................................................................................................. 119 5.5.4 Additional computational information regarding the catalyzed reduction of
HCOOBcat to CH2O and derivatives ...................................................................... 120 5.5.5 Additional computational information regarding the catalyzed reduction of CH2O
and derivatives to CH3OBcat .................................................................................. 121 5.5.6 Additional computational information regarding the possible involvement of IM3C
in catalysis ............................................................................................................... 122 5.5.7 Additional computational information regarding the uncatalyzed reduction of CO2 to
CH3OBcat ................................................................................................................ 123 5.6 6 Conclusions and perspectives ........................................................................................... 89 Conclusion and perspectives ........................................................................................... 126 Metal-free catalytic hydrosilylation of CO2 to methanol ........................................................ 131 6.1 Base catalyzed hydroboration of CO2 ............................................................................. 131 6.2 Base catalyzed hydrosilylation of CO2............................................................................ 132 ix 6.3 Research article: Phosphazenes: Efficient Organocatalysts for the Catalytic
Hydrosilylation of Carbon Dioxide ................................................................................. 136 6.3.1 Résumé .................................................................................................................... 136 6.3.2 Abstract ................................................................................................................... 136 6.3.3 Introduction ............................................................................................................. 137 6.3.4 Results and discussion ............................................................................................. 139 6.3.5 Conclusions ............................................................................................................. 145 6.4 Experimental ................................................................................................................... 146 6.4.1 General experimental: ............................................................................................. 146 6.4.2 Initial test experiment: ............................................................................................. 147 6.4.3 Rearrangement of 1 to 4 .......................................................................................... 147 6.4.4 Rearrangement of 2 to 5 .......................................................................................... 148 6.4.5 Rearrangement of 3 to 6 .......................................................................................... 149 6.4.6 Catalytic reduction using 13CO2............................................................................... 150 6.4.7 Following the reaction in DMF, with various loadings of silane: ........................... 150 6.4.8 Reaction under 5 atm. of CO2...................................................................................... 151 6.4.9 6.5 7 Conclusions and perspectives .......................................................................................... 156 Metal-free hydrogenation of carbon dioxide ........................................................................... 159 7.1 Introduction to FLP mediated hydrogenations ................................................................ 159 7.2 Understanding FLP hydrogenation.................................................................................. 160 7.3 Research article: Intramolecular B/N Frustrated Lewis Pairs and the Hydrogenation of
Carbon Dioxide ............................................................................................................... 169 7.3.1 Résumé .................................................................................................................... 169 7.3.2 Abstract ................................................................................................................... 169 7.3.3 Introduction ............................................................................................................. 170 7.3.4 Results and discussion ............................................................................................. 171 7.3.5 Conclusions ............................................................................................................. 178 7.4 x General procedures for catalytic tests: .................................................................... 152 Experimental ................................................................................................................... 179 7.4.1 General experimental: ............................................................................................. 179 7.4.2 Synthesis of compounds 1, 2 and precursors........................................................... 180 7.4.3 HD scrambling with 1 ............................................................................................. 183 7.4.4 Synthesis of 3 from 1............................................................................................... 183 7.4.5 Hydrogenation of CO2 - J Young NMR Tube Experiments .................................... 184 7.4.6 Synthesis of 4 from 2 (Bigger scale hydrogenation) ............................................... 186 7.4.7 Computational details: ............................................................................................ 188 7.4.8 DFT calculation for the protodeborylation of 1 and 2............................................. 189 7.4.9 DFT study of the possible isomers of 3................................................................... 193 7.4.10 DFT study of the hydrogenation of CO2 ................................................................. 194 7.5 8 Conclusion and perspectives ................................................................................................... 199 8.1 General conclusions ........................................................................................................ 199 8.2 Ongoing and future work ................................................................................................ 202 8.2.1 Non-CO2 related chemistry ..................................................................................... 202 8.2.2 CO2 related chemistry ............................................................................................. 203 8.3 Philosophical discussion ................................................................................................. 205 8.3.1 Carbon dioxide ........................................................................................................ 205 8.3.2 Frustrated Lewis Pairs ............................................................................................. 208 8.4 9 Conclusions and perspectives ......................................................................................... 198 Personal bibliogaphy ....................................................................................................... 210 Bibliography............................................................................................................................ 213 xi List of Tables
Table 1-1: Summary of key isolated transition metal CO2 complexes. ............................................ 12 Table 2-1: NMR nuclei of relevance to this project. ......................................................................... 34 Table 4-1: Reduction of CO2 with various hydroboranes. ................................................................ 77 Table 4-2: Crystal data and structural refinements for compound 1. ................................................ 87 Table 5-1: Crystal data and structural cefinements for compound 2. ............................................. 125 Table 6-1: Catalytic hydrosilylation of CO2 by phosphazene bases. .............................................. 144 Table 7-1: Hydrogenation of carbon dioxide by 1 and 2 ................................................................ 175 Table 7-2: Hydrogenation of CO2 with 1 and 2 .............................................................................. 185 xiii List of Figures
Figure 1-1: a) CO2 concentrations in the atmosphere b) Global temperature anomalies c) Trend of
global temperature variation. .............................................................................................................. 3 Figure 1-2: Conceptual representation of the use of CO2 to produce energy vectors. ........................ 6 Figure 1-3: Representation of the ambiphilic character of the CO2 molecule. ................................... 8 Figure 1-4: Walsh diagram of the CO2 molecule ................................................................................ 9 Figure 1-5: Transition state in the process of CO2 reduction by CO dehydrogenase.*..................... 15 Figure 1-6: Proposed mechanistic pathways for the base-catalysed hydrosilylation of CO2. ........... 19 Figure 1-7: Proposed mechanism for the FLP-mediated hydrosilylation of CO2. ............................ 23 Figure 1-8 Reported CO2-FLP adducts- October 2010. .................................................................... 25 Figure 1-9: Reported CO2-FLP adducts- June 2011. ........................................................................ 26 Figure 2-1: A glovebox workstation. ................................................................................................ 31 Figure 2-2: a) A Schlenk line b) Schlenk flasks................................................................................ 32 Figure 2-3: The cluster that was used for calculations: Colosse. ...................................................... 38 Figure 3-1: Catalysts for the reduction of CO2 a) Ni pincer catalyst b) Ir Pincer catalyst c) Pd or Pt
pincer catalyst. .................................................................................................................................. 40 Figure 3-2: New reported FLP systems for CO2 binding. ................................................................. 42 Figure 3-3: Qualitative thermodynamic analysis of Lewis pair reactivity a) Classical Lewis pair
reactivity b) Frustrated Lewis Pair reactivity. ................................................................................... 43 Figure 3-4: Structures of the previously reported ambiphilic ligands. ............................................. 48 Figure 3-5: ORTEP drawing of the first molecule of 1 in the asymmetric unit cell. ........................ 51 Figure 3-6: Optimized structure of 2 using DFT calculations. ......................................................... 53 Figure 3-7: Optimized structure of 2 using DFT calculations. ......................................................... 55 Figure 3-8: Turnover number (TON) for the formation of CH3OBcat. ............................................ 57 Figure 3-9:ORTEP drawing of one independent molecule of 5........................................................ 60 Figure 4-1: ORTEP drawing of 1 in the asymmetric unit cell. ......................................................... 72 Figure 4-2: Turn-over numbers (TON) for the formation of CH3OBcat. ......................................... 74 Figure 4-3: Turn-over numbers (TON) for the formation of (CH3OBO)3. ....................................... 75 Figure 4-4: Enthalpy profile (in kcal mol-1) for the reduction of CO2 by 1 and catecholborane. ...... 79 Figure 4-5: Generation of a formaldehyde adduct from 1. ............................................................... 80 Figure 4-6: Assignment of spectra for 1. .......................................................................................... 83 Figure 5-1: New metal based catalytic systems for the reduction of CO2......................................... 92 Figure 5-2: New FLP systems for the stoichiometric binding of CO2. ............................................. 93 xv Figure 5-3: Important intermediates and transition states for the catalyzed reduction of CO2 to
HCOOBcat. ..................................................................................................................................... 105 Figure 5-4: Relative energies of transition states and intermediates for the reduction of HCOOBcat
to CH2O or catBOCH2OBcat........................................................................................................... 110 Figure 5-5: ORTEP drawing of 2. ................................................................................................... 111 Figure 5-6: Relative energies of transition states and intermediates for the reduction of CH2O to
CH3OBcat. ....................................................................................................................................... 114 Figure 5-7: New transition states for CO2 reduction involving a formaldehyde adduct as catalyst .
......................................................................................................................................................... 128 Figure 6-1: New catalysts for the hydroboration of CO2................................................................. 131 Figure 6-2: Assignment of spectra for 4. ......................................................................................... 147 Figure 6-3 Product ratio over time for Test A ................................................................................. 150 Figure 6-4 Product ratio over time for Test B ................................................................................. 151 Figure 7-1: Qualitative analysis of hydrogen splitting by FLP systems.......................................... 160 Figure 7-2: Schematic representation of an hydrogenation reaction mediated by the Noyori catalyst.
......................................................................................................................................................... 164 Figure 7-3: General strategy for the hydrogenation of CO2 by FLP systems. ................................. 165 Figure 7-4: A CO2 molecule with the partial charges. .................................................................... 165 Figure 7-5: Hydrogen transfer from ammonia-borane to CO2. ....................................................... 166 Figure 7-6: DFT study of the thermodynamics of H2 splitting and CO2 binding by a variety of aryl
bridged FLP systems. ...................................................................................................................... 167 Figure 7-7: ORTEP depiction of 4. ................................................................................................. 176 Figure 7-8: Geometry of TS for reaction of C with CO2. ................................................................ 177 Figure 7-9: First protodeborylation step for 1. ................................................................................ 189 Figure 7-10: Second protodeborylation step for 1. ......................................................................... 190 Figure 7-11: First protodeborylation step for 2. .............................................................................. 191 Figure 7-12: Second protodeborylation step for 2. .......................................................................... 192 Figure 7-13: DFT optimized structures of the possible isomers for compound 3. .......................... 193 Figure 7-14: Hydrogenation of CO2 by 1. ....................................................................................... 194 Figure 7-15: Hydrogenation of CO2 by 1 (continued)..................................................................... 195 Figure 7-16: Hydrogenation of CO2 by 2. ....................................................................................... 196 Figure 7-17: Hydrogenation of CO2 by 2 (continued)..................................................................... 197 Figure 8-1: Timeline of the important systems for CO2 reduction to energy rich material. ............ 201 Figure 8-2: New FLP framework for C=O bond hydrogenation. .................................................... 204
xvi List of Schemes
Scheme 1-1: Schematic representation of the amine scrubbing process. ............................................ 4 Scheme 1-2: Generation of Aresta's complex. .................................................................................. 10 Scheme 1-3: Aerobic combustion of methane. ................................................................................. 13 Scheme 1-4: Stepwise CO2 hydrogenation. ..................................................................................... 14 Scheme 1-5: Cascade catalytic hydrogenation of CO2 to methanol. ................................................. 16 Scheme 1-6: Catalytic hydroboration of CO2 by a Ni pincer complex. ............................................ 16 Scheme 1-7: Dual interaction of reduced CO2 with an Ir pincer complex. ....................................... 17 Scheme 1-8: Lewis acid catalyzed hydrosilylation of CO2. .............................................................. 18 Scheme 1-9: Formation of a classical Lewis adduct. ........................................................................ 20 Scheme 1-10: Trapping of benzyne by a phosphine-borane Lewis pair. .......................................... 21 Scheme 1-11: Trapping of hydrochloric acid using an amino-borane Lewis pair. ........................... 21 Scheme 1-12: Representation of Frustrated Lewis Pair reactivity and hydrogen splitting. .............. 22 Scheme 1-13: Reversible metal-free activation of CO2 by FLPs. ..................................................... 22 Scheme 1-14: FLP mediated CO2 hydrogenation and formation of methanol. ................................. 23 Scheme 1-15: FLP mediated stoichiometric reduction of CO2 by ammonia borane. ....................... 24 Scheme 1-16: Preparation of ambiphilic compounds 1 and 2. .......................................................... 27 Scheme 1-17: Reactivity of ambiphilic molecules 1 and 2 with CO2. .............................................. 28 Scheme 1-18: Formation of a new spirocylic CO2 activation product. ............................................. 29 Scheme 3-1: CO2 hydrogenation by Ru catalyst. .............................................................................. 39 Scheme 3-2: Hydroboration of CO2 by a Ruthenium polyhydride catalyst. ..................................... 41 Scheme 3-3: Proposed dual activation strategy for the catalytic hydroboration of CO2 ................... 45 Scheme 3-4: Decomposition of 1 into 2. ........................................................................................... 45 Scheme 3-5: Synthesis of Al(C6H4(o-PPh2))3 (1). ............................................................................. 50 Scheme 3-6: Generation of 3 upon exposure of 1 to CO2. ................................................................ 54 Scheme 3-7: Proposed protection of the phosphine moieties of 1. ................................................... 58 Scheme 3-8: Synthesis of species 5. ................................................................................................. 59 Scheme 4-1: Reduction of CO2 in presence of HBcat and catalyst 1................................................ 73 Scheme 4-2: Synthesis of 1. .............................................................................................................. 83 Scheme 4-3: Catalytic reduction of methylformate by 1 and HBcat. ............................................... 85 Scheme 5-1: Catalytic hydroboration of CO2 after ring-expansion of an NHC-9-BBN adduct. ...... 94 Scheme 5-2: Schematic representation of the stepwise hydroboration of CO2 to methoxyboranes using
hydroboranes (H[B]). ........................................................................................................................ 99 xvii Scheme 5-3: Reaction of 1 with CO2, generating IM0 illustrating the potential binding sites for HBcat.
......................................................................................................................................................... 100 Scheme 5-4: Pathway A: hydroboration reaction of CO2 through a classical 4-membered transition
state. [B] = Bcat. .............................................................................................................................. 101 Scheme 5-5: Pathway B: hydroboration through coordination of HBcat to the catechol fragment
followed by intramolecular hydride delivery. [B] = Bcat. .............................................................. 102 Scheme 5-6: Pathway C: hydroboration through simultaneous Lewis base activation of the borane
and Lewis acid activation of CO2. [B] = Bcat. ................................................................................ 103 Scheme 5-7: Pathway D: CO2 reduction through the generation of a boronium / hydridoborate ion
pair. [B] = Bcat. ............................................................................................................................... 104 Scheme 5-8: Calculated catalyst-free mechanism for the hydroboration of HCOO[B]. ................. 106 Scheme 5-9: Experimental verification for the hydroboration of formic acid by catecholborane.
[B]=Bcat. ......................................................................................................................................... 107 Scheme 5-10: Attempt to isolate HCOOBcat, leading to the exclusive formation of 2 (IM2C’). [B]=
Bcat.................................................................................................................................................. 107 Scheme 5-11: Possible interactions and rearrangements of HCOOBcat with catalyst 1. [B]=Bcat.
......................................................................................................................................................... 108 Scheme 5-12: Suggested pathway for the catalyzed reduction of HCOO[B] involving the catalyst
[B]=Bcat. ......................................................................................................................................... 109 Scheme 5-13: Formation of IM2C’(2) through the rearrangement of catBOCH2OBcat to CH2O.
[B]=Bcat. ......................................................................................................................................... 112 Scheme 5-14: Experimental verification for the catalytic role of 1 in the hydroboration of 4bromobenzaldehyde by catecholborane. [B]=Bcat.......................................................................... 113 Scheme 5-15: Catalyzed reduction of formaldehyde to CH3OBcat, regenerating the catalyst. [B]=Bcat
......................................................................................................................................................... 113 Scheme 5-16: Proposed mechanistic pathway including important transition states for the reduction
of CO2 to CH3OBcat by 1. [B]=Bcat. .............................................................................................. 116 Scheme 5-17: Calculated pathways for the reduction of CO2 to HCOOBcat. [B]=Bcat................. 119 Scheme 5-18: Calculated pathways for the reduction of HCOOBcat to CH2O and derivatives.
[B]=Bcat. ......................................................................................................................................... 120 Scheme 5-19: Calculated pathways for the uncatalyzed reduction of CO2 to CH2O. [B]=Bcat. .... 121 Scheme 5-20: Calculated pathways for the catalyzed reduction of CO2 to HCOOBcat from IM3C.
[B]=Bcat. ......................................................................................................................................... 122 xviii Scheme 5-21: Proposed mechanism for the catalyst free reduction of carbon dioxide to CH3OBcat.
[B]=Bcat. ......................................................................................................................................... 123 Scheme 5-22: Labelling experiments with 13CH2O......................................................................... 127 Scheme 6-1: CO2 capture by dimethylamine. ................................................................................. 132 Scheme 6-2: Base catalzyed reduction of CO2 to methylamines. ................................................... 133 Scheme 6-3: Protonation of a phosphazene superbase.................................................................... 134 Scheme 6-4: Proposed strategy for the hydrosilylation of CO2 using a phosphazene catalyst. ...... 135 Scheme 6-5: Rearrangement of phosphazenes in the presence of CO2 with proposed intermediate.
......................................................................................................................................................... 140 Scheme 6-6: Product ratio after the catalytic reduction under 5 atm of CO2 and after addition of an
extra loading of silane in the absence of CO2. Yield based on Si-H. .............................................. 142 Scheme 6-7: Preparation of 4 from 1. ............................................................................................. 147 Scheme 6-8: Preparation of 5 from 2. ............................................................................................. 148 Scheme 6-9: Assignment of signals for 5. ...................................................................................... 148 Scheme 6-10: Preparation of 6 from 3. ........................................................................................... 149 Scheme 6-11: Assignment of signals for 6. .................................................................................... 149 Scheme 7-1: Hydrogen activation by a BCF/Et2O Lewis Pair. ....................................................... 162 Scheme 7-2: Catalytic hydrogenation by a BCF/ET2O Lewis Pair................................................. 162 Scheme 7-3: FLP mediated reduction of CO at a rhenium center using a strong phosphazene base in
combination with a weak Lewis acid. ............................................................................................. 164 Scheme 7-4: Preparation of 1-2....................................................................................................... 171 Scheme 7-5: DFT study of possible isomers of 3. .......................................................................... 172 Scheme 7-6: DFT study of H2 activation and protodeborylation events. ........................................ 173 Scheme 7-7: Synthesis of 3 from 1. ............................................................................................... 183 Scheme 7-9: Synthesis of 4 from 2 (bigger scale hydrogenation). ................................................. 186 Scheme 7-10: Assignment of the NMR spectra for 4. .................................................................... 187 Scheme 8-1: Reversible splitting of formic acid by a novel, robust FLP system. .......................... 203 xix List of Abbreviations
{1H} = proton decoupled
13
CO2 = carbon 13 labeled carbon dioxide
6-31+G** = a basis set for DFT calculations
9-BBN = 9-borabicyclo[3.3.1]nonane
Å= angstrom
ADH = alcohol dehydrogenase
AlF = Al(C6F5)3
B3PW91 = a DFT functional
B97D = a DFT functional
BCF = B(C6F5)3
BMe3 = trimethylborane
BPh3 =triphenylborane
COSY = NMR experiment correlating 1H and 1H nucleus
d = doublet
DCM = dichloromethane
dd = doublet of doublets
DFT = Density Functional Theory
DMSO = dimethylsulfoxide
Et2O = diethyl ether
FalDH = formaldehyde dehydrogenase
FateDH = formate dehydrogenase
FID = flame ionization detector
FLP = Frustrated Lewis Pair
FTIR = Fourier transform infrared spectroscopy
GC = gas chromatography
gHSQC = NMR experiment correlating 13C and 1H nucleus
HBcat = catecholborane
xxi HBPin = pinacolborane
HMB = hexamethylbenzene
HOMO = highest occupied molecular orbital
LUMO = lowest unoccupied molecular orbital
Lutidine =2,6-dimethylpyridine
m = multiplet
Me = methyl
MeOH = methanol
Mes = 2,4,6,-trimethylphenyl
Mes’ = 2,4,5,-trimethylphenyl
Ms = mass spectroscopy
Mt = mega tonne
MTO = methanol to olefin process
NADH = nicotinamide adenine dinucleotide
NBO = natural bond order
NMR = nuclear magnetic resonance
NOx = mono-nitrogen oxides
Ph = phenyl
PNN = (2-(di-tert-butylphosphinomethyl)-6-(diethylaminomethyl)pyridine)
PPh3 = triphenylphosphine
ppm = parts per million
r.t = room temperature
s = singlet
SDD = a basis set for DFT calculations that was developed by Stuttgart and Dresden
t = triplet
THF = tetrahydrofuran
TMP = 2,2,6,6-tetramethylpiperidine
TOF = turnover frequency
xxii TON = turnover number
TS = transition state
w-B97XD = a DFT functional
δ = chemical shift
ΔH = enthalpy
ΔHform = enthalpy of formation
Π = pi orbitals
σ = sigma orbitals
xxiii To my parents, Serge and Lise
‘‘The more original a discovery, the more obvious it seems afterwards’’
- Arthur Koestler
xxv Acknowledgements
First and foremost, I would like to thank my supervisor, Frédéric-Georges Fontaine for a
wonderful time in the Fontaine group. Thank you for the continuous support, for believing
in me more than I believed in myself, for always backing up my crazy ideas and suggesting
even crazier ones. Thank you for pushing me a little bit further every single time, for sending
me to countless conferences and for the helpful discussions about life, chemistry, scotch,
chess, and of course Game of Thrones. To be sure, there is so much more to say, but let me
put it that way: if it could be possible to start a second Ph.D. in your group, I would do it
again without the slightest hesitation.
I feel like I have grown incredibly in the past five years, and that is mostly due to the amazing
people I have been working with. Thank you Ambreen, Maria, Etienne, Fred, Marc and
Nicolas for being so awesome and for the countless discussions, whether it was about politics,
religion, the meaning of life or about how robots will take over the world, every moment
spent with you guys is cherished and you will be dearly missed. Special thanks to the new
guys: Étienne, Nicolas, Julien and Matthew for being awesome colleagues and for
collaborating on many of the new projects, you guys have a great future ahead.
I don’t know how I could possibly use words to express my gratitude for meeting and
working with my partner Marc-André Légaré. It was not always easy, but through the years,
projects and experiences, we have done so much together and have become so close. I’m not
sure how it’s going to be when it comes to the point where our roads have to split, but I’d
rather not think about that and reflect on the awesome times drinking pivo, playing games,
discovering chemistry together and all the rest…. Thank you so much for everything!
Of course, none of this would have been possible alone and I would like to thank every other
person who have contributed to the projects and/or worked in the Fontaine lab. Jérémie,
Simon, Celia, Sékou, Guillaume and the new interns Lydia and Thomas, it was a pleasure to
get to know you all.
xxvii Special thanks also to Laurent Maron and Christos Kefalidis, not only for welcoming me in
Toulouse and teaching me the dark arts of DFT, but also for the good times. Without you,
my Ph.D would have never been the same. Special thanks also go to Douglas W. Stephan
for being such an open minded, down to earth person. It is really a great honor and pleasure
to have the opportunity to work with such an outstanding person and chemist. Similarly, huge
thanks to Alex Pulis for the time invested in the hydrogenation project, it was a real pleasure
collaborating with you and I sure hope we will have the opportunity to do so again in the
future. Special thanks also to the team in Toulouse (Didier Bourissou, Ghenwa Bouhadir and
Richard Declerq) for the collaborative effort on pinpointing the role of the formaldehyde
adduct. I am also highly indebted to the research professionals in the department. Pierre
Audet and Wenhua Bi for their continuous help with the NMR and X-ray experiments.
I would also like to thank the members of my thesis committee, professors Douglas W.
Stephan, Peter H. McBreen and Stephen Westcott for taking some of their precious time for
reading my thesis and coming to my defense.
None of this would have been possible without the support of my family throughout the
course of my undergraduate and graduate studies. In fact, without this support, I don’t think
I would have even set foot in a university! Thank you so much, mom, dad and France for
believing in me. I was doing this for you guys and I still am, but thanks to your continuous
encouragement to push forward, I finally found what I was meant to do. Also thanks to all
the Aubut and Courtemanche family, my aunts, uncles, and everyone else! I cannot continue
without thanking my friends who made the last 7 years in Québec city seem like nothing at
all. Thanks for the all the fun Antoine, Félix, Pierre, Oli, Vince, Marc-Alex, Frank, J-R,
Thomas and all the others… Last but not least, the continuous support through the good and
hard times, the encouragement as well as the understanding of my life partner Maude is what
keeps me going every day. Warmest thanks to you my love.
xxviii Inserted research articles and author contributions:
The following section lists the published work that is included in every chapter as well as the
contribution of authors for every publication. All of the publications that are included in this
thesis have been published at the time this thesis is being deposited.
Chapter 2: Courtemanche, M.-A.; Larouche, J.; Légaré, M.-A.; Wenhua, B.; Maron, L.;
Fontaine, F-G
A Tris(triphenylphosphine)aluminum Ambiphilic Precatalyst for the
Reduction of Carbon Dioxide with Catecholborane. Organometallics, 2013, 32, 6804-6811.
Invited contribution
Author contributions: Most of the experimental work was made by MAC but MAL and
Jérémie Larouche (JL) helped with the synthesis of compounds. JL performed the V.T.
experiment. The manuscript writing and editing was mostly done by MAC and FGF. The
manuscript was reviewed and edited by MAL. All of the DFT calculations was done by
MAC. Laurent Maron (LM) supervised the DFT work.
Chapter 3: Courtemanche, M.-A.; Légaré, M.-A.; Maron, L.; Fontaine, F-G. A Highly Active
Phosphine–Borane Organocatalyst for the Reduction of CO2 to Methanol Using
Hydroboranes J. Am. Chem. Soc. 2013, 135, 9326-9329.
Author contributions: The catalyst was discovered, isolated and crystallized by MAC. While
the catalytic activity was discovered and the first catalytic experiments were done by MAC,
most of the experimental screening and the in-situ generation of the formaldehyde adduct
were done by MAL. The catalytic tests with methyl formate and the DFT calculations were
performed by MAC. The manuscript was written by MAC, FGF and MAL. LM supervised
the DFT work.
Chapter 4: Courtemanche, M.-A.; Légaré, M.-A.; Maron, L.; Fontaine, F-G. Reducing CO2
to Methanol using Frustrated Lewis Pairs: On the Mechanism of Phosphine-Borane
Mediated Hydroboration of CO2. J. Am. Chem. Soc. 2014, 136, 10708-10717.
Author contributions: All of the DFT calculations as well as all the experimental work were
performed by MAC. MAL participated in key discussions involving thermodynamics,
xxix kinetics, data analysis and possible reaction pathways. The manuscript was written by MAC,
with revisions and edits by FGF and LM. LM supervised the DFT work.
Chapter 5: Courtemanche, M.-A.; Légaré, M.-A.; Rochette, E.; Fontaine, F.-G.
Phosphazenes: efficient organocatalysts for the catalytic hydrosilylation of carbon dioxide
Chem. Commun, 2015, 51, 6858-6861.
Author contributions: The general concept of the reaction was developed by MAC and MAL.
The isolation of the reaction products with CO2 was made by MAC. The catalytic screening
experiments were performed by MAC and Étienne Rochette (ER). The manuscript was
written by MAC, then reviewed and edited by all the authors.
Chapter 6: Courtemanche, M.-A.; Pulis, A. P.; Rochette, R.; Légaré, M-A.; Stephan, D. W.;
Fontaine, F.-G. Intramolecular B/N Frustrated Lewis Pairs and the Hydrogenation of
Carbon Dioxide. Chem Commun, 2015, 51, 9797-9800
Author contributions: The concept and original reactivity of CO2 hydrogenations were
developed by MAC. Alexander P. Pulis (APP) and MAC performed most of the experimental
work, with occasional synthetic help from ER and MAL. The manuscript was written by
MAC and heavily edited by Douglas W. Stephan (DWS) after discussions. All the authors
reviewed and edited the manuscript prior to publication. Most of the DFT calculations were
performed by ER under the supervision of MAC.
Note: the data collection and refinement as well as resolving of all the crystal structures were
made by Wenhua Bi.
xxx 1 Introduction
1.1 A note on chronology
The field of carbon dioxide reduction that is being covered by this thesis has been evolving
at an incredible pace since the beginning of my doctoral studies. As such, it would be difficult
for the reader to properly situate the present work if all of the modern systems were presented
in the introduction. For these reasons, many results by other research groups will be reported
in the main chapters according to the chronological order in which the work was published.
Therefore, the literature review in the introduction will be mostly limited to contributions
dating before the first results contained in this thesis were published. Again, this is so that
the reader may have a better sense of the mindset in the field at the time the work was taking
place. Before each new chapter, a short introduction section describing the advances in the
field will be included in order to get the reader up to speed on the developments.
Occasionally, anachronisms relating to future systems may be included for a more in-depth
analysis. In such cases, the timescale will be explicitly discussed as to avoid confusion.
1.2 A note on compound identification
In this thesis, every chapter will present a unique numbering of compounds as to avoid the
necessity of going back and forth through chapters to be reminded of the structure of a certain
compound. In the cases where compounds appear in more than one chapter, they will be
explicitly presented once again and they will be assigned a new number for the chapter.
1 1.3 General introduction
The industrial revolution is one of the most important events in human history.1 The transition
to new industrial processes involving the use of steam or coal powered machines and large
scale chemical processes allowed the rapid development of manufacturing industries. For
instance, the large scale production of acids such as sulfuric acid or polyvalent bases like
sodium carbonate proved critical in the early development of the textile, soap and paper
industries.2 From that point, the fields of science and technology started evolving
dramatically. Continuous industrialization, development and evolution has consistently led
to a significant increase of living standards and consequently, to an uninterrupted increase of
global population. While the world population today stands around 7 billion, it is expected
to increase to up to 11 billion by 2050, a 57% increase in less than 40 years. 3
Such a striking expansion inevitably leads to an augmented need for primary resources. In an
increasingly technological society, energy has become one of the most important of such
resources. In 2004, the worldwide energy consumption was about 15 terawatts (TW). In
2025, it is expected that this number will increase to 21 TW and even reach 30 TW in 2050,
doubling the energy consumption in less than 50 years. The production of such an enormous
amount of energy would require 30 000 nuclear power plants of 1 gigawatt (GW) annual
output, meaning that one such power plant would need to be built every single day for the
next 50 years in order to meet the energetic demand.4
Today, over 86% of the energy that is being used originates from some form of fossil fuel
such as petroleum, oil, natural gas, coal and others.5 The limited amount of these resources
stands in contrast to the ever growing energy demand of society, up to a point where the
earth’s natural reserves are threatened of depletion.6 The burning of tremendous amounts of
fossil fuels is not only a short term solution, but is also at the origin of many serious
environmental problems. Thus, one of the most critical issues of the 21st century is to identify
how the energetic needs of our society can be satisfied without jeopardizing the future of our
planet.
2 1.4 Carbon dioxide and the greenhouse effect
Burning hydrocarbons generates primarily water (H2O) and carbon dioxide (CO2), which are
released in earth’s atmosphere. While CO2 is essential to life, it can absorb thermal radiation
from the surface of the earth and re-radiate this energy to the lower atmosphere, resulting in
a net increase of earth’s temperature, a process called the greenhouse effect.7 Emissions of
this greenhouse gas (GHG) have been increasing constantly since the industrial revolution
and the concentration in the atmosphere has recently reached over 400 ppm for the first time
in the history of mankind (Figure 1-1 a).8 Up until now, most of the negative effects from the
increase in CO2 concentration have been mitigated by natural CO2 sinks such as oceans, trees
and soil.9 Still, the increased amount of CO2 in the atmosphere has already led to an overall
rise of temperature on a planetary scale (Figure 1-1 b,c).10
Figure 1-1: a) CO2 concentrations in the atmosphere b) Global temperature anomalies c)
Trend of global temperature variation.
3 The consequences of this global warming, which will worsen as the capacity of the natural
sinks reach their limit, include the retreat of glaciers, the rise of sea levels, changes in
precipitation patterns, ocean acidification and species extinction due to a quick shift of
temperature regime.9 While they are difficult to predict, the ramifications of such
environmental disasters will certainly engender serious consequences on the lifestyle of
mankind.
The sequestration and underground storage of carbon dioxide was proposed as a possible
avenue to reduce the atmospheric concentration of this GHG. Unfortunately, such a solution
remains temporary and presents serious security issues.11 Another solution that comes to
mind is the limitation of CO2 emissions. Fortunately, technologies toward the direct recovery
of carbon dioxide from large industrial emitters have made significant progress during the
past few decades, making it possible to significantly reduce emissions. One of the most
promising avenues for CO2 capture was first invented by R. R. Bottoms in 1930.12 While the
process was continuously optimized until today, the core concept relies on the same principle:
CO2 is transformed into carbamates by reaction with an amine and is subsequently released
by heating (Scheme 1-1).13
Scheme 1-1: Schematic representation of the amine scrubbing process.
Unfortunately, except for the carbon tax concept, of which the development is stagnating,
there is very little economic incentive for the industry to recover carbon dioxide emissions
as this waste product does not have significant commercial value. However, the possibility
of transforming it in value added products could provide an incentive for industrial
companies to recover the greenhouse gas since the recovery process would become
lucrative.14 In fact, a number of processes utilize CO2 directly as a feedstock for the synthesis
of fine chemicals, but since these are mostly carried out on relatively small scales, they are
of limited interest in the fight against global warming.15
4 The industrial processes that uses by far the most carbon dioxide are the synthesis of urea
and propylene polycarbonate.16,17 In 2004, 72 megatonnes (Mt) of CO2 were used to produce
urea and as much as 100 Mt were used to produce propylene polycarbonate.18 Looking at the
bigger picture and comparing these numbers to the anthropogenic emissions of 25.6
gigatonnes (Gt) per year, we find that the combined processes utilize barely 0.7% of the
excess CO2 that is produced by humans every year.
In a serious analysis of CO2 transformation products, the economical aspect must not be
neglected. In order for the market to stay balanced, the offer must not exceed demands, but
in a scenario where gigatonnes of products must be used to have an actual impact on the
atmospheric concentrations, finding a product in continuous demand on such a scale leaves
very few candidates. Even though the transformation of CO2 into value added products such
as carbonates, esters, lactones, carbamates, pyrones, isocyanates and many more commodity
chemicals provides a promising, green alternative to the actual processes, the scale at which
these products are used is not sufficient to have a significant impact on the CO2 concentration
in the atmosphere.19 For this reason and to remain succinct, this thesis will mostly be directed
towards the transformation of CO2 to highly energetic material such as methane, methanol
and others.
1.5 Using CO2 as an energy vector: the methanol economy
Indeed, one of the only fields in which value added products derived from carbon dioxide
could be used on a scale sufficient to have an impact on the atmospheric concentrations is
the field of energy. In fact, while most of the energy sources used by mankind produce either
heat or electricity, the latter cannot be efficiently stored for prolonged periods of time.
Furthermore, the transportation of electricity over long distances is inefficient and wasteful.20
Consequently, there is much room for improvement in terms of energy storage, transportation
and distribution. Hydrogen (H2) has been suggested as a clean, renewable and green
alternative energy medium.21 However, despite intense investments from governmental and
private instances as well as many efforts in terms of research and development, a number of
underlying fundamental problems associated with the use of H2 remain unsolved. Hydrogen
is a highly volatile and explosive gas, raising safety concerns for its large scale transportation
or storage. Furthermore, current automotive technologies are not adapted to use hydrogen as
5 a fuel and the required technology transfer to replace service-stations would engender
astronomical costs.20
On the other hand, the use of CO2 to produce energy vectors, that is the storage of energy in
carbon dioxide through chemical transformations followed by the release of said energy by
combustion, could provide an entirely carbon neutral strategy to the energy storage problem.
Indeed, the simple concept of transforming the greenhouse gas into highly energetic
hydrocarbons such as methane (CH4) or methanol (CH3OH) would provide a very practical
way to store energy (Figure 1-2).
Figure 1-2: Conceptual representation of the use of CO2 to produce energy vectors.
While the use of methane gas raises concerns similar to those that were described for
hydrogen, methanol would be the ideal candidate due to its similarity with the hydrocarbons
that are used on a massive scale every day. It is a relatively non-toxic liquid that can easily
be stored and transported on a large scale.22 The combustion of methanol generates only
water and carbon dioxide, making it a very green alternative to hydrocarbons which generate
polluting mono-nitrogen oxides (NOx), sulfur dioxide (SO2) and other polluting particles.
Furthermore, very little modifications would be required to incorporate the use of methanol
into everyday life. A simple, cheap reconfiguration of combustion engines would allow
modern cars to run exclusively on methanol fuel. An investment of only $50 000 would be
required to transform a regular hydrocarbon dispensing based station into a fully functional
methanol dispensing gas station, a sharp contrast with a transformation into a hydrogen based
gas station necessitating an investment of $1 000 000.22
6 The use of methanol as a fuel is far from being a new concept. In fact, one of the world’s first
cars, the Ford model T was powered by methanol fuel. With the technological advances of
the modern world, methanol can now be used from public transportation buses to highly
performant race cars.20
On the other hand, hydrocarbon based fossil fuels offer much more than energy. From
plastics to synthetic materials, from polymers to pharmaceuticals, hydrocarbon derivatives
are essential to everyday life. Fortunately, most of these can easily be derived and produced
from methanol, making it possible to imagine an economic system based solely on the use of
methanol as a raw material to replace fossil fuels, a concept developed and promoted notably
by Nobel laureate George A. Olah and defined as ‘’The methanol economy’’.20,22–24 Through
the methanol to olefin process (MTO), methanol can be converted to ethylene or propylene,
generating the required basic building blocks for most of the petrochemical based products.25
Furthermore, methanol can be used as synthetic material for the generation of commodity
chemicals such as polymers, acetic acid, paints, dyes, adhesives and much more. Exploiting
the concept of the methanol economy to its full potential would allow us to become entirely
independent from fossil fuels and stop emitting carbon dioxide in the atmosphere, slowing
down global warming significantly. All of this without affecting our capacity to produce
hydrocarbon based commodity chemicals.
1.6 Challenges in the transformation of carbon dioxide
In order to become entirely independent from fossil fuels and generate a sustainable
economy, the energy must come from a cheap, green and renewable source. Much research
is ongoing towards the use natural sources of energy such as solar, hydroelectric, wind and
geothermal processes.26 These technologies need to be developed to provide the world with
the required energy, but while this problem is of fundamental importance, it is beyond the
scope of this thesis to discuss the details of such technologies.
Indeed, another underlying problem needs to be solved before clean energy can be
implemented into the methanol economy. It is not a coincidence that such a tremendous
amount of CO2 ends up in the atmosphere. Factually, carbon dioxide is a very stable molecule
(ΔHform=-392.5 kJ.mol-1). Therefore, its transformation often requires highly reactive
catalysts, high pressures and high temperatures.
7 Such prerequisites are problematic as a lot of energy is being wasted. For instance, methanol
is currently produced on a very large scale (100 Mt/year), mostly by steam reforming, a nonsustainable process that combines non-renewable methane from natural gas and water to
generate methanol, hydrogen and water. The process involves a Cu/ZnO/Al2O catalyst and
operates at high temperature (250 °C) and high pressure (250 MPa).27 Other technologies
involving the use of hydrogen or geothermal heat are emerging, but they are still at the stage
of early development and both still require high temperature and pressures to effect reduction
to methanol. Consequently, new, more efficient processes for CO2 reduction must be
uncovered in order to facilitate the storage of energy into the small molecule. To develop
such processes, the basic properties and reactivity patterns of CO2 must be understood. Only
then can new concepts and avenues towards its reduction can be revealed and it is around the
context of opening such new avenues that this thesis is constructed.
1.7 The carbon dioxide molecule
Carbon dioxide is a thermodynamically stable molecule that belongs to the D∞h point group
and is linear in its ground state, making it non-polar. However, due to the possible resonance
forms of carbon dioxide and the electronegativity of the oxygen atoms, the molecule
possesses an ambiphilic character. Indeed, the oxygen atoms are slightly nucleophilic, while
the carbon atom itself is electrophilic (Figure 1-3). Therefore, two main strategies can be
employed in order to promote a reaction with the molecule. First, the oxygen atoms could
bind to an electron deficient active center, or an electron rich atom could interact with the
carbon atom.
Figure 1-3: Representation of the ambiphilic character of the CO2 molecule.
8 The Walsh diagram of CO2, which plots the change of molecular orbital energy levels with
respect to the molecule’s geometry is depicted in Figure1-4. A close look at the diagram
reveals that upon bending of the CO2 molecule, the energy variation of the π orbitals is much
more pronounced than that of the σ orbitals. In fact, the sigma (σ) orbitals show little to no
destabilization, while the degeneracy of the pi (π) orbitals are split upon bending (2πu to 6a1).
For the highest occupied molecular orbital (HOMO), it is clear that the bending is not
favorable since both components are energetically higher upon bending, meaning that if one
binds the CO2 molecule through oxygen interaction with an electron deficient molecule, the
molecule will remain essentially linear.
Figure 1-4: Walsh diagram of the CO2 molecule.
9 In contrast, while one component of the lowest unoccupied molecular orbital LUMO remains
largely unaffected, the out of plane component (corresponding to the anti-bonding orbital
localized on the carbon atom) is drastically stabilized, to the point where it starts mixing with
other bonding orbitals. Translating this observation in terms of reactivity reveals that
introduction of electron density in this LUMO orbital will lead to bending of the CO2
molecule, drastically changing its fundamental properties. The ambiphilic nature of carbon
dioxide and the consequences that stem from this unique characteristic will be underlined
throughout the introduction since it constitutes the foundation of the research project
motivating the work described herein.
1.8 Organometallic binding of carbon dioxide
While many believe that Aresta was the first to trap carbon dioxide using the now well-famed
electron-rich Ni0 species depicted in Scheme 1-2,28 the coordination of carbon dioxide to
organometallic species was discovered almost a decade before this seminal discovery. In a
1972 review, Volpin already describes a number of transition metal CO2 adducts, albeit with
very limited knowledge of the involved coordination modes.29 One year before, the first
crystallographically characterized complex of CO2 was reported. Unfortunately,
the
limitations of the method at the time did not allow the authors to unambiguously determine
the coordination mode of the carbon dioxide moiety.30 In 1974, Floriani was already taking
advantage of the ambiphilic nature of CO2 and reported a cobalt salen complex that was
proposed to bind CO2 through nucleophilic interaction of the electron rich cobalt center to
the electrophilic carbon atom of CO2. The negative charge on the oxygen atoms was proposed
to be stabilized by the sodium counter-ion. 31 Nonetheless, Aresta’s complex was the first
completely structurally characterized metal complex of CO2 and constitutes a major
breakthrough in the field (Scheme 1-2).
Scheme 1-2: Generation of Aresta's complex.
10 In this complex, the nucleophilic and electrophilic properties of CO2 are both exploited. In
the years that followed, the rich reactivity of carbon dioxide with metal complexes was
slowly uncovered. The key results as well as the coordination modes for a variety of transition
metal centers are summarized in Table 1-1. 19
11 Table 1-1: Summary of key isolated transition metal CO2 complexes.
Coordination mode Structure
Metalref
C-O bond length (Å)
η1-O
U32
1.122(4), 1.277(4)
η1-C
Rh33, Ir34
1.20(2), 1.25(2)
η2-C-O
Ni,35–37 Rh,38 Fe,39 Pd40 1.17, 1.22
μ2-η2
O
Pt 41,42 Ir/Zr,43 Ir/Os,44
O M2
Rh,45 Ru46,47
M1 C
μ2-η3
Re/Zr,48 Ru/Zr,49
1.229(12), 1.306(1)
1.285(5), 1.281(5)
(Ru/Ti, Fe/Zr, Fe/Ti)50
μ2-η3
Re/Sn,51 Fe/Sn52
1.269(1), 2.257(7)
1.252(3), 2.394(2)
μ3-η3
Os,53,54 Re55
1.276(5), 1.322(5)
1.28, 1.25
μ3-η4
Co31,56,57
1.20(2), 1.24(2)
μ4-η4
Ru58
1.283(2), 1.245(2)
μ4-η5
Rh/Zn59
1.29(14), 1.322(1)
12 Looking at Table 1-1, it can be determined that by increasing the coordination number of
CO2, the C-O bond order diminishes, resulting in longer C-O bond lengths. Even with such
a wide range of unique CO2 complexes, none exhibited catalytic activity in the reduction of
the small molecule. However, some metal complexes were developed for the catalytic
transformation of CO2 into value added chemicals. For an in-depth discussion on the
formation of value added products such as carbonates, esters, lactones, carbamates, pyrones
and isocyanates, the reader is directed to recent review articles.60,61 Indeed, as it was stated
before, the scope of this thesis is limited to the transformation of CO2 into energy rich
materials.
1.9 Carbon dioxide for energy storage: general aspects
Simply put, the combustion of hydrocarbons is an oxidation reaction. Let us take the
combustion of methane as an example: in the presence of heat, the carbon atom is oxidized
from its -4 state to +4 oxidation state, an exothermic reaction that generates energy in the
form of heat (Scheme 1-3).
Scheme 1-3: Aerobic combustion of methane.
It is logical that in order to store energy in the CO2 molecule, the opposite reaction must take
place. In order to achieve this, some form of energy is required. Also, due to the high stability
of the CO2 molecule, a catalyst is often required to accelerate the reaction. In Scheme 1-4,
the reduction of CO2 using hydrogen as primary energy source is depicted. By transferring
two, four, six or eight electrons to carbon dioxide, it would be possible to generate formic
acid, formaldehyde, methanol or methane respectively. 62,63
13 Scheme 1-4: Stepwise CO2 hydrogenation.*
*Energies are reported in kJ mol-1. Thermodynamic values are given for aqueous solutions of H2 and
CO2, while the products are in liquid phase.
This multi-step process is at the core of the challenge of CO2 reduction. In fact, a dichotomy
exists in this reaction: the inherent stability of the CO2 molecule mandates the use of highly
reactive catalysts, yet the catalyst must remain stable in the presence of all the fundamentally
different reaction products. Another major challenge resides in the design of a catalyst that
will selectively generate a single desirable reduction product.
1.10 Metal enzymes for CO2 reduction
As early as 1976, evidence for the possibility of a catalytic CO2 reduction process was put
forward. An enzyme, formate dehydrogenase (FateDH) was shown to catalytically reduce
carbon dioxide to formate by using a reduced form of nicotinamide adenine dinucleotide
(NADH) as an electron source.64 In 1999, the ingenious use of a cascade of three different
enzymes, namely FateDH, formaldehyde dehydrogenase (FalDH) and alcohol dehydrogenase
(ADH) catalyzed the reduction of CO2 all the way to methanol.65 However, the use of the
NAD/NADH couple for reduction is not a viable option in terms of large scale transformation
due to the high cost of this energy source.66 Nonetheless, an in-depth study of these biological
systems could lead to hints regarding the important parameters in CO2 reduction chemistry.
In 2007, a major breakthrough in enzyme mediated carbon dioxide reduction was made when
an intermediate of CO2 reduction was crystallized, allowing the postulation of a plausible
mode of action for the carbon monoxide dehydrogenase enzyme.67
14 In one of the key steps of the reaction, it was found that the ambiphilic activation of carbon
dioxide, which can be defined as the simultaneous activation by an electron rich and an
electron withdrawing fragment, was of crucial importance. In Figure 1-5, which represents a
key transition state, it is possible to observe that the electrophilic iron center interacts with
the oxygen atom of CO2 while the nucleophilic nickel center interacts with the carbon atom.
Figure 1-5: Transition state in the process of CO2 reduction by CO dehydrogenase.*
*Figure reproduced from (Jeoung, J.-H.; Dobbek, H. Science, 2007, 1461-1464). Reprinted with
permission from AAAS.
1.11 Transition metal catalyzed reduction of CO2 to energy rich materials
Meanwhile, a number of heterogeneous catalytic systems capable of hydrogenating carbon
dioxide to methanol have been reported. Yet, such catalysts require elevated operating
pressure (40-50 atm) and temperatures (200-250°C). Not only does the high temperature
regime consume important amounts of energy, but it also limits the theoretical yield of the
reaction since the products are entropically disfavored.62
These limitations encouraged chemists to develop homogeneous catalysts capable of
operating at lower temperatures and pressures, but the task proved challenging. Perhaps
inspired by the enzymatic cascade stratagem, Sanford reported in 2011 that the use of three
homogeneous catalysts (PMe3)4Ru(Cl)(OAc), Sc(OTf)3, and (PNN)Ru(CO)(H) PNN = (2(di-tert-butylphosphinomethyl)-6-(diethylaminomethyl)pyridine) operating in sequence
could lead to the formation of methanol (Scheme 1-5).68 While interesting in concept, the
turnover number (TON) remained limited (TON = 21) and the reaction conditions were still
far from ideal (135 °C, 10 atm H2, 30 atm CO2).
15 13CO
2
A, B , C
12CH OH
3
13CH OH +
3
H2O
P(t-Bu)2
PMe3
Me3P
PMe3
Ru
Sc(OTf)3
Me3P
OAc
Cl
(A)
(B)
N
Ru CO
N
(C)
Scheme 1-5: Cascade catalytic hydrogenation of CO2 to methanol.
Utilizing hydrosilanes as a primary source of chemical energy and an iridium catalyst,
Eisenberg managed to reduce CO2 to methoxysilanes, products that can be readily
hydrolyzed to methanol, albeit with limited efficiency.69 Then, it was shown that the use of
a zirconium catalyst in combination with the highly Lewis acidic B(C6F5)3 (BCF) promoted
the hydrosilylation of carbon dioxide all the way to methane, reaching a TON of 225.70 In
2010, Guan and collaborators broke the record in efficiency for the reduction of CO2 to
methanol. Using catecholborane (HBcat) as a reducing agent, they were able to show that a
nickel pincer catalyst could reduce CO2 to methanol derivatives with an impressive turnover
frequency (TOF) reaching 495 h-1 (Scheme 1-6).71,72
Scheme 1-6: Catalytic hydroboration of CO2 by a Ni pincer complex.
16 While the reduction of CO2 to energy-rich materials was still only emerging, a number of
highly active systems had been reported for the hydrogenation of CO2 to formates.73–79 Even
though the detailed description of such systems is beyond the scope of this thesis, one
particular system must be discussed in more detail.
In 2011, Hazari and Crabtree reported an iridium based pincer catalyst exhibiting impressive
reactivity for the hydrogenation of CO2 to formic acid (Scheme 1-7).80 As was discussed
earlier, the generation of formic acid is thermodynamically disfavored and must be trapped
with a stoichiometric amount of base. Nonetheless, the system reached turnover numbers up
to 348 000 with a TOF of 18 000 h-1. Once again, drawbacks of the system include high
operating temperature 185 °C and high pressure (55 atm).
Scheme 1-7: Dual interaction of reduced CO2 with an Ir pincer complex.
However, the key observation in this system was the particular mode of action of the catalyst
where CO2 interacts through hydrogen bonding with the N-H moiety of the catalyst. The
authors suggest that this weak interaction is key in promoting efficient catalysis, once again
underlining the importance of synergistic interactions with CO2.
1.12 Photochemical reduction of carbon dioxide
The ability of certain metal centers or semiconductors to absorb sunlight has led to
development of catalysts for the photochemical conversion of CO2 to energy enriched
reduction products using water as the primary electron source. Much advance has been made
in the last two decades and the subject has been recently reviewed.81,82 It is unfortunate that
in most cases, low light efficiency, low quantum yields and limited selectivity are observed.
17 While the field represents a very interesting strategy for cheap, efficient recycling of CO2,
much remains to be done to achieve the activities required for large scale applications
1.13 Transition-metal free transformations of carbon dioxide
As early as 1950, it was reported that lithium borohydride could reduce carbon dioxide to a
formate derivative which could then be hydrolysed to formic acid (HCOOH).83 Shortly after,
it was reported that sodium borohydride could also promote the reduction of CO2 to a mixture
of formate products and that the product distribution was highly dependent on the employed
conditions.84,85 In 1967, a similar strategy was employed, but this time for the stoichiometric
reduction of CO2 to formic acid in aqueous conditions.86 Very recently, this aqueous
reduction chemistry was revisited, but the activity could not be expanded beyond
stoichiometric transformations.87,88 A drastically different strategy has also been used to
effect stoichiometric CO2 reduction. Indeed, it was recently reported that in the presence of
CO2, highly Lewis acidic silylium ions react to yield a mixture of benzoic acid, formic acid
and methanol after aqueous workup.89
Pushing the silylium ion strategy a step further, a catalytic system for the Lewis acid
catalyzed reduction of CO2 could be developed. By using a high loading of AlEt2+ (10 mol%),
the catalytic hydrosilylation of CO2 to methane and several solvent alkylation by products
was carried out by Wehmschulte and co-workers. 90 Using control experiments, the authors
were able to obtain evidence for the catalytic role of the Lewis acidic aluminum species as
independently synthesized silylium ions exhibited drastically inferior reactivity. While
interesting in concept, the system suffers from a number of drawbacks including slow activity
(50% conversion after 14h at 80 °C) and lack of selectivity (Scheme 1-8). A second
generation catalyst was developed by the same research group, but the use of a highly
sterically congested catalyst [Al(2,6-Mes2C6H3O)2Al]+ (Mes=2,4,6,-trimethylphenyl) still
led to important selectivity problems. 91
Scheme 1-8: Lewis acid catalyzed hydrosilylation of CO2.
18 In contrast to the Lewis acid activation strategy, an antagonistic approach relying on Lewis
base activation was developed for the catalytic reduction of CO2 to high energy content
molecules. The use of an N-heterocyclic carbene (NHC) catalyst presumably promoted the
hydrosilylation of CO2 to methoxysilanes with over 90% selectivity. The authors proposed
the reduction to occur either through coordination of the NHC moiety to the reducing silane,
generating a nucleophilic hydride capable of CO2 reduction (Figure 1-6, pathway A), or
through coordination of the NHC to CO2, rendering the molecule more reactive towards
reduction. (Figure 1-6, pathway B).
Figure 1-6: Proposed mechanistic pathways for the base-catalysed hydrosilylation of CO2.
Wang et al. supported that pathway A was more plausible in a theoretical study,92 arguing
that such a mechanism would account for the experimental evidence that bulkier silanes
exhibited lower reactivity. While the record-breaking TON and TOF of 1840 and 25.5 h-1
exceed the performance of most catalysts, the important solvent effects that were observed
experimentally remained mostly unexplored. Indeed, a recent report by our group raises
certain concerns regarding the reported turnovers. For an in-depth discussion, the reader is
directed to chapter 6.
Still, these results suggest that carbon dioxide reduction can be promoted both by Lewis acids
and bases. The general strategy that is exploited in this thesis relies on using both of these
activations in a synergistic fashion. But before getting into more details, an introduction to
Frustrated Lewis Pair (FLP) chemistry is mandatory.
19 1.14 A short history of Frustrated Lewis Pair chemistry
Molecules were first classified as electron donors or acceptors by Gilbert Lewis, thus creating
the foundation for the concept of Lewis acids and bases.93 Lewis acids are generally
compounds possessing an empty orbital that can accept electron density from an electron rich
molecule. Alternately, a Lewis base will typically possess a free doublet of electrons capable
of giving electron density to an electron poor molecule. A more general way of classifying
these molecules would be to say that Lewis acids have a low lying LUMO while Lewis bases
have a high lying HOMO. In most cases, when combined, Lewis pairs form Lewis adducts,
with the electron density of the Lewis base fulfilling the electronic deficiency of the Lewis
acid, resulting in significant thermodynamic stabilization (Scheme 1-9).
Scheme 1-9: Formation of a classical Lewis adduct.
In 1942, Brown and co-workers made an interesting observation: lutidine (2,6dimethylpyridine) forms a Lewis adduct when combined with trifluoroborane (BF3) but does
not react in the presence of trimethylborane (BMe3). The lack of reactivity between lutidine
and BMe3 was attributed to the steric congestion that prevented the formation of a classical
Lewis adduct.94 More than a decade later, Wittig and Benz noted that triphenylphosphine
(PPh3) and triphenylborane (BPh3) showed no evidence of adduct formation when combined.
Nonetheless, when a mixture of the two compounds was exposed to benzyne which was
generated in-situ from 2-iodobromobenzene and Mg0, a zwiterrionic phosphonium-borate
(C6H4)(PPh3)(BPh3) was isolated (Scheme 1-10).95
20 Scheme 1-10: Trapping of benzyne by a phosphine-borane Lewis pair.
In 2003, Piers and Roesler introduced an interesting concept by synthesizing a phenylene
bridged ambiphilic molecule in which the Lewis acidic and Lewis basic centers were
precluded from forming an adduct by geometric constraints (Scheme 1-11).96 In this original
publication, Piers suggests the utilization of these Lewis pairs to activate hydrogen and trap
reactive intermediates. Unfortunately, the compound proved unreactive towards hydrogen,
but did afford zwiterrionic species in the presence of hydrochloric acid (HCl) or water.
Scheme 1-11: Trapping of hydrochloric acid and water using an amino-borane Lewis pair.
Perhaps the most interesting part of this report is the passage where the authors conclude the
following: ‘‘The basicity of the nitrogen center in an aminoborane would have to be
significantly higher in order to thermodynamically favor the formation of a dihydrogen
adduct…’’. Indeed, this conclusion will be confirmed almost a decade later by Repo and coworkers who managed to activate H2 by simply changing the phenyl groups on nitrogen for
methyl groups.97
21 A few years later, Stephan and Welch used a different approach to make the ground-breaking
discovery that it was indeed possible to activate molecular hydrogen without the use of
transition metals. In the original publication, they were able to show that highly reactive
Lewis acids and bases could be prevented from reacting with each other by introduction of
steric bulk around the active centers. In such cases, the active centers could act synergistically
to split hydrogen into two fragments: a proton and a hydride (Scheme 1-12).98
Scheme 1-12: Representation of Frustrated Lewis Pair reactivity and hydrogen splitting.
Following pioneering reports on the splitting of molecular hydrogen using FLPs, this class
of species has been shown not only to activate several other substrates, but also to catalyze a
number of hydrogenation reactions.99–102
Exploiting the ambiphilic nature of the carbon dioxide molecule, Erker and Stephan
demonstrated that FLPs consisting of bulky phosphines and highly Lewis acidic boranes
could bind carbon dioxide. One of such complexes and its reaction with carbon dioxide is
illustrated in Scheme 1-13.103 Interestingly, these species release carbon dioxide at higher
temperature.
t-Bu
P
t -Bu
t -Bu
C6 F5
B
C6F5
C6F5
CO 2, 25°C
80°C
O
C O
(t-Bu3 )P
B(C6 F5)3
Scheme 1-13: Reversible metal-free activation of CO2 by FLPs.
22 1.15 Carbon dioxide activation and reduction using Frustrated Lewis Pairs
Less than 6 months after this seminal discovery, a first example of stoichiometric CO2
reduction by FLPs was reported by Ashley and O’Hare. By combining the highly Lewis
acidic B(C6F5)3 and TMP (TMP=2,2,6,6-tetramethylpiperidine), in the presence of CO2 and
H2, a formatoborate product was formed upon heating at 110 °C for several hours. Providing
further energy to the system by heating at 160 °C for several days allowed the authors to
isolate methanol in 25% yield after distillation of the reaction mixture (Scheme 1-14).63
Scheme 1-14: FLP mediated CO2 hydrogenation and formation of methanol.
Shortly after, the general strategy was exploited by Piers and co-workers to design the first
catalytic system for CO2 reduction using FLPs. Indeed, by adding an extra equivalent of BCF
to the system along with a hydrosilane reducing agent, the formatoborate species could be
reduced all the way to methane.
104
The observation of multiple reaction intermediates by
NMR spectroscopy allowed the authors to propose a stepwise mechanism which was latter
supported computationally (Figure 1-7).105 Unfortunately, the very limited turnovers ensured
that FLP systems remained subservient to other catalytic processes at the time.
Figure 1-7: Proposed mechanism for the FLP-mediated hydrosilylation of CO2.
23 In another key study, Ménard and Stephan reported a FLP system consisting of PMes3/AlX3
(X=Cl or Br) featuring interactions of the CO2 fragment with two Lewis acids that could
promote CO2 reduction in the presence of ammonia-borane (NH3BH3) to generate
methoxyaluminum species. 106 Unfortunately, hydrolysis was required to liberate methanol,
destroying the FLP system in the process and thus precluding catalysis (Scheme 1-15).
Scheme 1-15: FLP mediated stoichiometric reduction of CO2 by ammonia borane.
Theoretical studies 107,108 and further experimentation with a variety of related Lewis acids
and amine–boranes109 allowed the scientific community to gain more insight into this
reduction process. Although these systems did not exhibit any catalytic activity, the
introduction of hydroboranes as potential reducing agents in FLP mediated transformations
proved to be a turning point. Interestingly, related aluminum based systems were later shown
to perform the stoichiometric reduction of CO2 to CO,110 but the details of these
transformations are beyond the scope of this discussion.
1.16 Original hypotheses and motivation
Much before Frustrated Lewis Pair chemistry started to gain in popularity, the Fontaine group
had been involved the use of ambiphilic ligands (intramolecular Lewis Pairs) for applications
in homogeneous organometallic catalysis. Indeed, the first example of an ambiphilic ligand
in catalysis was reported by Fontaine and Zargarian in 2004.111 In this contribution, the
authors laid the groundwork for a field that would become rich and diverse in the years to
come.112–115
24 Motivated by the recent developments in the Frustrated Lewis Pair chemistry involving
carbon dioxide, it is in 2011 that the Fontaine group set out to develop the comprehension of
these emerging systems. It is important to note that at the time that this work was taking
place, only three reports of metal-free CO2 binding had been published (Figure 1-8 a-c).
103,106,116
Figure 1-8: Reported CO2-FLP adducts- October 2010.
In an FLP system, two components will dictate the reactivity: the Lewis acid and the Lewis
base. However, when it comes to FLPs in which the active centers are covalently attached
together, the bridging framework gains a central function in terms of reactivity.
Intramolecular FLPs present clear advantages over their intermolecular counter parts. First,
the entropic challenge associated with the combination of a Lewis acid, a Lewis base and a
reagent in the same transition state is overcome by fixing the Lewis acidic and basic moieties
in a single framework. Consequently, fine-tuning of the framework gives the synthetic
chemist the opportunity to manage the orientation, distance, steric environment and
electronics of the active components. Thus, the incorporation of a bridging framework inbetween the active sites leads to an extra dimension of control on the reactivity of FLP
systems.
25 It was also hypothesized that in the case of some intramolecular FLPs, the important steric
bulk of FLPs was not at all necessary, an idea that went against the fundamental principal of
FLPs. Yet, taking the ambiphilic nature of CO2 into account, one finds that the molecule is
naturally predisposed to interact with ambiphilic systems. While the work described in the
next section was also motivated by the desire to develop the first intramolecular aluminum
containing FLP, Uhl and co-workers reported such a system during the preparation of our
manuscript, albeit using a fundamentally different approach (Figure 1-9 a).117 Two more
intermolecular FLP reports of CO2 activation were also published during the preparation of
the following study (Figure 1-9 b-c). 117–119 illustrating how fast the field was moving.
a)
b)
O
c)
C O
Al(t-Bu)2
(Mes)2P
Ph
Uhl
O BX2
(t-Bu)3P C
O
(t-Bu)3P C
O BX2
X=Cl,C6F5
Stephan
O
B(C6F5)2
O
O B(C6F5)2
(t-Bu)3P
Figure 1-9: Reported CO2-FLP adducts- June 2011.
26 C O R=cycloalkyl
B(C6F5)2R
Stephan
1.17 Non-frustrated aluminum based ambiphilic molecules for CO2 capture
The first efforts of the Fontaine group were directed towards the study of methylene bridged
FLP systems of the general formula (R2PCH2AlMe2)2 where R=Me or Ph.120 Indeed, the
coordination chemistry of these species to a rhodium center had been studied by the Fontaine
group in the past and the species have been known to exhibit dynamic reactivity despite their
dimeric nature.121,122
The ambiphilic compounds (Me2PCH2AlMe2)2 (1) and (Ph2PCH2AlMe2)2 (2) were
synthesised by transmetallation of ClAlMe2 with LiCH2PR2 (Scheme 1-16). Compound 1
had previously been reported by Karsch using a similar strategy.123
Scheme 1-16: Preparation of ambiphilic compounds 1 and 2.
Complete characterization of both compounds confirmed that they were dimeric both in
solution and in the solid state. Indeed, the absence of steric bulk around the active centers
allows dimerization of the compounds, a sharp contrast with all the previously reported FLP
systems. Another key distinction from the preceding systems is the sp3 hybridized methylene
bridging group.
Even though the compounds are dimeric, they both react instantly with carbon dioxide at
room temperature. Interestingly, a reaction intermediate could be trapped by following the
reaction at -35 °C in dichloromethane-d2. Labelling experiments with
13
CO2 allowed the
unambiguous characterization of the reaction products (Scheme 1-17).
27 Scheme 1-17: Reactivity of ambiphilic molecules 1 and 2 with CO2.
It was shown that independently of the substituents on phosphorus, the first intermediate is a
simple CO2 activation product (3 or 4). Indeed, the multinuclear NMR characterization as
well as infrared spectroscopy revealed properties are very similar to those of the previously
reported FLP CO2 adducts. However, in contrast with the previously reported systems which
release CO2 when warmed up, warming of the solution of 3 or 4 leads to a rearrangement of
the CO2 adducts into dimeric aluminium carboxylate species 5 or 6.
Previous studies have shown that the CH2-Al bond in Cp*RhMe2(Me2PCH2AlMe2.DMSO)
where DMSO = dimethylsulfoxide could be cleaved selectively as the methylene group acts
as an alkylating agent.122 Surprisingly, no such insertion could be observed with the
compounds reported by Stephan and Erker. One can assume that the sp3 methylene in 1 and
2 is more nucleophilic and better oriented towards the electrophilic carbon of CO2 than would
the sp2-carbon of the previously reported species. The absence of CO2 insertion into the AlMe bond further supports this argument. Interestingly, changing the methyl substituent on
the phosphorus for less electron donating phenyl groups did not have a significant impact on
the reactivity.
However, one important difference in reactivity was noted when the reaction was carried out
in the presence of a limited amount of CO2. Addition of a single equivalent of CO2 to dimeric
1 affords species 7 in approximately 50% yield after 18 hours. A similar reaction using 13CO2
results in the presence of a triplet at δ 97.1 (1JC-P = 99 Hz) in the 13C{1H} NMR spectrum, a
spectral region typical for acetal moieties, thus confirming that 7 results from the activation
of CO2 by two ambiphilic fragments (Scheme 1-18). The validity of the proposed structure
was further supported by DFT experiments. Similar reactivity was not observed for the Ph
28 analogue. It could be that the higher steric demands or the less electron-donating properties
of an aryl substituted phosphine are in part responsible for the difference in reactivity.
Scheme 1-18: Formation of a new spirocylic CO2 activation product.
Finally, the reaction of CO2 with compound 1 was studied in the solid state in order to prove
that the reactivity was not due to partly dissociated dimers in solution. Exposing a solid
sample of 1 to CO2 followed by characterization by MAS
31
P{1H} solid-state NMR
spectroscopy and infrared spectroscopy revealed the formation of a mixture of compounds
1, 3, 5, and 7.
This first incursion of the Fontaine group in the world of FLP chemistry effectively
demonstrated that the presence of bulky groups on the Lewis pair is not the only requirement
for CO2 activation. The ambiphilic nature of the molecules mitigates the entropic
contribution when compared to intermolecular FLP systems. Furthemore, the oxophilicity as
well as the ability of aluminum to become hyper-coordinate were deemed key factors. But,
the insertion of carbon dioxide into the aluminum-methylene bond is an undesirable process
in terms of catalytic applications. Yet, probably one of the most important conclusions of this
work was that simple and stable arylphosphines were as effective as the more reactive
phosphine analogues.
29 1.18 Scope of thesis
As a whole, the thesis is directed towards developing a better understanding of CO2 reduction
processes and applying the attained knowledge to the design of efficient catalysts. A strong
emphasis is put on reducing CO2 to the high hydrogen content methanol molecule. In truth,
the original intent of the present research project was to develop new transition-metal free
systems for the activation of carbon dioxide. The first results towards that goal were already
presented in the introduction and will continue in chapter 3. Upon realisation of this
objective, the project was steered towards finding an efficient system for the catalytic
reduction of CO2. The work describing how this new undertaking was achieved is described
in chapters 3 and 4. On the other hand, chapter 5 deals with an in-depth study of the factors
governing the reactivity of the unique systems that were discovered. Having developed the
most efficient catalyst for the hydroboration of CO2, the knowledge that was gained through
the process was utilized to expand the scope of reducing agents to hydrosilanes, but a totally
different system had to be employed and is described in chapter 6. Finally, combining all that
was learned through years of developing CO2 reduction catalysts, the project was steered
towards the most challenging frontier in chapter 7: the development of a metal-free system
for the hydrogenation of CO2.
30 2 Experimental methods
2.1 Inert atmosphere chemistry
The work that is described in the present thesis often required handling of reactive material
under an inert atmosphere of dinitrogen in order to prevent decomposition by either
hydrolysis or oxidation reactions. While most of the work reported in chapter 3 involved the
manipulation of highly reactive organometallic species such as aluminum compounds, the
necessity of using inert atmosphere techniques became less and less common as the project
advanced due to the increased stability of the developed systems. Nonetheless, many of the
manipulations that are described required handling of some reagents under an inert
atmosphere and as such, these techniques will be briefly introduced.
2.1.1
Glovebox
Simply put, a glovebox is a very large chamber filled with an inert atmosphere of either argon
or nitrogen. A copper catalyst continuously purifies the atmosphere in order to keep the
concentration of oxygen below 0.1 ppm. A similar purification process involving molecular
sieves ensures that the concentration of water in the box remains under 0.1 ppm. Material is
introduced in the box through an antechamber in which the objects to be entered are subjected
to a series of purges, alternating between inert atmosphere and vacuum. This methodology
allows one to enter or exit the glovebox without ever exposing the atmosphere to the ambient
air.
Figure 2-1: A glovebox workstation.
31 2.1.2
Schlenk line
A schlenk line consists of a dual manifold with several ports. One of the ports is connected
to a vacuum pump while the other is connected to a source of inert gas such as nitrogen or
argon. Stopcocks located on the manifolds allows one to control the atmosphere in closed
vessels without exposing the contents of the vessels to the ambient atmosphere and thus
protecting them from moisture or oxygen. Volatile compounds can be easily evaporated
directly on the Schlenk line since a trap frozen in liquid nitrogen will condense the vapors
before they reach the vacuum pump. Special vessels commonly called Schlenk tubes or flasks
are utilized to handle sensitive materials. These vessels are composed of a closed reaction
chamber that may be connected to the Schlenk line through a sidearm with a stopcock,
allowing one to control the flow of gases to the chamber. J-young NMR tubes can also be
connected to the Schlenk line via a Teflon screw cap that also has the role of preventing the
contents of the tube from coming in contact with the ambient atmosphere.
Figure 2-2: a) A Schlenk line b) Schlenk flasks.
32 2.2 Working with gases
Since the Schlenk line is completely leak-free, it is the ideal tool to manipulate reagents in
the gas phase. Most of the manipulations involving gases such as CO2 or H2 have been carried
out using the Schlenk line. It is possible to determine the total volume of the Schlenk line
relatively easily. One can measure the volume of a sealed bulb using water and standard
glassware. Then, the bulb is connected to a Schlenk line which is under static vacuum. The
pressure variation in the Schlenk line upon opening the stopcock connecting the line and bulb
allows one to calculate the volume of the line precisely using Boyle’s law (P1V1=P2V2). By
using the same equation, one can then readily calculate, through pressure variations, the
amount of gas that is being condensed in a sealed vessel. This is particularly useful for
experiments involving 13CO2 or to calculate the exact amount of gas that is being introduced
in the reaction mixture. Therefore, by combining this technique with the use of J-young NMR
tubes, one can prepare NMR scale reactions that will give tremendous amounts of
information regarding the reactivity of the studied material.
It should be noted that care must be taken when working with highly pressurized gas. For
instance, J-Young NMR tubes and regular Teflon capped Schlenk flasks are limited to 5 atm
of pressure. Going beyond the pressure limits of the vessel may result in violent explosions
with dangerous glassware scattering. Fischer-porter tubes can be used to reach up to 10
atmospheres of pressure, but anything beyond that requires the use of specialized equipment.
33 2.3 Nuclear Magnetic Resonance (NMR) spectroscopy
Experiments carried out in J-young NMR tubes followed by analysis of the reaction mixture
by NMR spectroscopy proved to be the most valuable technique to obtain the results
presented included in this thesis. Indeed, NMR spectroscopy is an incredible tool for organic
and organometallic chemistry. It can provide clues regarding the chemical environment of
proton and carbon nucleus, often allowing one to identify the structure of a molecule using
solely this method However, NMR spectroscopy is not limited to proton and carbon. For this
project, other nucleus such
11
B,
31
P, and were used to obtain a number of important clues
regarding the structures and solution behaviors of the reported compounds. Not only can the
acquired spectra reveal important information about the chemical environment of the nuclei,
but one can also make deductions based on the coupling with other atoms such as 1H or 13C.
Table 2-1: NMR nuclei of relevance to this project.
Nuclei Abundance Spin Relative Receptivity 1
99.985% +1/2 62.89904 13
1.108 +1/2 1.00000 11
80.1 +3/2 2.07734 31
100 +1/2 4.17243 H C B P Furthermore, NMR spectroscopy allows the in-situ monitoring of reactions and as such, it is
possible to obtain important kinetic data by recording spectra as the reaction progresses. It is
also possible to slow down or accelerate exchange processes using variable temperature
techniques. The use of this method allowed us to obtain important information on the
dynamics of the systems presented in this thesis. The use of 13C labeled CO2 also allowed
unambiguous characterisation of most of the CO2 derived products due to the strong
resonance and coupling patterns of the 13C nucleus.
34 2.4 X-ray crystallography
While NMR can provide important information regarding molecules, it is often difficult to
extract very precise structural data from an NMR spectra. Fortunately, the bond lengths and
angles as well as the general conformation of the molecule can be determined using single
crystal X-ray diffraction techniques. The general method consists of exposing a single crystal
of the desired compound to an X-ray beam. If the incoming X-ray beam approaches the
crystal at an angle that satisfies Bragg’s law (nλ=2dsinϴ), the beam will be diffracted and
the scattered beam will be detected. Since crystals consist of a regular arrangement of atoms,
the relative position of the diffracted beams can be used to solve the crystal structure of the
material. Solving the crystal structure allows one to identify the precise conformation of the
molecule in the solid state since it allows you to see the actual molecule. Furthermore, a lot
of key information can be extracted from the bond lengths and angles.
While X-ray diffraction is extremely powerful and provides a lot of information, it also has
its limitations. In order to obtain a crystal structure, a single crystal is required. It is often
very difficult to obtain a single crystal with the quality that is required for a diffraction
analysis. Some compounds actually do not crystallize at all while others require dozens of
crystallization attempts.
Another limitation of the method is that while we can solve the
structure of a molecule, it is important to keep in mind that a crystal structure is only a
snapshot of the molecule in a conformation that is defined by the crystal packing. Molecules
have a dynamic behavior in solution and as such, one must still take this into account when
characterizing a new compound. For these reasons, X-ray diffraction is a tool that is
complementary to NMR spectroscopy. It should be noted that all the structures presented in
this thesis have been solved by Wenhua Bi.
35 2.5 DFT calculations
A complementary technique that has gained a lot of popularity in the last decade is the use
of computational methods to obtain structural information. Density Functional Theory is the
most common and also the most widely accepted method for performing calculations to
predict the chemical properties, thermodynamics and structure of a molecular system. In
contrast with most classical computational methods, DFT does not aim to solve complex
many electron wavefunctions. In fact, these complex calculations are completely bypassed
due to the fact that it is possible to approximate the ground state energy of a system based
only on the electron density using the Hohenberg-Kohn theorems. Simply put, this bypassing
method allows one to save enormous amounts of computation time, making it possible to
modelize complex systems rather accurately. Indeed, through an iterative computation
process, DFT can be used to predict the ground-state (the most stable) structure of practically
any imaginable compound.
A DFT method which is often referred to as a level of theory consists of two main parameters:
the functional and the basis-sets. The functional is the approximation that is used to calculate
the energy of the system. Simply put, it is the actual calculation method that is being used.
In this thesis, the B3PW91 functional is used in the first few chapters due to its proven ability
to predict the thermodynamic properties of main group compounds. However, a thorough
benchmark study of the relative accuracies of a large number of functionals in predicting the
properties of main group compounds was performed by Grimme and co-workers,124
prompting us to switch to the more accurate ω-B97xD functional.
36 The basis-set on the other hand is a set of functions which are combined to create molecular
orbitals. In a molecule, each atom must be assigned at least one-basis function, but more can
be added for a more precise computation. A common addition is a polarization function
which is denoted in this thesis by asterisks (*). One asterisk means that a polarization function
is added to every atom except hydrogen while two of them means that a polarization function
was also added to hydrogen atoms. While the addition of these functions results in increased
computation time, they allow a better flexibility of the calculated orbitals and thus allows a
better description of complex bonding situations.
Similarly, diffuse functions which are denoted by a plus (+) sign can be added to better
describe the orbitals that are far from the nuclei. These functions are very important when
considering charged systems. The calculations performed in this thesis are always performed
at a very high level of theory using the basis-set 6-31++G** which include both the diffuse
and polarization functions. All of the calculations included in this thesis have been performed
by myself or by Étienne Rochette under my supervision.
While the calculation of an optimized structure does not match a crystal structure in terms of
experimental validation, a DFT optimized structure can provide similar information. Still,
the biggest advantage of DFT methods is that it is possible to calculate the electronic energy
of most compounds. It is also possible to calculate the energy of transition states, which
represent an intermediate that is being formed during a chemical transformation. By
compiling this thermodynamic data, it is possible to obtain an energetic profile for a certain
reaction pathway. The calculation and comparison of the energetic profiles for multiple
reaction pathways can be compared, allowing the determination of plausible reaction
mechanisms. It goes without saying that DFT can also be used to predict reactivity trends
and orient the synthetic chemist in the design of molecular systems, making it a very versatile
tool for the modern chemist.
37 Figure 2-3: The cluster that was used for calculations: Colosse.
While DFT is an incredible tool, one must take care of how the results are interpreted. Indeed,
there is always the possibility that some unanticipated reaction may occur in solution.
Furthermore, most calculations are performed in the gas-phase and the behavior of molecules
in solution often greatly differ from their behavior in the gas phase. For this is reason, DFT
methods are best used as a tool that supports experimental evidences.
2.6 Gas chromatography
In some large scale reactions, gas chromatography (GC) was employed to properly quantify
the amount of methanol that was produced during some of the catalytic reductions. In a
typical GC experiment, a carrier gas will carry the mobile phase that contains the analyte (in
our case methanol) through a stationary phase. Depending on the affinity of the molecules
for the stationary phase, different molecules will spend different amounts of time in the
column. This allows for a very efficient separation of different compounds and the molecules
can thus be characterized by their retention time. A flame ionization detector (FID) will detect
the products that come out of the column and allow us to quantify the amount of methanol
that was present in the sample.
38 3 Aluminum based ambiphilic molecules fort the capture of carbon
dioxide.
3.1 Advances in the transition-metal mediated reduction of CO2
During the time that this work was being carried out, there was a series of very important
breakthroughs in the field of transition-metal catalyzed reduction of carbon dioxide. Firstly,
Leitner and Klankermayer reported a ruthenium catalyst capable of CO2 hydrogenation to
methanol (Scheme 3-1).125
Scheme 3-1: CO2 hydrogenation by Ru catalyst.
Drawbacks of the system included the necessity of using high pressures (10 atm CO2/ 50 atm
H2) and high temperature (140 °C). Nonetheless, an impressive TON of 221 was observed
with a single catalyst, an order of magnitude over the previous record holding system
involving three metal centers that generated only 21 turnovers. In the following years (20112015), the catalyst was adapted to flow systems126 and multiphase catalysis.127 The catalyst
also proved active in the reductive methylation of imines128 and amines129 through CO2
hydrogenation.
Employing the same pincer framework that was previously reported for the hydroboration of
CO2, but with an iridium metal center instead of nickel, Brookhart and co-workers were able
to catalyse the hydrosilylation of CO2 to methane with impressive turnovers (Figure 3-1b).130
Soon after, Turculet and co-workers reported a related system consisting of a different pincer
ligand framework used in combination with palladium or platinum metal centers (Figure 31 c).131 While the activity is slightly lower than Brookhart’s system, it is much beyond the
values of previously reported systems.
39 Figure 3-1: Catalysts for the reduction of CO2 a) Ni pincer catalyst b) Ir pincer catalyst c)
Pd or Pt pincer catalyst.
Both authors suggest similar mechanistic pathways where the strongly Lewis acidic metal
center abstracts a hydride from the silane, generating a highly reactive silylium ion. These
very oxophilic ions can then effectively coordinate to the oxygen atoms of CO2. Then,
delivery of the metal hydride to the activated carbon atom regenerates the catalyst. This mode
of action is reminiscent of the metal-free reduction of CO2 to methane by Lewis acidic
[AlEt2]+ and silylium ions. 89,90 It can be hypothesized that the higher stability of the metal
centers open access to rapid, long-lived reduction systems. Shortly after these reports, Piers,
Maron and Eisenstein reported a very complete study of strongly Lewis acidic scandium
catalyzed hydrosilylation of CO2 to methane.132 While turnovers remained limited,
mechanistic studies once again confirmed that activation of the silane by the Lewis acid was
a key step in such reduction systems. Unfortunately, these transition metal catalysts do not
allow the isolation of the highly desirable methanol product due to systematic over-reduction
to methane that is invariably associated with the use of these highly Lewis acidic catalysts.
Further progress was also made in the field of transition-metal mediated hydroboration of
CO2. Sabo-Etienne and co-workers reported that a ruthenium polyhydride catalyst could
generate a complex mixture of reduction products when the catalyst was exposed to CO2 in
the presence of pinacolborane (HBPin). Labelling experiments with
13
CO2 allowed the
authors to identify five different reduction products, including a methoxide derivative,
species which can readily be hydrolysed to methanol (Scheme 3-2).133
40 Scheme 3-2: Hydroboration of CO2 by a ruthenium polyhydride catalyst.
While the report of new CO2 reduction products is of fundamental interest, in no case could
the system generate more than 10 turnovers of CO2 reduction. Furthermore, the lack of
selectivity limited the applicability of the system.
Altogether, the development of new and very efficient transition metal based reduction
systems demonstrated that efficient catalytic reduction of CO2 was possible, propelling the
field of CO2 reduction to a point it was not possible to envision just a year earlier.
Nonetheless, in order to justify the use of rare and very expensive transition metals such as
ruthenium, iridium, platinum or palladium in combination with complex ligand frameworks,
much better reduction rates would need to be achieved. As was amply discussed before, the
use of abundant non-metal elements represents an enticing alternative, both from economic
and ecological standpoints.
3.2 Advances in the binding of CO2 by FLPs
Fortunately, the field of FLP chemistry for CO2 binding was also evolving rapidly. Stephan
reported CO2 fixation using the highly Lewis acidic Al(C6F5)3 (AlF) (Figure 3-2a),134 while
Erker reported increasingly complex phosphine systems (Figure 2-2b).135 Also of particular
interest was the binding of CO2 using original Lewis basic centers such as variety of
phosphinimines (Figure 3-2c)136 or bis(dialkylphosphino)amines of the general formula
HN(PR2)2 (Figure 3-2d),137 expanding the scope of available FLP systems. Using a very
original
approach,
Stephan
and
collaborators
also
prepared
an
aryl
bridged
amido/fluorophosphonium FLP system capable of CO2 binding (Figure 3-2e).
41 Figure 3-2: New reported FLP systems for CO2 binding.
A quick look at the previously reported systems reveals that the use of very Lewis acidic
moieties such as BCF or AlF is recurrent in almost every FLP. While this is certainly a valid
strategy to achieve stoichiometric CO2 binding, the design of a catalytically active system
remained elusive.
3.3 Fundamental aspects of CO2 chemistry: the general concept
A critical look at the fundamental thermodynamics of FLP systems allows us to identify the
underlying factors governing this recent chemistry. Classically, a combination of unhindered
Lewis acids and bases will result in a significant stabilization, quenching the reactivity of the
active centers (Figure 3-3a, red). With Frustrated Lewis Pairs on the other hand, the
components are prevented from quenching each other, thus allowing the respective active
fragments to retain their activity and instead, react in a synergic fashion with another
molecule to reach a certain transition state (TS) that would engender a chemical
transformation (Figure 3-3a, blue). Thus, the role of the steric bulk is simply to prevent
thermodynamic stabilization.
Still, most reports continue to utilize highly Lewis basic and highly Lewis acidic fragments
in order to ensure maximal ‘‘activation’’ of CO2. The inherent problem with this approach is
that a strong activation of CO2 will lead to the same thermodynamic well that was being
avoided at first (Figure 3-3b red). Furthermore, the use of very active centers will limit the
system’s tolerance to other functional groups and/or reaction intermediates. For this reason,
we propose that the use of less reactive active centers, even though they could bind CO2 less
strongly, should actually lead in increased reactivity (Figure 3-3b, blue)
42 Figure 3-3: Qualitative thermodynamic analysis of Lewis pair reactivity a) Classical Lewis pair
reactivity b) Frustrated Lewis Pair reactivity.
The terminology that is classically being used in FLP chemistry may at times be somewhat
inaccurate. In fact, most of the systems considering CO2 binding utilize the term ‘‘activation’’
to designate CO2 binding. While this terminology is straightforward, it is important to
differentiate CO2 binding and CO2 activation. It is clear that the binding of CO2 to either a
Lewis base or acid will change the electron distribution of the CO2 molecule. It is most likely
that in some cases, this redistribution leads to different reactivity patterns and as such, the
molecule could be considered activated. This concept is well illustrated by the metal-free
reduction of CO2 by NHCs or by strong Lewis acids.90,138 However, in the case of some FLP
systems, it can be difficult to draw the line between activation and de-activation. The use of
a strong Lewis acid for CO2 binding will most likely generate an important electron
deficiency at the carbon atom, but by filling this reactive orbital with the lone orbital pair of
a Lewis base, the CO2 is being thermodynamically stabilized. The same situation can be seen
the other way around, where the Lewis base interaction increases electron density on the
oxygen atoms, favoring the strong binding of a Lewis acidic moiety.
43 3.4 Overview of the project
While the work presented in chapter 4 was published before the work in this chapter, it is
important to note that the results presented in the present chapter were obtained much before
those that are presented in chapter 4. The chronology of the development of this project is
important in order to fully appreciate the working strategies that ensued. For these reasons,
it was thought best to give an overview of the developments described in chapter two and
three in a chronological order with respect to the evolution of the project. Therefore, the
contents of chapters 2 and 3 are summarized in the next section. For an in-depth discussion,
the reader is directed to the respective research articles that are presented in chapters 2 and
3.
It goes without saying that CO2 binding by FLPs can, in ideal situations, lead to activation of
the small molecule. In fact, the hypotheses that are being put forward in the following chapter
are based on finding the right equilibrium that will promote CO2 reactivity while avoiding
insurmountable thermodynamic stabilization. While aryl bridged ambiphilic molecules were
extensively used in coordination chemistry, none of these versatile systems were studied for
CO2 capture and transformation, motivating us to develop a new aryl bridged FLP for such
applications.139
Our attempt to generate the monophosphine analogue ortho-Ph2PC6H4AlMe2 was dampened
by the observation that such species undergo rearrangement under synthetic conditions to
generate the triphosphine species Al(C6H4(o-PPh2))3 (1). This unusual decomposition
product was not the initial FLP we were aiming for, but it presented very interesting features.
For one, it was still an ambiphilic molecule that should be able to activate carbon dioxide.
But, the most interesting property was the presence of the two sterically accessible pendant
Lewis basic moieties within the secondary coordination sphere of the aluminum center.
44 The initial hypothesis was that by simultaneously activating CO2 and a reducing agent, it
would be possible to promote catalytic CO2 reduction without the use of transition metals.
Due to the availability of the lone pair of the pendant Lewis base, a sterically unhindered
hydroborane such as catecholborane seemed like a substrate of choice for the reaction
(Scheme 3-3).
Scheme 3-3: Proposed dual activation strategy for the catalytic hydroboration of CO2.
While the new system did exhibit important catalytic activity for CO2 reduction, we soon
realized that the aluminum compound decomposed in the presence of hydroboranes to form
a simpler ambpihilic phosphine-borane species 2. We hypothesized that this new species
could have been responsible for the catalytic activity of the system, but interestingly enough,
it did not show any sign of reactivity towards either CO2 or catecholborane. Still, we decided
to push the investigation further and combine species 1, catecholborane and CO2.
Gratifyingly, 2 proved to be an incredibly active catalyst for the hydroboration of CO2 to
methoxyboranes (Scheme 3-4).
Scheme 3-4: Decomposition of 1 into 2.
In order to remain succinct, the phosphine-borane was first reported as an active catalyst,
while the decomposition of the phosphine-alane system was reported later in an invited
45 contribution. The following chapter deals with the development and reactivity study of the
phosphine-alane system, while the development of the phosphine-borane catalytic system is
reported later in chapter 3. Since chapters 2 and 3 are inherently linked together as the work
was taking place somewhat simultaneously, the conclusions and perspectives section will be
reported at the end of chapter 3. Similarly, no introduction or literature update will be given
for chapter 3, as to ensure a continuous flow of ideas.
46 3.5 Research article: A Tris(triphenylphosphine)aluminum Ambiphilic
Precatalyst for the Reduction of Carbon Dioxide with Catecholborane
3.5.1
Résumé
L’espèce ambiphile Al(C6H4(o-PPh2))3 (1) a été synthétisée et complètement caractérisée,
entre autres par diffraction des rayons-X. L’espèce 1 adopte une géométrie pseudobipyramide trigonale causée par deux interactions Al-P. La molécule peut réagir avec le CO2
pour générer un adduit observé de façon courante dans la chimie d’activation du CO2 par les
PLFs. Cette espèce ambiphile sert de précatalyseur pour la réduction du CO2 en présence de
catécholborane (HBcat) pour générer CH3OBcat, qui à son tour, peut facilement être
hydrolysé en méthanol. Une analyse du mélange réactionnel confirme qu’en présence de
HBcat, l’espèce 1 génère le catalyseur connu 1-Bcat-2-PPh2-C6H4 (2) ainsi qu’une espèce
d’aluminium insoluble. En effet, il a été possible d’isoler un monocristal de Al(κ2O,O(MeO)2Bcat)3 (5), supportant cette hypothèse. De plus, l’analogue protégé de 1, Al(C6H4(oPPh2.BH3))3 (4), ne réagit pas en présence de catécholborane, suggérant que les phosphines
jouent un rôle important dans la transformation de 1 à 2.
3.5.2
Abstract
Ambiphilic species Al(C6H4(o-PPh2))3 (1) was synthesized and fully characterized, notably
using X-ray diffraction. Species 1 exhibits pseudo-bipyramid trigonal geometry caused by
the two Al-P interactions. 1 reacts with CO2 to generate a CO2 adduct commonly observed
in the activation of CO2 using Frustrated Lewis Pairs (FLPs). This ambiphilic species serves
as a precatalyst for the reduction of CO2 in presence of catecholborane (HBcat) to generate
CH3OBcat, which can be readily hydrolyzed in methanol. The reaction mixture confirms that
in presence of HBcat,1 generates known CO2 reduction catalyst 1-Bcat-2-PPh2-C6H4 (2) and
intractable catecholate aluminum species. It was however possible to isolate a single-crystal
of Al(κ2O,O-(MeO)2Bcat)3 (5) supporting this hypothesis. Also, the borane protected analogue
of 1, Al(C6H4(o-PPh2.BH3))3 (4),
does not react with catecholborane, suggesting the
influence of the pending phosphines in the transformation of 1 into 2.
47 3.5.3
Introduction
Ambiphilic compounds containing both Lewis acid and Lewis base moieties in the
same molecular framework, most notably containing group XIII and XV elements, have
generated growing interest in the past decade for a myriad of applications. These molecules,
including the well-known “Frustrated Lewis Pairs (FLPs)”,99 have been used as multi-center
catalysts,140–143 as precursors for nonlinear optical materials and sensors,144–155 for activating
small molecules,98,103,117,156–160 and as ligands for transition metals.111,114,115,121,122,161–177 A
particularly interesting design of ambiphilic molecule is the phosphine borane derivatives
that have been developed over the past few years mostly by Bourissou et al.178–195 These
mono-, bis-, and tris-phosphine species demonstrated versatile coordination modes to
transition metal centers, notably as Z-type ligands for transition metals (Figure 3-4).112 More
recently, monophosphine boranes have shown impressive reactivity in singlet dioxygen
activation,190 in the catalytic Michael addition reaction,191 and in trapping reactive
intermediates of organic transformations.194 Whereas borane ambiphilic molecules have been
extensively studied over the past decade, the reactivity of aluminum-containing ambiphilic
molecules has not garnered much attention, notably because of the synthetic challenges
associated with their synthesis and their kinetic instability.111,117,121,122,156,196–198 Nevertheless,
such compounds have demonstrated interesting potential in the activation of small molecules
and in catalysis.
Figure 3-4: Structures of the previously reported ambiphilic ligands.
48 One area where ambiphilic molecules could have a significant impact is in the
activation and functionalization of carbon dioxide. The latter molecule possesses both an
electrophilic carbon atom and nucleophilic oxygen atoms, thereby acting as an ambiphilic
substrate. In a seminal report, Stephan and Erker have demonstrated that ambiphilic FLPs
are efficient in the activation of carbon dioxide.103 Since then, a large variety of ambiphilic
and FLP systems have been shown to be active in carbon dioxide capture.156103,117 However,
there are only a handful of ambiphilic systems able to functionalize carbon dioxide into value
added chemicals. It has been shown that the FLP system consisting of PMes3/AlX3, (Mes =
mesityl, X=Cl, Br) can reduce CO2 to methanol using BH3•NH3 as hydrogen source.106 It was
also shown that that CO2 could be hydrogenated using TMP/B(C6F5)3 (TMP = 2,2,6,6tetramethylpiperidine).63 However, these two systems have limited use since they require
stoichiometric amounts of FLPs. Using TMP and B(C6F5)3, Piers has demonstrated the
catalytic reduction of carbon dioxide using Et3SiH, but with low turnovers.104 In a recent
breakthrough, our research group demonstrated that molecule 1-Bcat-2-PPh2-C6H4 (1) (cat =
catechol), with modest Lewis acid and basic ambiphilicity, acts as a catalyst for CO2
reduction in the presence of hydroboranes to generate methoxyboranes which can be readily
hydrolyzed into methanol.199 In order to broaden the scope of such systems, we were curious
to see the effect of using an aluminum center enclosed in a tris-arylphosphine scaffold.
Herein, we report the preparation of a novel tri-phosphine organoalane Al(C6H4(o-PPh2))3
(1) which can reversibly bind carbon dioxide under ambient conditions. Species 1 can be
used as a precatalyst for the reduction of CO2 in presence of cathecolborane (HBcat),
generating catalyst 2 and aluminum catecholate species, including Al(κ2O,O-(MeO)2Bcat)3 (5)
which was structurally characterized.
49 3.5.4
Results and discussion
By using the reaction pathway illustrated in Scheme 3-5, the ambiphilic species Al(C6H4(oPPh2))3 (1) was synthesized in 76% yield from previously reported o-lithiated
triphenylphosphine.199 Colorless crystals of 1 suitable for X-ray diffraction studies were
obtained from toluene at -40 °C.
Scheme 3-5: Synthesis of Al(C6H4(o-PPh2))3 (1).
According to the results of single-crystal X-ray diffraction analysis, 1 crystallized in the P-1
space group with two crystallographic independent molecules in the asymmetric unit (one of
the two independent molecules is shown in Figure 3-5). 1 adopts a distorted trigonal
bipyramid geometry in the solid state with two Al-C-C-P four-membered rings. The trigonal
bipyramid of AlC3P2 in the first molecule (Al1) is more distorted than that of the second one
(Al2). The Al-C bond lengths in the first molecule range from 1.976(2) to 2.005(3) Å, which
are comparable to those observed in similar structures, and the two Al-P bond lengths are
significantly different from each other (2.630(1) and 2.820(1) Å). The P1-Al1-P3 angle of
155.56(3) ° is far from the ideal value of 180 °. The distance between Al1 and the third
phosphorous atom (3.408(1) Å) is too long to exhibit significant bonding interaction. The
most important difference between the two independent molecules is that the longer of the
two bonding Al-P interactions is significantly shorter in the second molecule (2.726(1) Å)
than in the first one (2.820(1) Å). As a consequence, the coordination environment around
Al2 is more crowded than in the first molecule which leads to the third phosphine moiety
being pushed away from the metal centre (Al2-P5 distance of 3.438(1) Å). The other bond
lengths in the second molecule are quite similar to the first one.
50 Figure 3-5: ORTEP drawing of the first molecule of 1 in the asymmetric unit cell.*
*Anisotropic atomic displacement ellipsoids shown at the 50% probability level. Hydrogen atoms are
omitted for clarity. Selected bond lengths [Å] and angles [°]: Al1-P1 = 2.8197(9); Al1-P3 = 2.6305(9);
Al1-C19 = 1.976(2); Al1-C18 = 1.992(2); Al1-C37 = 2.005(2) Å; C19-Al1-C18 = 128.11(1); C19Al1-C37 = 113.71(9); C18-Al1-C37 = 117.52(9); C19-Al1-P3 = 102.05(7); C18-Al1-P3 = 104.27(7);
C37-Al1-P3 = 67.52(7); C19-Al1-P1 = 101.61(7); C18-Al1-P1 = 64.82(6); C37-Al1-P1 = 97.20(7);
P3-Al1-P1 = 155.56(3).
51 Müller and colleagues published one similar structure, Al{C6H4[o-CH2P(Ph)2]}3, in which
the trigonal bipyramid of AlC3P2 was less distorted, with a P1-Al1-P2 bond angle of
164.78(8)o.123 Such discrepancy can be attributed to the flexibility of the five-membered rings
around the central aluminum atom in Al{C6H4[o-CH2P(Ph)2]}3 compared to the fourmembered rings in 1. The P1-Al and P2-Al bond distances of respectively 2.676(3) Å and
2.782(2) Å in the complex reported by Müller are quite comparable to that of 1, but the P3Al distance (4.440(6) Å) is much longer. Similar bond distances of 2.66 and 2.78 Å have also
been reported by Bourissou.183 Only a handful of structures with two or three four-membered
rings containing aluminum and non-metal elements have been reported.200–208
The 31P{1H} NMR spectrum of 1 in a benzene-d6 solution shows one sharp singlet at -1.5
ppm indicating a fast exchange process between all the phosphines. In order to assess this
fluxional process, variable temperature NMR experiments were performed using a 0.07 M
solution of 1 in toluene-d8 down to -100 ° C. It was observed that the single peak for 1 started
to broaden at -90 °C, indicative of the slowing of the exchange process. However, it was not
possible to go to lower temperature because of the freezing temperature of toluene.
Nevertheless, it is possible to compare this data with the report by Müller on Al{C6H4[oCH2P(Ph)2]}3.123 Indeed, in the latter complex, the broadening appears at -80 ° C suggesting
that the tension in the four-membered rings in 1 induce a more dynamic behaviour than in
the species with five-membered rings. In order to explore the bonding feature in 1, DFT
calculations at the B3PW91, 6-31G** level of theory were performed. As shown in Figure
3-6, the shortest P-Al bond is of 2.73 Å in the optimized structure, while the distance between
the trans P atom and Al is slightly longer (2.82 Å). The non-interacting P atom is 3.47 Å
away from the aluminum centre. These values are strikingly similar to those found in the
solid state and reflect the accuracy of the model. According to the NBO calculations, the
bond orders were of 0.3293, 0.2536, and 0.0639 for the shortest to the longest Al-P bond
lengths, respectively. It indicates that the two shorter Al-P distances have partial bonding,
whereas no significant bond is observed with the last phosphine.
52 Figure 3-6: Optimized structure of 2 using DFT calculations.*
*Level of theory: B3PW91, 6-31G**. Selected bond lengths [Å] and angles [°]: Al1-P1 = 2.85; Al1P3 = 2.69; Al1-C19 = 2.00; Al1-C18 = 2.00; Al1-C37 = 2.01 Å; C19-Al1-C18 = 123.6; C19-Al1C37 = 114.9; C18-Al1-C37 = 120.6; C19-Al1-P3 = 104.0; C18-Al1-P3 = 105.5; C37-Al1-P3 = 67.0;
C19-Al1-P1 = 101.0; C18-Al1-P1 = 64.6; C37-Al1-P1 = 96.7; P3-Al1-P1 = 154.2.
This interesting fluxional behavior prompted us to study the reactivity with carbon dioxide
since the phosphine centers could still be available for CO2 coordination. When a 0.03 M
solution of 1 in benzene-d6 was exposed to 1 atmosphere of CO2 at room temperature, a
reaction took place immediately leading to two new signals in the 31P{1H} NMR spectrum,
a doublet at -5.1 ppm and a triplet at -6.2 ppm integrating for two and one phosphorous atoms,
respectively, both with a small JP-P of 3.2 Hz. To our knowledge, this species represents the
first report of an aryl bridged o-aryl ambiphilic CO2 adduct.
53 Scheme 3-6: Generation of 3 upon exposure of 1 to CO2.
In order to find out if one or two CO2 molecules could coordinate to 1, the synthesis of 3 was
carried out in presence of
13
CO2, generating 3-13C. The resulting
31
P{1H} NMR spectrum
showed splitting of the triplet signal to a doublet of triplets with 1JC-P= 102 Hz while the
doublet remained unchanged, thus suggesting that only one CO2 molecule can interact with
the aluminum species (Scheme-3-6). The 13C{1H} NMR spectrum exhibited a new signal at
159.5 ppm, reminiscent of activated CO2, also with a 1JP-C of 102 Hz, confirming that the new
product arises from a single CO2 coordination.120 However, as no additional coupling with
the other two phosphorous atoms were observed for the 13C atom, we were curious to see
what caused the additional 3.2 Hz JP-P coupling in 3. Since the isolation at the solid state of 3
was not possible, the structure of the CO2 adduct was determined using DFT. The enthalpy
profile determined by DFT calculations showed that the formation of 3 is favorable by 4.1
kcal/mol with respect to free 1 and CO2, which is in accordance with the experimental results
(Figure 3-7). The NBO analysis did not show any significant bonding interaction to be
present between the phosphine centres and the aluminum core (values of 0.0372, 0.0131, and
0.0874). Therefore, the origin of the coupling is not arising from 3JP-P or 5JP-P scalar couplings
through Al-P interactions, but presumably from a 6JP-P scalar interaction through the
P-C=C-Al-C=C-P backbone, although through-space interactions cannot be excluded.i
i
The possibility of long range coupling was considered and cannot be excluded, but several observations go
against it. First, the low symmetry of 3 in the optimized structure (P1-P2 and P2-P3 bond lengths of 4.39 and
5.03 Å, respectively) would suggest a different coupling constant between the P atoms. Also, long-range
coupling would be possible if electronic communication between the nuclei is present, but we did not observe
any orbital suggesting that the three phosphorous atoms are able to communicate through space.
54 Figure 3-7: Optimized structure of 2 using DFT calculations.*
*Level of theory: B3PW91, 6-31G** Selected bond lengths [Å] and angles [°]: Al1-P1 = 3.48; Al1P3 = 3.64; Al1-C19 = 2.01; Al1-C18 = 2.00; Al1-C37 = 2.00; Al1-O1 = 1.86; P2-C0 = 1.94; C0-O1
= 1.27; C0-O2 = 1.21 Å; C19-Al1-C18 = 107.7; C19-Al1-C37 = 122.9; C18-Al1-C37 = 109.6; O1C0-O2 = 130.9; O1-C0-P2 = 116.3; O2-C0-P2 = 112.6; O1-Al1-C18 = 111.6; O1-Al1-C19 = 103.2;
O1-Al1-C37 = 101.4.
55 After removing the volatiles from the J-Young NMR tube under vacuum for 1 hour and
dissolving in ca. 0.6 mL of benzene-d6, multinuclear NMR spectroscopy revealed complete
conversion of 3 back to 1. Alternately, leaving a solution of 3 under nitrogen for 12 hours
resulted in a mixture of 1 and 3 in approximately 5:1 ratio (according to the values of NMR
integration). These results demonstrate that the CO2 adduct is reversible, and suggest a
fluxional behavior between the CO2 adduct and free 3, a well-known behavior in the
ambiphilic activation of carbon dioxide.103 More interestingly, none of the previously
reported aluminum based ambiphilic molecules are known to give-up CO2 at room
temperature in solution.156,117,209 Leaving the solution under nitrogen for another 12 hours
resulted in complete conversion back to 1.
Because the calculations indicate that this CO2 adduct is much more favorable than in the
case of the previously reported organocatalyst 1-Bcat-2-PPh2-C6H4 (1), (-4.1 kcal.mol-1 vs
+9.9 kcal.mol-1), we were curious to see if catalytic reduction of carbon dioxide could be
achieved through this novel CO2 adduct. The
31
P{1H} and 1H NMR monitoring of the
reaction between a 5 mM solution of 1 containing 49 equivalents of HBcat under one
atmosphere of CO2 at 60°C did show the presence of catalytic activity. In the first 30 minutes
after the addition of CO2, the 1H NMR spectrum did not show any significant change.
However, after 30 minutes, the 1H NMR did show a signal corresponding to CH3OBcat at
3.37 ppm. As observed in Figure 3-8, the concentration of CH3OBcat increased rapidly once
the induction period was over and then started slowing down with diminishing concentration
of HBcat. The addition of 1000 equiv of HBcat to a 1.5 mM solution of 1 in benzene-d6 under
2 atm. of CO2 yielded quantitative formation of CH3OBcat over a 72 hour period at 70° C.
56 Turnover Number (TON)
35
30
25
20
15
10
5
0
0
50
100
150
200
250
Time (minutes)
Figure 3-8: Turnover number (TON) for the formation of CH3OBcat.*
*Results from a 5 mM solution of 1 in benzene-d6 in the presence of 49 equiv. of HBcat under 2 atm
of CO2 at 60 °C. The TONs are based on the number of hydrogen atoms transferred to CO2.
In order to assess possible rearrangements of the precatalyst during catalysis, the
reaction of 1 with 4 equiv of catecholborane at room temperature was monitored using
multinuclear NMR spectroscopy. A complex set of signals in the
31
P NMR spectrum
suggested that several rearrangements were occurring, but none of the signals corresponded
to 1. However, heating the solution at 70 °C for 3 hours did reduce the number of observable
species with a predominant signal at -4.6 ppm reminiscent of species 2. The aromatic signals
observed in the 1H NMR of the reaction mixture were also consistent with the formation of
organocatalyst 1-Bcat-2-PPh2-C6H4 (2). Most notable was the apparition of a downfield
doublet at 8.08 ppm that is assigned to the resonance of the proton ortho to the borane moiety
of 2. However, a large amount of refractory white precipitate appeared in the solution, which
could be attributed to the presence of catecholate aluminum species. Species 2 was
successfully isolated in measurable amounts (40 % yield) by reacting 1 with 4 equivalents of
HBcat in toluene at 70 °C for 3 hours. The crystallization from the reaction mixture afforded
quality crystals of 2.
57 Although the isolation of the aluminum-containing products formed under large excess of
HBcat was not possible, it is possible to speculate that the degradation of HBcat in the
presence of the Lewis basic moieties of 1 can be at the origin of the formation of 2. Indeed,
Westcott et al. demonstrated that in the presence of phosphine moieties, HBcat can degrade
into
several
boron-containing
products,
including
notably
B2cat3,
BH3
and
[(PR3)2BH2][Bcat2].210 Therefore, one possible pathway would be for nucleophilic boron
containing species such as the [Bcat2] anion to attack the oxophilic aluminum centre, thus
generating a highly nucleophilic phenoxide anion, that would in turn attack the boron of
another cathecolborane to form species 2. In order to support such assumption, the synthesis
of the BH3 protected adduct of 1 was carried out. Reacting 1 with an excess of BH3.SMe2
followed by removal of the volatiles in vacuo generated derivative 4, as demonstrated by the
single broad peak at 22.5 ppm in the
11
31
P{1H} NMR spectrum (See Scheme 3-7).ii
B{1H}NMR also showed a broad signal at -33 ppm, confirming the interaction with
phosphorous. Interestingly, the addition of 30 equiv of HBcat to a solution of 4 in benzened6 did not yield any new compound nor catalytic activity when exposed to CO2 even after
heating at 70 °C for 12 hours.
Scheme 3-7: Proposed protection of the phosphine moieties of 1.
ii
The addition of BH3 to PPh3 yields the BH3.PPh3 adduct having the following spectral properties: the
B{1H}NMR showed a broad peak at -37 ppm. The 31P{1H} showed a broad peak at -21 ppm. 11
58 It was also possible to observe that 4 was also generated when six equivalents of MeOBcat
were added to 1. As it was observed in the presence of HBcat, the only other species observed
were refractory materials. Fortunately, it was possible to isolate a single crystal that was
identified as species Al(κ2O,O-(MeO)2Bcat)3 (5) (Scheme 3-8). Unfortunately, all attempts to
purify and isolate this compound in a significant quantity failed. However, the presence of
such species in solution supports our hypothesis that rearrangements caused by nucleophilic
anionic borate species can be at the origin of the generation of species 2 and consequently of
the catalytic activity. The ORTEP representation of 5 is shown in Figure 3-9.
Scheme 3-8: Synthesis of species 5.
According to the result of the X-ray diffraction analysis, the structure of 5 is refined
to the P-1 space group. Two crystallographic independent molecules and one solvent
molecule of toluene, disordered in two different orientations in a 50:50 ratio, are present in
the asymmetry unit. As shown in Figure 3-9, the central Al atom in each molecule of 5 is
coordinated by six o-CH3 groups forming a slightly distorted octahedron. In complex 5, the
Al-O bond lengths range from 1.859(1) to 1.884(1) Å in the first molecule, and from 1.862(1)
to 1.874(1) Å in the second molecule, distances that are comparable with those in the closest
structures reported, which consist of the aluminum species [Al(κ2O,O--OR)2Al(OR)2)3] (R =
alkyl).211 Because of the presence of the four-membered rings, the octahedron is distorted
with O-Al-O angles ranging from 158.76(6)° to 161.82(6)°. While the pseudo-D3 geometry
is commonly found in aluminum chemistry, aluminum containing products having three fourmembered rings with non-metal elements is quite rare.
59 Figure 3-9: ORTEP drawing of one independent molecule of 5.
*Anisotropic atomic displacement ellipsoids are shown at the 50% probability level. The hydrogen
atoms are drawn using an arbitrary sphere size for clarity. The hydrogen atoms are drawn using an
arbitrary sphere size for clarity. Selected bond lengths [Å] and angles [°] : Al1-O1 = 1.8593(1); Al1O3 = 1.8606(1); Al1-O4 = 1.8691(1); Al1-O5 = 1.8737(1); Al1-O6 = 1.8747(1); Al1-O2 = 1.8842(1);
O1-Al1-O3 = 94.66(6); O1-Al1-O4 = 159.10(7); O3-Al1-O4 = 70.65(6); O1-Al1-O5 = 98.34(6); O3Al1-O5 = 96.92(6); O4-Al1-O5 = 98.16(7); O1-Al1-O6 = 98.84(6); O3-Al1-O6 = 162.51(7); O4Al1-O6 = 98.77(6); O5-Al1-O6 = 70.22(6); O1-Al1-O2 = 70.38(6); O3-Al1-O2 = 100.59(6); O4Al1-O2 = 97.08(6); O5-Al1-O2 = 159.80(6); O6-Al1-O2 = 94.42(6).
60 Generation of 5 provides significant thermodynamic stability to the system by generating
strong Al-O and B-O bonds. This reaction is a good example of the relative kinetic instability
of arylaluminum species in the presence of oxygen sources. Although such aluminum
compounds are efficient in the activation of carbon dioxide, it is necessary to account for
their reactivity with potential reduction products in order to avoid catalyst decomposition. As
such, aryl bridged alanes are probably not suitable catalysts for the carbon dioxide reduction
into methoxyboranes using hydroboranes.
3.5.5
Conclusion
A novel triphosphine alane ambiphilic molecule with interesting coordination geometry has
been synthesized and characterized by X-ray diffraction. Its fluxional behavior in solution
indicates a rapid process where phosphine moieties dissociate from the metallic center,
leaving the aluminum open for further coordination. Remarkably, 1 is able to bind CO2 at
room temperature in a reversible way to form a new CO2 coordination complex.
Decomposition of 1 into an interesting hexavalent aluminum species and reported active
catalyst 2 by reaction with HBcat and CO2 reduction product CH3OBcat suggests that even
though
phosphine-alane
compounds
seem
like
promising
alternatives
for
the
activation/reduction of carbon dioxide, their sensitivity limits the scope of their application
in catalysis. In fact, reduction with borane reagents will ultimately lead to decomposition of
the catalyst. Studies to stabilize the system and broaden the scope of reducing agents to
alkylsilanes are currently underway.
61 3.6 Experimental Section
3.6.1
General procedure
Unless otherwise specified, manipulations were carried out under an atmosphere of
dinitrogen, using standard glovebox and Schlenk techniques. Reactions were carried either
in a sealed J-Young NMR tube, in which case NMR conversions are indicated, or in standard
flame dried Schlenk glassware. All solvents were distilled from Na/benzophenone, benzened6 and toluene-d8 were purified by vacuum distillation from Na/K alloy. Toluene was stored
on Na/K alloy. Bone dry CO2 was purchased from Praxair and used as received. 13CO2 (99%
isotope label) was purchased from Cambridge Isotope Laboratories and also used as received.
Lithiated o-bromophenyldiphenylphosphine212 and CH3OBcat71 were synthesized in good
yields by following literature procedures. The full characterization of species 2 was already
reported elsewhere.199
NMR spectra were recorded on a Agilent Technologies NMR spectrometer at 500 MHz (1H),
125.758 MHz (13C), 202.456 MHz (31P) 160.46 MHz (11B), on a Varian Inova NMR AS400
spectrometer, at 400.0 MHz (1H), 100.580 MHz (13C), 161.923 MHz (31P), or on a Bruker
NMR AC-300 at 300MHz (1H), 75.435 MHz (13C), 121.442 MHz (31P). 1H NMR and 13C
{1H} NMR chemical shifts are referenced to residual protons or carbons in deuterated
solvent. The temperatures of the VT NMR experiments were measured using a thermocouple
inside the probe, which was calibrated with methanol prior to use. Multiplicities are reported
as singlet (s), broad singlet (s, br) doublet (d), triplet (t), multiplet (m), or virtual triplets (vt).
Chemical shifts are reported in ppm. Coupling constants are reported in Hz. gHMQC,
gHSQC, NOESY 2D, COSY and 1H{31P} NMR experiments were performed in order to
properly assign spectra.
62 3.6.2
Synthesis of compounds
Al(C6H4(o-PPh2))3 (1) : 2-(Diphenylphosphino) phenyllithium diethyletherate (1.035g, 3.00
mmol) was placed in a Schlenk vessel with 0.134g of AlCl3 (0.134g, 1.00 mmol). Toluene
(c.a. 20ml) was then added directly on the solid reactants resulting in a sudden rise of
temperature indicating an exothermic reaction. The mixture was then heated at 50°C for 12
hours. The solution was collected by cannula filtration and the filtrate was concentrated in
vacuo yielding 0.64 g of 1 as a colorless solid (76% yield). The material was further purified
by dissolving in toluene, filtering and storing at -40 °C for 3 days after which the product
precipitated as colorless crystals suitable for X-ray diffraction (0.33g, 39% yield).
1
H NMR (500 MHz, benzene-d6): δ 8.11 (d, 3H, 3JH-H = 7.2 Hz), 7.36-7.51 (bm, 1H), 7.30-
7.33 (m, 3H), 7.26 (m, 12H), 7.17-7.18 (m, 2H), 7.06-7.09 (m, 3H), 6.88-6.95 (m, 18H).
13
C{1H} (126MHz, benzene-d6): 144.6 (d, 3JP-C = 2.4 Hz), 144.6 (dd, 2JP-C = 18.6Hz, 3JP-C =
1.0 Hz), 144.6 (s), 138.9 (ddd, 2JP-C = 21.6 Hz, 2JP-C = 12.4 Hz, 3JP-C = 2.4 Hz), 136.7 (s),
134.2 (d, 2JP-C = 19.6 Hz), 133.6 (dd, 2JP-C = 10.4 Hz, 2JP-C = 5.2 Hz), 132.0 (s), 129.3 (s),
128.9 (s), 128.8 (d, 2JP-C = 6.7 Hz), 128.7 (bs, width = 8.5 Hz), 128.6 (s), 128.5 (dd, 3JP-C =
4.8 Hz, 3JP-C = 2.4 Hz), 128.4 (bs, width = 6.6 Hz).31P{1H} (203 MHz, benzene-d6): δ -1.5.
Al(C6H4(o-PPh2))3(CO2) (3): A solution of 1 in benzene-d6 (2.8 mg in c.a. 0.6ml, 5.7
mmol/L) was placed in a J-young NMR tube. This tube was frozen in liquid nitrogen on the
schlenk line and put under vacuum for 30 minutes. An atmosphere of CO2 was then allowed
to flow in the tube for 5 to 10 seconds, and the tube was closed. The liquid nitrogen bath was
then gradually removed to let the solution slowly thaw. 96% yield based on NMR integration.
1
H NMR (500 MHz, benzene-d6,): δ 8.48 (bs, 1H), 8.01 (d, 2H, 3JH-H= 7.0 Hz), 7.41-7.32 (m,
12H), 7.26 (bs, 1H) 7.11-6.92 (m, 18H), 6.90-6.85 (m, 2H), 6.80-6.62 (m, 6H).
13
C {1H}
(126MHz, benzene-d6) δ145.3 (s), 141.7 (dt, 2JP-C = 19.7 Hz, 3JP-C = 7.7Hz), 140.7 (d, 2JP-C
= 13.1 Hz), 139.9 (d, 2JP-C = 12.2 Hz), 139.7 (d, 2JP-C = 25.9 Hz), 134.4 (d, 2JP-C = 9.9 Hz),
134.0 (d, 2JP-C = 17.5 Hz), 133.8 (s), 133.7 (s), 133.6 (d, 3JP-C = 2.7 Hz), 132.6 (d, 2JP-C = 13.3
Hz), 131.1 (d, 3JP-C = 3.8 Hz), 129.4 (s), 127.8 (d, 2JP-C = 11.0 Hz), 127.0 (d, 2JP-C = 12.6 Hz),
120.9 (1JP-C = 78 Hz).31P{1H} (203 MHz, benzene-d6): -5.1 (d, 2P, 6JP-P=3.2 Hz); -6.2 (t, 1P,
6
JP-P=3.2 Hz).
63 Al(C6H4(o-PPh2.BH3))3 (4) : 110 mg (110mg, 0.13mmol) of 1 were introduced into a small
schlenk. Toluene was then directly added to this product and the resulting mixture was stirred
until a clear solution was obtained. An excess of BH3.SMe2 (0.1 ml, 1.1 mmol, 8.1 equiv)
was added to the solution and the reaction was left stirring for an hour. The solvent and excess
reagents were then removed in vacuo. The product was obtained without further purification
as a very static white powder. Reaction is quantitative. 1H NMR (500 MHz, benzene-d6): δ
8.07 (bs, 3H), 7.62 (t, 3H, 3JP-H = 8.6 Hz), 7.53 (t, 9H, 3JH-H = 8.6 Hz), 7.06 (t, 3H, 3JH-H = 8.2
Hz), 7.02-6.89 (m, 18H), 6.77 (bs, 6H, width = 40.7 Hz), 2.15 (bs, 9H, width = 380.7 Hz).
13
C{1H} (126MHz, benzene-d6) δ 141.4 (d, 2JP-C = 18.6 Hz), 134.0 (d, 2JP-C = 9.5 Hz), 134.0
(d, 1JP-C = 65.3Hz), 133.6 (d, 2JP-C = 9.5 Hz), 133.6 (s), 131.1 (d, 3JP-C = 2.4 Hz), 130.7 (d,
1
JP-C = 57.7 Hz), 130.7 (d, 3JP-C = 2.4 Hz), 129.3 (d, 3JP-C = 2.4 Hz), 128.9 (d, 2JP-C = 10.5
Hz), 128.6 (d, 2JP-C = 10.0 Hz), 128.3 (s), 126.6 (d, 2JP-C = 9.1 Hz). 11B (161 MHz, benzened6) δ -33.0 (s). 31P (202 MHz, benzene-d6) δ 22.3 (s).
[Al(κ2O,O--OMe)2Bcat)3] (5): A solution of CH3OBcat in toluene was added to a solution
of 1 in toluene via cannula. A white precipitate was instantly formed. This mixture was left
stirring under a flow of nitrogen for an hour at room temperature. The reaction was filtered
and washed thoroughly with toluene. Few crystals of 5 were isolated, enough for one
crystallographic study.
3.6.3
CO2 reduction catalytic tests
Al(C6H4(o-PPh2))3 (1) (2.0 mg 2.5 X 10-3 mmol) was dissolved in ca. 0.6 mL of benzene-d6.
HBcat (14.7 mg, 0.12 mmol, 49 equiv) was added to the solution and the mixture was
introduced in a J-young NMR tube. The J-young NMR tube was frozen in a liquid nitrogen
bath after which the headspace of the J-young was filled with 1atm. of CO2. The reaction was
then followed by NMR spectroscopy. Yields are reported by 1H NMR integration using
hexamethylbenzene (1.35 mg, 8.3 X 10-3 mmol) as an internal standard.
64 3.6.4
Crystallographic studies
Nice single crystals with suitable size for both compounds (1 and 5) were mounted on
CryoLoops with Paratone-N and optically aligned on a Bruker SMART APEX-II X-ray
diffractometer with 1K CCD detector with the aid of a digital camera. Initial intensity
measurements were performed using a fine-focused sealed tube, graphite-monochromated,
X-ray source (Mo Kα, λ = 0.71073 Å) at 50 kV and 30 mA. Standard APEX-II213 software
package was used for determining the unit cells, generating the data collection strategy, and
controlling data collection. SAINT214 was used for data integration including Lorentz and
polarization corrections. Semi-empirical absorption corrections were applied using SCALE
(SADABS215). The structures of all compounds were solved by direct methods and refined
by full-matrix least-squares methods with SHELX-97216 in the SHELXTL6.14 package. As
the solvent molecules in 1 are highly disordered, the SQUEEZE subroutine in PLATON217
software suit was used to remove the scattering contributions from the highly disordered
guest molecules. The resulting new HKL file was adopted to further refine the structural
model. All of the H atoms (on C atoms) were generated geometrically and refined in riding
mode. Crystallographic information for all obtained phases is summarized in Table 3-1.
Atomic coordinates and additional structural information are provided in the CIF file of the
Supporting Information.
65 Table 3-1: Crystal data and structural refinements for compounds 1 and 5.
Compound
1
5
Formula
C54H42AlP3
C25.75H31.75O16B3Al
Mr (g mol−1)
810.77
592.67
Crystal size
0.20 × 0.16 × 0.06
0.42 × 0.40 × 0.38
Crystal system
Triclinic
Triclinic
Space group
P -1
P -1
T (K)
150(2)
150(2)
a (Å)
11.0020(2)
12.49960(1)
b (Å)
13.1765(2)
13.9779(2)
c (Å)
31.9414(6)
16.8699(2)
α (o)
87.5550(1)
95.1040(1)
β (o)
82.8900 (1)
96.6310(1)
γ (o)
77.3160 (1)
90.8690(1)
V (Å3)
4482.16(1)
2915.19(6)
Z
4
1
ρcal. (g cm−3)
1.201
1.350
F(000)
1696
1241
μ (mm−1)
0.188
0.131
θ for data collection (o)
1.9 to 26.4
1.5 to 26.4
Reflection collected
18131
46769
Data /parameters
18131 / 1045
11901 / 770
GOOF
1.049
1.038
R1(wR2) (I > 2σ(I))
0.0526 (0.1277)
0.0431 (0.1049)
R1(wR2) (all data)
0.0878 (0.1402)
0.0615 (0.1151)
largest diff. peak / hole (eÅ-3)
0.347 / -0.473
0.384 / -0.303
66 3.6.5
Computational details
Calculations were performed with the GAUSSIAN 03 suite of programs.218 The B3PW91219
functional was used in combination with the 6-31+G** basis set for B,C,H, and O
atoms,220,221 and the SDD basis set with an additional polarization function (one d function
with a 0.34 exponent and a 1.0 contraction coefficient) for P atoms,222 and for aluminum.223
The stationary points were characterized as minima by full vibration frequencies calculations
(no imaginary frequency). All geometry optimization were carried out without any symmetry
constraints.
Cartesian coordinates and free energies as well as additional experimental data including
NMR spectra are available free of charge via the ACS website: http://pubs.acs.org/
67 4 Metal-Free catalytic Reduction of CO2 to methanol
4.1 Research article: A Highly Active Phosphine-Borane Organocatalyst for
the Reduction of CO2 to Methanol using Hydroboranes.
4.1.1
Résumé
Dans cet ouvrage, nous rapportons que l’organocatalyseur 1-Bcat-2-PPh2-C6H4 ((1); cat =
catéchol) agit à titre de catalyseur pour la réduction du dioxyde de carbone en présence
d’hydroboranes (HBR2 = HBcat (catécholborane), HBPin (pinacolborane), 9-BBN (9borabicyclo[3.3.1]nonane), et BH3.SMe2), pour générer des produits du type CH3OBR2 ou
(CH3OBO)3 qui peuvent facilement être hydrolysés en méthanol. Les rendements peuvent
atteindre 99% avec formation exclusive de CH3OBR2 ou (CH3OBO)3 avec des efficacités
catalytiques atteignant 2950 tours catalytiques à une fréquence de 853 tours catalytiques par
heure. De plus, le catalyseur possède un caractère vivant: une fois la première charge
catalytique consommée, l’activité reprend immédiatement à l’ajout d’une charge
additionnelle de réactifs.
4.1.2
Abstract
In this work, we report that organocatalyst 1-Bcat-2-PPh2-C6H4 ((1); cat = catechol) acts as
an ambiphilic metal free system for the reduction of carbon dioxide in presence of
hydroboranes (HBR2 = HBcat (catecholborane), HBPin (pinacolborane), 9-BBN (9borabicyclo[3.3.1]nonane), and BH3.SMe2) to generate CH3OBR2 or (CH3OBO)3, products
that can be readily hydrolysed to methanol. The yields can be as high as 99% with exclusive
formation of CH3OBR2 or (CH3OBO)3 with TON (turn-over numbers) and TOF (turn-over
frequencies) reaching >2,950 and 853 h-1, respectively. Furthermore, the catalyst exhibits
“living” behavior: once the first loading is consumed it resumes its activity on adding another
loading of reagents.
69 4.1.3
Introduction
It is widely known that carbon dioxide is a green-house gas and one of the most important
contributors to global warming and several political initiatives have been put forward to
reduce carbon dioxide emissions.14 Most of the current systems known to catalyze the
reduction of CO2 into valuable products use transition metals,224,225 including notably the
reverse water-gas shift reaction to generate carbon monoxide which in turn can be
transformed into several useful chemicals.226 Recently, some homogeneous organometallic
systems have shown promise in generating valuable chemicals.68,70,71,80,104,125,130–133,227–229
The most active systems to date for the reduction of CO2 into high hydrogen content
molecules include a ruthenium phosphine complex125 and a nickel pincer complex71, using
respectively H2 and HBcat (HBcat = catecholborane), to generate MeOH from CO2, and an
iridium catalyst that can reduce CO2 into methane using hydrosilanes as a hydrogen source
with turn-over numbers (TON) up to 8,300.130
Recently, a variety of transition metal-free systems have emerged for carbon dioxide
activation and functionalization. Indeed, it has recently been shown that Lewis acidic Et2Al+
species can catalytically reduce carbon dioxide to methane.90 Similarly, silyl cations can
catalytically reduce CO2 to a mixture of benzoic acid, formic acid and methanol.89 However,
both systems greatly lack in selectivity and generate undesirable alkylation by-products. An
avenue of interest for carbon dioxide activation is the use of “Frustrated Lewis Pairs” (FLP),
work pioneered by Stephan and Erker.103 Since this initial discovery, many ambiphilic
systems have been shown to be active in the stoichiometric fixation of CO2.117,156 Piers
demonstrated elegant use of this concept for the catalytic reduction of CO2 into methane
using the robust TMP/B(C6F5)3 (TMP = 2,2,6,6-tetramethylpiperidine) system with Et3SiH,
albeit with limited turnovers.104 It has been shown that the FLP system consisting of
PMes3/AlX3, (Mes = mesityl, X=Cl, Br) not only binds CO2 but also reduces it to methanol
using BH3.NH3 as hydrogen source.106 Alternatively, O’Hare and Ashley demonstrated that
CO2 could be hydrogenated using TMP/B(C6F5)3.63
70 Unfortunately, the two last systems require stoichiometric amounts of FLP. Although
interesting in concept, none of the FLP or ambiphilic systems reported to date demonstrate
good catalytic activity for carbon dioxide reduction. The only efficient organocatalytic
system reported to date for the reduction of CO2 into methanol use highly Lewis basic Nheterocyclic carbene catalysts and diphenylsilane as hydrogen source with turn-over
frequencies (TOF) of 25 h-1 at 25°C.138
Our research program targets ambiphilic systems with little “frustrated” character and/or
weak Lewis acidity and basicity.115,121 One ambiphilic system of interest is that of aryl
bridged phosphine-boranes extensively studied by Bourissou and collaborators.179,184 These
molecules have been shown to be quite robust, stable and easy to synthesize. More recently,
they have been used in the activation of singlet oxygen190 and as organocatalysts for the
Michael addition reaction,191 but to our knowledge the activity of these molecules for carbon
dioxide reduction has not been investigated. Here we report that the 1-Bcat-2-PPh2-C6H4
ambiphilic system is one of the most active catalysts for the selective catalytic reduction of
carbon dioxide to methanol.
71 4.1.4
Results and discussion
Although several ambiphilic phosphine-boranes were prepared by Bourissou,179,184,190,191 the
synthesis of the catecholborane derivative 1-Bcat-2-PPh2-C6H4 (1) was never reported. The
air-stable product is easily synthesized in 80% yield from previously reported o-lithiated
triphenylphosphine using a known synthetic pathway 212
Multinuclear NMR characterization of species 1 demonstrates this molecule to be monomeric
in solution having no observable P-B interaction. The 31P{1H} and 11B{1H} NMR chemical
shifts are respectively of -4.57 and 33.1 ppm. The solid state structure (Figure 4-1) does not
show any evidence of P-B interaction, the latter distance being quite long (3.28 Å).
Figure 4-1: ORTEP drawing of 1 in the asymmetric unit cell.*
*Anisotropic atomic displacement ellipsoids shown at the 50% probability level. Selected bond
lengths [Å] and angles [°] : P(1)-C(13) 1.8429(1), C(18)-B(1) 1.553(3), C(13)-C(18) 1.412(3), C(17)C(18)-B(1) 114.71(17), C(13)-C(18)-B(1) 127.07(16), C(18)-C(13)-P(1) 119.67(13), C(14)-C(13)P(1) 120.86(1).
72 Unsurprisingly, exposing 1 to 1 atmosphere of CO2 at room temperature resulted in no
spectroscopic change in solution (1H, 31P, and 11B NMR spectroscopy). Although no adduct
was observed between CO2 and 1, the addition of 100 equivalents of HBcat to a 9 mM
solution of 1 in benzene-d6 in a J-Young NMR tube under one atmosphere of CO2 resulted
in the formation of a white precipitate after 24 hours. This was characterized as catBOBcat
based on spectroscopic comparison with the independently synthesized product (Scheme 41).
Scheme 4-1: Reduction of CO2 in presence of HBcat and catalyst 1.
Monitoring of the solution using 1H NMR spectroscopy showed the presence of a single new
peak at 3.37 ppm attributed to CH3OBcat by comparison to the independently synthesized
product. Hydrolysis of the latter product produces methanol, which was confirmed using GCFID. As expected, carrying out the same reaction under an atmosphere of 13CO2 shows the
formation of 13CH3OBcat with the expected 1JC-H of 145 Hz.71
73 Figure 4-2: Turn-over numbers (TON) for the formation of CH3OBcat.*
*Results from a 9 mM solution of 1 in benzene-d6 in the presence of 100 equivalents of HBcat under
2 atmospheres of CO2. The TON’s are based on the number of hydrogen atoms transferred to CO2.
Reactions were carried out at (♦) 23°C and (■) 70 °C.
Monitoring the reduction of CO2 in presence 100 equivalents of HBcat and 1 using 1H and
31
P{1H} NMR spectroscopy showed an induction period of 30 minutes where no
spectroscopic change was observed in the solution. However, after the induction period the
reaction started readily and after 2 hours a 34% yield (TON = 34, TOF = 17 h-1) of CH3OBcat
was obtained (Figure 4-2, ♦). The rate of the reaction diminished as the reaction progressed,
suggesting that conversion is dependent on the concentration of HBcat in solution. Indeed,
50% conversion to CH3OBcat was obtained in less than 5 hours and a yield of 69% of
CH3OBcat was observed after a period of 24 hours. The reduction of CO2 also proceeded in
the presence of 100 equivalents of BH3.SMe2 to generate (CH3OBO)3 but a longer induction
period was observed (> 2 hours; Figure 4-3, ♦).
74 Figure 4-3: Turn-over numbers (TON) for the formation of (CH3OBO)3.
*Results from a 9 mM solution of 1 in benzene-d6 in the presence of 100 equivalents of BH3•SMe2
under one atmosphere of CO2. The TON’s are based on the number of hydrogen atoms transferred
to CO2. Reactions were carried out at (♦) 23°C and (■) 70 °C.
Nevertheless, the conversion to the methoxyborane species is rapid once catalysis starts,
obtaining respectively 108 and 200 TON at 2 and 5 hours after the induction period
(respective TOF of 54 and 40 h-1). After a period of 14 hours, a TON of 257 was obtained.
The TON numbers being greater than 100 suggests that all hydrogen atoms from BH3•SMe2
are available for the reduction of CO2. To our knowledge, it represents the first time that BH3
is used as a hydrogen source for the catalytic reduction of CO2 to methanol. It is interesting
that the catalyst remains active even if BH3 is known to coordinate phosphine moieties, which
could inhibit catalysis; it is thus logical to presume that the longer induction period is caused
by a competition between BH3 and CO2 for coordination to the catalyst. BH3 is of great
interest since it has the highest hydrogen content of any hydroborane.
75 A factor that dramatically increased the efficiency of the catalytic system was temperature.
Heating a solution of 1 with 100 equivalent of HBcat to 70 °C under one atmosphere of CO2
generated CH3OBcat without any observable induction period (Figure 4-2, ■). After 36
minutes, a TON of 86 was observed (TOF = 143 h-1), which increased to 92 after a period of
90 minutes (Table 4-1, entries 1-2). After letting the solution rest for a 24-hour period,
another loading of 100 equivalents of HBcat was added and the solution reheated to 70 °C.
The catalytic reaction resumed, but with a rate that seemed somewhat slower (an overall TON
of 136 after 30 minutes), possibly due to the presence of a large quantity of precipitate in the
solution (catBOBcat) that reduced the homogeneity of the solution. However, 60 minutes
after the addition of the second loading a TON of 185 was measured (Table 4-1, entries 3-4),
a similar yield to that observed in the first run. Such behaviour is reminiscent of a durable
and “living” catalyst. Under similar conditions, BH3•SMe2 proved to be an excellent
hydrogen source, generating 90% yield of (CH3OBO)n in 67 minutes (TON = 271). A TON
of 211 was obtained after only 13 minutes, representing a TOF of 973 h-1 (Figure 4-3, ■).
The latter result is remarkable since the highest TOF reported for the reduction of CO2 for a
methanol derivate is 495 h-1 by a homogeneous nickel catalyst using HBcat as an hydrogen
source.4d The reaction was also carried out using other hydroborane sources. The
significantly lower activity of pinacolborane compared to HBcat is not surprising since it is
known that HBPin is less reactive for the hydroboration reaction.230 Similarly, 9-BBN only
showed 34 TON in a 3-hour period (entry 7).
76 Table 4-1: Reduction of CO2 with various hydroboranes.
a
Entry
Borane
#eq.
Time (min)
TONb
TOF (h-1)
1
HBcat
100
36
86
143
2
HBcat
100
98
92
56
3
HBcat
100+100c
30
136
72
4
HBcat
100+100c
60
185
85
5
BH3.SMe2
100
67
271
242
6
HBPin
100
174
60
21
7
9-BBN
50d
174
34
12
8e
HBcat
300
60
145g
145
9f
HBcat
1,000
240
664g
166
10e
BH3.SMe2
300
60
853g
853
11f
BH3.SMe2
1,000
240
>2,950g
>737
12f
BH3.THF
300
60
340g
340
Reaction conditions : Unless noted otherwise, 2.0 mg (0.0053 mmol) of 1 in 0.6 mL of benzene-d6
at 70 °C b Based on mole of B-H consumed per mole of 1, determined by 1H NMR integration using
hexamethylbenzene as internal standard for entries 1-7, and determined by GC-FID with iPrOH as
standard for entries 8-11. c A second addition of 100 equivalents of HBcat was added 24 hours after
the first addition. d Limited at 50 equivalents because of low solubility of 9-BBN. e 2.0 mg (0.0053
mmol) of 1 in 3 mL of benzene at 70 °C under ca. 2 atmosphere of CO2. f 2.0 mg (0.0053 mmol) of 1
in 9 mL of benzene at 70 °C under ca. 2 atmosphere of CO2.
g
Quenched with excess H2O and
analyzed by GC-FID with iPrOH as an standard.
77 Since diffusion problems could limit the rate of the reaction when carried out in NMR
conditions, catalytic tests were carried out on a larger scale using Fisher-Porter bottles under
ca. 2 atmospheres of CO2. The products obtained were hydrolyzed to methanol and the turnover numbers were calculated based on the concentration of methanol using gas
chromatography with a flame ionisation detector. As can be observed in Table 4-1, the
activities observed at the NMR scale can be reproduced at larger scale and lower catalyst
loading. Indeed, the reduction of CO2 using 300 equivalents of HBcat and BH3.SMe2 gave in
one hour methanol in 48% and 95% yield, giving respectively TOF of 145 and 853 h-1 (Table
4-1, entries 8-9). It is notable that the TOF observed under large loading of hydroboranes are
consistent to those observed at the NMR scale at low conversion that were respectively of
143 and 973 h-1 for HBcat and BH3.SMe2. Catalysis using a 0.1% catalyst loading (1000
equivalents of substrate) in a four hour period also gave impressive results. In the presence
of HBcat, a TON of 664 was observed, which indicates that the rate of reaction remains the
same during the four-hour period even with a lower catalyst loading and a lower catalyst
concentration (Table 4-1, entry 10, TOF = 166 h-1). In presence of BH3.SMe2, all of the
substrate was consumed since the conversion to methanol was quantitative (Table 4-1, entry
11), once more suggesting that the TOF observed after one hour is conserved over a longer
reaction time. The reaction works also with BH3•THF, albeit less efficiently (Table 4-1, entry
11, TOF = 340 h-1).
78 Density functional theory studies at the B3PW91 6-31G** level of theory were performed to
obtain further insight at the mechanistic pathway, using HBcat as the hydrogen source. It
should be noted that only potential intermediates were considered in the following and the
results are summarized in Figure 4-4. Also, we did not account for the fact that HBCat is
known to degrade in presence of Lewis bases since control experiments have shown this
process to be marginal in our system.210 As observed experimentally, the coordination of CO2
to 1 to generate intermediate IM1 is disfavoured by 9.9 kcal.mol-1, in line with a weak
coordination of carbon dioxide as reflected by its geometry. Indeed, despite the bending of
the molecule (indicative of CO2 activation), the C-O bonds appear to be only slightly
elongated compared to free CO2 (1.28 and 1.21 Å). Nevertheless, this adduct can undergo
addition of HBcat to yield a novel species whose formation is favourable by 14.4 kcal.mol-1
compared to 1. Once the formation of the complex IM2 is achieved, the second reduction to
generate the formaldehyde-1 adduct (IM3) and catBOBcat is downhill by 33.3 kcal.mol-1.
The third reduction to regenerate the catalyst as well as CH3OBcat is an even more
exothermic process (34.0 kcal.mol-1). To summarize, as soon as the difficult coordination of
CO2 has taken place, the reduction is thermodynamically highly favourable.
Figure 4-4: Enthalpy profile (in kcal mol-1) for the reduction of CO2 by 1 and
catecholborane.
79 In order to confirm these computational results, 1 was reacted with methylformate in attempt
to generate an analogous compound to IM2, namely species IM2mf. In line with the DFT
results, where the adduct is predicted to be 3.9 kcal.mol-1 higher in energy than 1, no product
could be observed by NMR spectroscopy. However, upon the addition of 3 equivalents of
catecholborane without the presence of CO2, a 90% conversion to CH3OBCat was observed
after 20 hours at room temperature. This latter results suggest that although the formation of
the adduct IM2mf is thermodynamically slightly disfavoured, the reduction occurs in
presence of a hydroborane. It is also interesting to note that the intermediate IM3 is proposed
to be formed in both reduction pathways. A similar formaldehyde intermediate was identified
as a key intermediate in previous systems,71,138 but could not be observed experimentally.
While monitoring the reduction of CO2 in presence of hydroboranes and catalyst 1, only one
resonance in the 1H NMR spectra, a broad singlet at 5.20 ppm, could not be assigned to the
starting materials or products. Running the experiment in presence of
13
CO2 allowed the
observation of a 1JC-H of 151 Hz, suggesting that this species arise from the reduction of CO2.
In the latter experiment, a 1JP-C of 52 Hz was also observed both in the 13C{1H} and 31P{1H}
NMR spectra. The latter species could not be isolated from the catalytic mixture, being in too
small concentration in solution. However, when a solution of 1 was reacted with
paraformaldehyde and heated at 70 °C for 15 minutes, the same product was observed to be
formed with 74% conversion, as characterized by multinuclear NMR spectroscopy as IM3
(Figure 4-5).
Figure 4-5: Generation of a formaldehyde adduct from 1.
80 4.1.5
Conclusions
In summary, we have reported a metal free system for the reduction of carbon dioxide to
methanol using a borane as the reducing agent. The system is a robust living catalytic system
and generates TOF up to 973 h-1 and TON up to 2950 at 70°C under 1 atm of CO2, although
larger TON can be expected by additional loadings of hydroboranes. The key aspect of this
reported system compared to the other metal free systems for the activation of CO2 is the
weak interaction between the catalyst and carbon dioxide. Indeed, contrary to most
ambiphilic and FLP systems reported to date, no adduct formation is observed between 1 and
CO2. Nevertheless, CO2 being an ambiphilic molecule with its electrophilic carbon atom and
nucleophilic oxygen atoms available, does not require significant bonding interaction with
an ambiphilic catalyst to undergo reduction with hydroboranes. Once the first reduction has
occurred, following reductions occur readily to generate CH3OBR2. Preliminary results
demonstrate that the BPin analogue 1-BPin-2-PPh2-C6H4191 is an active catalyst for the CO2
reduction using BH3.SMe2, albeit working less efficiently than 1 (TOF of 24 h-1 in conditions
similar to entry 2 of Table 4-1). Current work focuses on optimizing the steric and electronic
properties at boron and phosphorous centers to obtain optimal catalytic activity.
Computational studies to unveil the full reaction mechanism are also well underway.
81 4.2 Experimental
4.2.1
General experimental
Unless otherwise specified, manipulations were carried out under an atmosphere of
dinitrogen, using standard glovebox and Schlenk techniques. Reactions were carried either
in a sealed J-Young NMR tube, in which case NMR conversions are indicated, or in standard
flame dried Schlenk glassware. Catalytic reactions over 1 atm of pressure were carried out in
a sealed Fischer-Porter bottle. All solvents were distilled from Na/benzophenone, benzened6 was purified by vacuum distillation from Na/K alloy. 1,2 Dibromobenzene was purchased
from Sigma-Aldrich and used as received. B(OMe)3 was dried on Na0 and distilled prior to
use. Bone dry CO2 was purchased from Praxair and used as received. 13CO2 (99% isotope
label) was purchased from Cambridge Isotope Laboratories and also used as received.
Lithiated o-bromophenyldiphenylphosphine, CH3OBcat and catBOBcat were synthesized in
good yields by following literature procedures.71,212,231
NMR spectra were recorded on a Agilent Technologies NMR spectrometer at 500 MHz (1H),
125.758 MHz (13C), 202.456 MHz (31P), 160.46 MHz (11B), on a Varian Inova NMR AS400
spectrometer, at 400.0 MHz (1H), 100.580 MHz (13C), 161.923 MHz (31P). 1H NMR and 13C
{1H} NMR chemical shifts are referenced to residual protons or carbons in deuterated
solvent. 11B {1H} was calibrated using an external reference of B(OMe)3. Multiplicities are
reported as singlet (s), broad singlet (s, br) doublet (d), triplet (t), multiplet (m). Chemical
shifts are reported in ppm. Coupling constants are reported in Hz. gHMQC, gHSQC, NOESY
2D, gDQCOSY and 1H{31P} NMR experiments were performed in order to properly assign
spectra. GC spectra were recorded on a Hewlett Packard GC-FID 6890 Series with an HP-5
(Crosslinked 5% PHME siloxane) column, using an isotherm at 50°C. Injection volumes
were 1 µL.
82 4.2.2
Synthesis of 1
Scheme 4-2: Synthesis of 1.
503 mg (1.47 mmol) of 2-(diphenylphosphino)phenyllithium was placed in a Schlenk vessel
and dissolved in ca. 10 mL of THF. The solution was cooled to -78°C and 1.0 mL (8.97
mmol) of B(OMe)3 was added at once. The solution was slowly warmed to r.t. and left to
react for 24 hours. The solution was filtered from the precipitated lithium salts and evaporated
to dryness in order to remove excess B(OMe)3. The resulting oily residue was then dissolved
in ca. 10 mL of toluene and added to a solution of 162 mg (1.47 mmol) of catechol in ca. 10
mL of toluene. The mixture was heated at 90°C for a period of 3 hours. After cooling the
solution to room temperature, the solution was concentrated to half its original volume by
evaporation causing the unreacted catechol to precipitate. The solution was then filtered and
evaporated to dryness, yielding 450 mg (80%) of pure 1 as a white powder. Colorless needles
suitable for X-ray diffraction were grown from a saturated solution of toluene at -40°C.
(Scheme 4-2)
Figure 4-6: Assignment of spectra for 1.
83 1
H NMR (500 MHz, benzene-d6): δ 8.08 (d, 1H, 3JH-H=7.0 Hz; Hc), 7.46 (t, 4H, 3JH-H=7.3
Hz; Hh); 7.09-6.99 (m, 9H; Hd,e,f,g,i); 6.97 (dd, 2H, 3JH-H=5.8 Hz, 4JH-H=3.4 Hz; Ha), 6.71
(dd, 2H, 3JH-H=5.8 Hz, 4JH-H=3.4 Hz; Hb).
11
31
P{1H} (203 MHz, benzene-d6): δ -4.57 (s).
B{1H} (160.46 MHz, benzene-d6): δ 33.1 (s) 13C{1H} (101 MHz, benzene-d6): δ 148.9 (s;
C1), 145.8 (d, 1JP-C= 20.5 Hz ; C9), 138.6 (d, 2JP-C= 13 Hz; C11), 136.7 (d, 2JP-C= 10.0 Hz;
C5), 134.4 (d, 1JP-C= 20.0 Hz; C10), 133.9 (s; C6 or C7 or C8), 131.7 (s; C6 or C7 or C8),
128.78 (s; C13), 128.83 (d, 3JP-C= 2.9 Hz; C12), 128.4 (s; C6 or C7 or C8), 122.9 (s; C3),
112.8 (s; C2). M/Z (M+H+) = 380.1227 calc = 380.11.
4.2.3
General procedure for catalytic reduction of carbon dioxide
2.0 mg (5.3 μmol) of 1 was dissolved in ca. 0.6 mL of benzene-d6 containing an internal
standard of hexamethylbenzene. The reducing agent (HBcat, HBPin, 9-BBN or BH3.SMe2)
was added to the solution and the mixture was introduced in a J-Young NMR tube. The Jyoung NMR tube was frozen in a liquid nitrogen bath after which ca. 1atm. of CO2 was
introduced. The reaction was then followed by NMR spectroscopy. Subsequent loadings of
reductants and CO2 were introduced in the J-young NMR tube using the exact same
procedure. Yields are reported by
1
H NMR according to the integration of
hexamethylbenzene which was used as an internal standard. The full characterization of the
products obtained with HBcat and BH3.SMe2 is described below, but the characterization of
the reduction of CO2 with 9-BBN and HBPin is solely based on 1H NMR spectroscopy and
on the similarity with other reagents.
Important Notice: A 1:1 mixture of PPh3 and reducing agents (BH3•SMe2, HBcat) did not
show any activity after 24 h at 70°C.
84 4.2.4
General procedure for big scale catalytic reduction of carbon dioxide
2.0 mg (5.3 mmol) of 1 was dissolved in ca. 1 mL of C6H6 per 100 equivalents of reducing
agent (BH3.SMe2 or HBcat). The reducing agent was added to the solution and the mixture
was introduced in a Fischer-Porter bottle. The Fischer-Porter bottle was frozen in a liquid
nitrogen bath and put under vacuum for 15 min after which ca. 2 atm of CO2 was introduced
(in order to ensure sufficient amount of CO2 for complete reduction). The mixture was left
to react at 70°C for the specified amount of time after which the contents of the FischerPorter bottle were quenched using ca. 2 mL of H2O per 100 equivalents of reducing agent.
This was left to stir at r.t. for 1 hour to ensure complete quenching of the mixture. Yields
have been determined using GC-FID with isopropanol as internal standard. Methanol yields
have been determined using GC-FID with isopropanol as the internal standard. A calibration
plot was obtained using three aqueous solutions of isopropanol (1% v/v) containing methanol
(0.5, 1 and 2 % v/v) to ensure linearity of the iPrOH/MeOH signal.
4.2.5
Reactivity with methylformate
2.0 mg (5.3 mmol) of 1 was dissolved in ca. 0.6 mL of benzene-d6. 4 µL (65 mmol) of
methylformate was added to the solution and the mixture was introduced in a J-Young NMR
tube for NMR characterization. 26.8 mg (224 mmol) of HBcat was then added to the solution
and left to react at 23° C for 20 hours. Yield is reported by NMR integration using toluene
as internal standard. Yield: approx. 90% conversion (Scheme 4-3).
Scheme 4-3: Catalytic reduction of methylformate by 1 and HBcat.
85 4.2.6
In-situ preparation of IM2
8.7 mg (22.9 mmol) of 1 was dissolved in ca. 0.6 mL of benzene-d6 along with excess (2.3
mg) of paraformaldehyde. The mixture was heated for 15 min at 70 °C in a J-Young NMR
tube before NMR characterization. Crude Yield (approx 74% by NMR integration). Not
isolated.
4.2.7
Crystallographic information
A colorless block crystal of 1 with approximate dimensions of 0.44 x 0.38 x 0.32 mm was
mounted on CryoLoops with Paratone-N and optically aligned on a Bruker SMART APEXII X-ray diffractometer with 1K CCD detector using a digital camera. Initial intensity
measurements were performed using a fine-focused sealed tube, graphite-monochromated,
X-ray source (Mo Kα, λ = 0.71073 Å) at 50 kV and 30 mA. Standard APEX-II213 software
package was used for determining the unit cells, generating the data collection strategy, and
controlling data collection. SAINT214 was used for data integration including Lorentz and
polarization corrections. Semi-empirical absorption corrections were applied using SCALE
(SADABS215). The structures of all compounds were solved by direct methods and refined
by full-matrix least-squares methods with SHELX-97217 in the SHELXTL6.14 package.. All
of the H atoms (on C atoms) were generated geometrically and refined in riding mode.
Crystallographic information for all obtained phases is summarized in Table 4-2. Atomic
coordinates and additional structural information are provided in the .cif files of the
Supporting Information. Crystallographic data have been deposited with CCDC (CCDC No.
936490). These data can be obtained upon request from the Cambridge Crystallographic Data
Centre, 12 Union Road, Cambridge CB2 1EZ, UK, E-mail: deposit@ccdc.cam.ac.uk, or via
the Internet at www.ccdc.cam.ac.uk.
86 Table 4-2: Crystal data and structural refinements for compound 1.
Compound
1
Formula
C24 H18BO2P
Mr (g mol−1)
380.16
Crystal size
0.44 x 0.38 x 0.32 mm
Crystal system
Triclinic
Space group
P -1
λ (Å)
0.71073
T (K)
150(2)
a (Å)
10.492(2)
b (Å)
10.674(2)
c (Å)
11.004(2)
α (o)
93.029(9)
β (o)
114.593(9)
γ (o)
116.839(8)
V (Å3)
954.8(3)
Z
2
ρcal. (g cm−3)
1.201
F(000)
396
μ (mm−1)
0.161
θ for data collection (o)
1.9 to 28.4
Reflection collected
19023
Data /parameters
19023 / 4709
GOOF
1.078
R1(wR2) (I > 2σ(I))
0.0537 , (0.1504)
R1(wR2) (all data)
0.0687, (0.1705)
87 4.2.8
Computational Details:
Calculations were performed with the GAUSSIAN 03 suite of programs.218 The
B3PW91219functional was used in combination with the 6-31+G** basis set for B,C,H, and
O atoms,220,221 and the SDD basis set with an additional polarization function (one d function
with a 0.34 exponent and a 1.0 contraction coefficient) for P atoms,222 . The stationary points
were characterized as minima by full vibration frequencies calculations (no imaginary
frequency). All geometry optimization were carried out without any symmetry constraints.
Cartesian coordinates and free energies as well as additional experimental data including
NMR spectra are available free of charge via the ACS website: http://pubs.acs.org/
88 4.3 Conclusions and perspectives
A number of ideas and aspects of importance have been discussed in chapters 3 and 4.
Chapter 3 dealt with the development of a unique arene bridged FLP system bearing three
pendant Lewis basic phosphine groups. Even though these non-bulky and rather unreactive
phosphines are able to coordinate to the aluminum center, the system as whole still exhibits
FLP reactivity. In fact, it is the first aluminum-based FLP system to exhibit reversible binding
of carbon dioxide. Unfortunately, no direct evidence was found to support the validity of the
initial strategy involving simultaneous substrate and CO2 activation. In addition, the
aluminum system decomposed readily in the reaction conditions, underlining the problems
associated with the use of highly oxophilic aluminum-based catalysts in the reduction of CO2.
On the other hand, careful examination of the decomposition products led to the discovery
of a very active organocatalyst for the transformation of carbon dioxide.
Indeed, the identification, synthesis and characterization of the phosphine-borane reduction
product 1 allowed us to study its reactivity and consequently, to discover the first metal-free
system for the catalytic reduction of CO2 using hydroboranes, a process that was thought to
be exclusive to transition metals. It should be emphasized that the catalyst selectively
produces methoxyboranes, species which are readily hydrolysed to methanol, a very
desirable end product. Furthermore, the catalyst is the most active ever reported for the
conversion of CO2 to methanol, including transition metal based systems. It is also the first
system to employ the cheap, high hydrogen content borane BH3•SMe2 as reductant. One of
the most notable aspects of the catalyst is its simplicity. With a simple phosphine as a Lewis
base and a relatively unreactive catechol borane Lewis acid, the catalysts exhibits good
stability and is easily prepared on a large scale from readily available material.
89 However, the impact of this work is not limited to the properties of the catalytic system.
Indeed, this work has paved a new path for the metal-free catalytic hydroboration reaction,
opening the way for the development of future systems that would in turn allow a better
comprehension of CO2 reduction chemistry. Interestingly enough, the catalyst exhibits none
of the classical FLP features such as strongly Lewis acidic and basic centers or steric bulk.
Before this report, there was a general consensus in the field that the use of such extremes
was essential to FLP reactivity. Such a mindset limited the development potential of FLPs,
leading in most cases to stoichiometric activation or poor catalytic results. This work
illustrated that it was possible to use simple and stable molecules as effective FLP systems,
an advantageous alternative to classical FLPs. But, as simple as the catalyst appears to be,
the underlying reactivity had yet to be fully understood. It is in an effort to truly understand
its mode of action that we set out to conduct a thorough mechanistic investigation of the
catalytic system.
90 5 Mechanistic investigations
5.1 Metal mediated catalytic reductions of CO2
After we had reported the first isolation of formaldehyde directly produced from a CO2
reduction process through ambiphilic trapping of the intermediate,199 Sabo-Etienne and coworkers developed another trapping strategy, allowing them to generate formaldehyde from
CO2232 and later to generate it catalytically.233 The simple but ingenious maneuver of adding
a primary amine to the previously reported catalytic process133 allowed them to trap
formaldehyde as it was generated, forming an imine. Upon hydrolysis, the formaldehyde
could be released. An intriguing aspect of this report is that the use of the additive results in
better turnovers than those reported in the original system (TON=108 vs TON=20
respectively), presumably by eliminating competitive reaction pathways (Figure 5-1a).
Hill and co-workers were able to develop a new alkaline earth based catalytic system using
BCF as a co-catalyst and HBPin as a reductant (Figure 5-1b). Unfortunately, the system did
not generate more than 10 turnovers, even after 4 days at 60°C.234 In another contribution,
Piers revisited the strategy of using scandium based catalysts with pronounced Lewis acidity
for the hydrosilylation of CO2 by Et3SiH. Ingeniously, the over-reduction to methane could
be mitigated by stabilizing the BCF co-catalyst through B-O interaction with the catalyst,
leading to a very efficient system for the hydrosilylation of CO2 to a bis silyl acetal
Et3SiOCH2OSiEt3 (Figure 5-1c).235 In optimized conditions, a TON of 3400 was reached
after extended reaction time and multiple re-loadings of CO2.
91 Figure 5-1: New metal based catalytic systems for the reduction of CO2.
Undoubtedly, the field of metal-catalyzed reduction of CO2 was still evolving rapidly. By
refining their strategies, Sabo-Etienne and Piers were able to generate new derivatives from
CO2 reduction while simultaneously increasing the activities, a notable advancement. Still,
despite the novelty of using alkaline-earth metals for CO2 reduction, efficient systems for the
generation of methanol had yet to be uncovered.
92 5.2 Metal-free activation of CO2
While we were investigating the reaction mechanism of FLP mediated catalysis, a number
of advances have been made in the design of increasingly elaborate FLP systems for
stoichiometric CO2 activation. Figueroa reported a relatively complex (boryl)iminomethane
system capable of CO2 binding (Figure 5-2a).236 Wang also reported CO2 activation through
highly reactive amido/boronium centers, but without surprise, no further reactivity was
observed (Figure 5-2b).237 Similarly, Muller developed an entire library of highly Lewis
acidic silylium based Lewis pairs, but the reactivity seemed once again limited to
stoichiometric CO2 fixation (Figure 5-2c). Three years later, Uhl reported a system incredibly
similar to our original report of methylene bridged phosphine-alane (Figure 5-2d).
Unsurprisingly, similar reactivity was observed as the only difference was the replacement
of the methyl substituents on phosphorus by tert-butyl groups. Except Figueroa’s system
which uses a very reactive amide to make-up for the weakly acidic borane, all the systems
still shared the characteristic of highly active acceptor fragments. Perhaps that the choice of
these components could explain why no further reactivity was documented for these systems.
Nonetheless, with these reports to add to the library of FLPs capable of CO2 binding, new
ideas and concepts were being put forward, enriching the global available knowledge in the
field.
Figure 5-2: New FLP systems for the stoichiometric binding of CO2.
93 5.3 Metal-free reduction of CO2
Not long after the publication of the phosphine-borane catalyst by the Fontaine group, two
more FLP systems were reported for the hydroboration of carbon dioxide. First, the ring
expansion of a NHC/9-BBN adduct was shown to generate an active hydroboration catalyst.
The catalyst was able to reach a TON of 240 using the advantageous BH3•SMe2 as a reducing
agent after 11 hours at 60 °C (Scheme 5-1).238 While the activity of this system was
substantially lower than that of our phosphine-borane, this report convincingly reinforced the
concept that simple FLPs could act as active catalysts. Two more catalytic reports of basepromoted hydroboration using 9-BBN as reducing agent were also published in the
meantime.239,240 However, since the reactivity of similar systems will be developed in detail
during the next chapter, the discussions related to these systems are postponed to chapter 6.
Scheme 5-1: Catalytic hydroboration of CO2 after ring-expansion of an NHC-9-BBN
adduct.
With the possibility of using hydroboranes to reduce CO2 being less than one year old, much
remained to be understood. In fact, the true purpose behind the development of these
hydroboration catalysts was to gain as much insight as possible into the factors governing
CO2 reactivity in hope of one day being able to design a system capable of CO2 reduction in
a sustainable fashion. The following chapter is aimed at trying to understand the fundamental
reactivity patterns associated with the FLP mediated reduction of CO2.
94 5.4 Research article: Reducing CO2 to Methanol using Frustrated Lewis
Pairs: On the Mechanism of Phosphine-Borane Mediated Hydroboration
of CO2
5.4.1
Résumé
Le mécanisme complet pour l’hydroboration du CO2 par l’organocatalyseur hautement actif
1-Bcat-2-PPh2−C6H4 (cat = catechol) a été déterminé en utilisant des méthodes
computationnelles et expérimentales. Il a été démontré que la paire de Lewis intramoléculaire
comme est active à chaque étape de la réduction. Contrastant avec les systèmes existants
basés sur les PLFs, l’absence d’encombrement stérique autour du fragment basique permet
l’activation de l’agent réducteur alors que la réactivité modérée des centres actifs permet le
relargage des produits de réduction. L’activation simultanée de l’agent réducteur et du
dioxyde de carbone représente la clé pour une activité catalytique optimale à chaque étape
de la réduction.
5.4.2
Abstract
The full mechanism for the hydroboration of CO2 by highly active ambiphilic organocatalyst
1-Bcat-2-PPh2−C6H4 (cat = catechol) was determined using computational and experimental
methods. The intramolecular Lewis pair was shown to be involved in every step of the
stepwise reduction. Contrasting with traditional FLP systems, the lack of steric hindrance
around the Lewis basic fragment allows activation of the reducing agent while moderate
Lewis acidity/basicity at the active centers promotes catalysis by releasing the reduction
products. Simultaneous activation of both the reducing agent and carbon dioxide is the key
to efficient catalysis in every reduction step.
95 5.4.3
Introduction
The general concern over the increase of the CO2 concentration in the atmosphere and its
influence on climate change has led to several worldwide initiatives to control the emissions
of this green-house gas. Although several carbon capture technologies have been developed,
the possibility of using CO2 as a C-1 feedstock to synthesize valuable chemicals could be an
important financial incentive for reducing CO2 emissions.226 For these reasons, carbon
dioxide transformation has attracted much scientific attention over the past decade.14 Of
particular interest and at the core of the methanol economy is the transformation of CO2 into
high hydrogen content hydrocarbons since such technology could help generate “green”
energy vectors that are needed on a global scale to replace fossil fuels.22 Although most of
the reported systems use heterogeneous catalysts, some homogeneous transition-metal based
catalytic systems have been developed for the reduction of CO2 to formic
acid,75,77,241formate,70,130,131,229,242–245formaldehyde,233 methanol,68,71,125,227 methane70,104,130–
132
and acetals.235
Organocatalysts, as species not comprised of transition metals, are still scarce in the field of
CO2 functionalization to valuable chemicals.229 Notable systems include highly Lewis acidic
aluminum species90,246 and silyl cations89 which have been shown to reduce CO2 with low
selectivity to mixtures of products comprising methane, methanol and a number of alkylation
by-products. Pioneering work by Stephan and Erker demonstrated the capacity of FLPs to
bind CO2 which led to the subsequent discovery of a number of ambiphilic systems capable
of stoichiometric fixation103,117,119,137,156,237,247–249 However, except for the reduction of CO2
to CO by carbodiphosphoranes,160 no other catalytic reduction of CO2 was reported for these
systems. The PMes3/AlX3 (X=Cl, Br, C6F5) FLP mediated the stoichiometric reduction of
CO2 using NH3BH3, but the system had to be hydrolysed in order to free the methoxide
fragment and generate methanol.63,106,107,109,209 Piers also developed an FLP based catalytic
reduction of CO2 to methane by using hydrosilanes, albeit with limited turnovers.104,105 Ying
reported that N-heterocyclic carbenes could be used as catalysts to reduce CO2 to methanol
in the presence of hydrosilanes with a TOF of 25 h-1 at room temperature.138
Recently,
Cantat demonstrated that some nitrogen bases, such as guanidines and amidines, could be
used as catalysts for the reduction of CO2 to formamide using hydrosilanes or to
methoxyboranes using 9-borabicyclo[3.3.1]nonane (9-BBN) and HBcat.239 Stephan also
96 reported that phosphines could catalyze the reduction of CO2 to methoxyboranes using 9BBN as the reducing agent.240
Our group recently reported that organocatalyst 1-Bcat-2-PPh2-C6H4 (1), which can also be
generated by the addition of HBcat to precatalyst Al(2-PPh2-C6H4)3,139 is highly active for
the hydroboration of CO2 to methoxyboranes, species that can be readily hydrolyzed to
methanol, using a variety of hydroboranes.199,250 Using catecholborane or high hydrogen
containing BH3SMe2, a turn-over frequency (TOF) of 973 h-1 and turn-over numbers (TON)
over 2950 were observed at a temperature of 70 °C. In a recent contribution, Stephan et al.
reported a similar ambiphilic system to be catalytically active in the hydroboration of CO2.238
Both of these systems have in common the weak Lewis acidity of the borane compared to
the strong Lewis acids normally used in classical FLP systems. Understanding the
fundamental process of this catalytic system and identifying the important reaction
intermediates is therefore of prime importance in order to unveil the full potential of
ambiphilic molecules and Frustrated Lewis Pairs as efficient catalysts. In order to determine
the true role of the catalyst in every step of the reduction process, a thorough computational
study has been carried out and complemented by experimental studies. Herein, we report the
full mechanism for the first metal-free catalytic hydroboration of CO2 to methoxyboranes. A
closer look at the critical steps of the reaction underlines some of the key aspects of the
mechanism and offers an unprecedented insight and a novel way to approach ambiphilic
molecule and FLP-mediated catalysis.
97 5.4.4
Computational details
All the calculations were performed on the full structures of the reported compounds.
Calculations were performed with the GAUSSIAN 03 and GAUSSIAN 09 suite of
programs.251 While the wB97XD functional252 was qualified as promising by Grimme124
and was used to accurately describe the mechanism of FLP mediated hydrogenation of
alkynes ,253 its use for the modelization of 1 showed a very different geometry than the
reported crystal structure.199 Based on the accurate description of 1 with respect to the
reported structure, the B3PW91219 functional was used in combination with the 6-31G**
basis set for B, C, H, and O atoms220,221 and the SDD basis set with an additional polarization
function (one d function with a 0.34 exponent and a 1.0 contraction coefficient) for the P
atom.222 The transition states were located and confirmed by frequency calculations (single
imaginary frequency). Intrinsic reaction coordinate calculations (IRC) have been performed
to confirm the link between the transition states and the reactants/products. The stationary
points were characterized as minima by full vibration frequencies calculations (no imaginary
frequency). All geometry optimizations were carried out without any symmetry constraints.
The energies were then refined by single point calculations to include dispersion at the
B97D/6-31G** level of theory.254 The energies were further refined by single point
calculations to account for solvent effects using the SMD solvation model255 with benzene,
the experimental solvent. Since the entropic contribution in solution cannot be accurately
predicted by standard quantum mechanical calculations and are often greatly
overestimated,256–262 it was shown that enthalpy values are a better approximation. Thus, the
energies are reported in terms of enthalpy with the free energy reported between parentheses.
Bond rotations and their associated transition states were not calculated as it is clear that their
energy will be much lower than the energy barriers associated with the reduction steps in
such a system and are therefore trivial. All structures with their associated free enthalpy and
Gibbs free energies as well as their cartesian coordinates are fully detailed in the supporting
information.
98 5.4.5
General remarks
At this point, it is very important to mention that the entropic contributions for gas phase
calculations have been shown to be overestimated by 50-60 % for a two component
reaction.259 Thus, for the majority of the reported reactions (where three components come
together), the entropic contribution, and therefore the free energy, is expected to be greatly
overestimated. Some strategies have been used to better estimate the entropic contribution,
notably by performing the vibrational analysis at up to 1324 atm263 to account for better
entropy correction, but for this study the free energies are provided without any correction.
Even though entropic contributions are important and cannot be neglected, the enthalpy
values provide more accurate comparisons for similar intermediates.
The hydroboration of carbon dioxide to methoxyboranes is a stepwise process that occurs
through three subsequent reduction processes. First, CO2 is reduced to a formatoborate,
which is then reduced to formaldehyde. Finally, the formaldehyde is reduced to
methoxyboranes (Scheme 5-2). The upcoming sections will consider these three reductions
steps one by one in order to simplify the discussion.
Scheme 5-2: Schematic representation of the stepwise hydroboration of CO2 to methoxyboranes
using hydroboranes (H[B]).
99 5.4.6
First reduction step: CO2 to HCOOBcat
As expected, the direct reduction of carbon dioxide by catecholborane (HBcat) is kinetically
forbidden as the associated transition state TS1 was located at +34.2 (+47.7) kcal.mol-1
higher than the reactants. Experimental results support this hypothesis as heating HBcat in
the presence of 1 atm of CO2 at 70 °C for 48 hours did not yield any observable CO2 reduction
product, even in the presence of PPh3.199 As such, a catalyst is required to lower the energy
barrier and provide access to HCOOBcat (IM1, -11.0 (+1.5) kcal.mol-1).
As was previously reported, the adduct between CO2 and ambiphilic compound 1 (1-Bcat-2PPh2-C6H4) was never observed spectroscopically.199 Theoretical results suggest that the
adduct formation between 1 and CO2 is endothermic by +6.8 kcal.mol-1 (IM0, +6.8 (+20.8)
kcal.mol-1). The binding of CO2 induces a pyramidalization at the boron center, modifying
the coordination environment of the catalyst. In fact, while the sum of the angles around the
boron center in 1 is 359.9 °, indicative of a sp2 planar geometry, the sum of the same angles
in IM0 is 334.4 °. Intermediate IM0 counts four Lewis basic sites that can potentially bind
catecholborane (HBcat). Indeed, coordination of the hydroborane to a nucleophilic site is
required to promote the hydroboration of carbonyl-containing fragments.264
In order to
simplify the discussion, the Lewis basic sites were numbered 1 through 4 as illustrated in
Scheme 5-3.
Scheme 5-3: Reaction of 1 with CO2, generating IM0 illustrating the potential binding sites
for HBcat.
100 Firstly, no transition state (TS) could be located for the reduction of CO2 via the coordination
of HBcat to sites 1 and 3, mainly due to the geometric constraints that prevent the hydride
transfer to the carbonyl moiety. Consequently, all four possible pathways (labeled A through
D), involving coordination to the two remaining sites as well as direct coordination to the
phosphorus atom of 1 were studied for the initial reduction step and are described below. The
most direct reduction path (pathway A, Scheme 5-4) involves the coordination of HBcat to
site 4, generating the classical 4-membered ring hydroboration transition state (TS1A) as
suggested by Dimare for the reduction of a variety of aldehydes and ketones by
hydroboration.264 For such a process, the barrier was found to be relatively high, although
accessible at +24.4 (+55.7) kcal.mol-1, generating IM1A (-13.1 (+16.0) kcal.mol-1).
Therefore, pathway A does not appropriately represent the reduction of CO2 to HCOOBcat
by catalyst 1.
Scheme 5-4: Pathway A: hydroboration reaction of CO2 through a classical 4-membered
transition state. [B] = Bcat.
101 Coordination of HBcat to site 2 generates intermediate IM0B (-2.1 (+26.5) kcal.mol-1) which
is only slightly thermodynamically stabilized with respect to IM0 (Pathway B, Scheme 5-5).
From the adduct IM0B, the hydride can be transferred to the carbon atom of CO2 through a
6-membered ring transition state (TS1B, +16.7 (+46.0) kcal.mol-1), yielding IM1B (-17.7
(+12.2) kcal.mol-1). Such reactivity is reminiscent of the hydroboration mechanism observed
with oxazaborozilidine catalysts developed by Corey et al. where the coordination of the
borane to a Lewis base promotes intramolecular hydride delivery. 265,266 It should be noted
that pathway B is kinetically more accessible than pathway A since the TS is 7.7 kcal.mol-1
lower in energy.
Scheme 5-5: Pathway B: hydroboration through coordination of HBcat to the catechol
fragment followed by intramolecular hydride delivery. [B] = Bcat.
A third pathway can be considered in which both the reducing agent and CO2 are
simultaneously activated. The phosphorus atom activates catecholborane while CO2 is
activated by the boron fragment. The coordination of the Lewis base increases the electronic
density at the boron center, therefore making the hydride more nucleophilic.
102 In fact, the hydride activation of catecholborane by a variety of phosphines, including
triphenylphosphine, has been reported in the past and was shown to occur readily at room
temperature.210 Hence, pathway C (Scheme 4-6), involving TS1C, (+10.8 (+38.3) kcal.mol1
) and leading to IM1C (-17.8 (+12.0) kcal.mol-1), is even more energetically favorable than
pathways A and B. The simultaneous activation of the reducing agent and the substrate
drastically contrasts from the classical view of CO2 activation by FLP systems where the
emphasis is on the sole activation of carbon dioxide by both functionalities. Very bulky
groups on the catalyst framework, notably on the Lewis base, may restrict the interaction
with the hydride source, decreasing the reactivity of the system.
Scheme 5-6: Pathway C: hydroboration through simultaneous Lewis base activation of the
borane and Lewis acid activation of CO2. [B] = Bcat.
Consistently with the experimental results where no reaction was observed when heating
catecholborane in the presence of 1,199 no minimum was found on the potential energy
surface for the formation of an adduct between HBcat and the catalyst. However, further
theoretical investigation shows possible rearrangements leading to other plausible
intermediates.
103 Indeed, as represented in Scheme 5-7, HBcat can add to one of the B-O bonds of the catalyst
through TS0D (+19.0 (+35.7) kcal.mol-1) to generate intermediate IM0D (-0.2 (+18.4)
kcal.mol-1), that upon addition of CO2, generates intermediate IM0D’ (+4.7 (+34.4)
kcal.mol-1). The latter can be described as a hydridoborate/boronium bifunctional system
where the binding of CO2 is ensured by the assistance of the catecholboronium fragment
which makes CO2 more prone to nucleophilic attacks. At the same time, the phosphine moiety
acts as an anchor point, allowing the fixation of CO2 with an ideal orientation for hydride
delivery from the hydridoborate fragment. The hydride delivery occurs at TS1D (+12.5
(+43.3) kcal.mol-1), leading to the regeneration of the catalyst by release of HCOOBcat. This
completes an alternate reaction path for the initial step of CO2 reduction (pathway D, Scheme
5-7). Such reactivity is somewhat reminiscent of the catalytic reduction of imines by
boronium hydridoborate ion pairs reported by Crudden et al.267
Scheme 5-7: Pathway D: CO2 reduction through the generation of a boronium /
hydridoborate ion pair. [B] = Bcat.
104 Summing up the results for the reduction of CO2 to HCOOBcat (Figure 5-3), the direct
hydroboration through pathway A can be ruled out. Although the activation of HBcat by one
of the oxygen atoms of 1 (pathway B) or through hydride transfer from HBcat to the catalyst
(pathway D) are plausible, pathway C is the most easily accessible and yields IM1C with a
net energetic gain of 17.8 kcal.mol-1. The catalyzed reduction leads to a decrease of the
activation energy by 23.4 kcal.mol-1 when compared to the uncatalyzed system, making the
reduction kinetically manageable.
Figure 5-3: Important intermediates and transition states for the catalyzed reduction of CO2
to HCOOBcat.
105 5.4.7
Second reduction step: from HCOOBcat to CH2O and derivatives
Before determining the possible role of the catalyst in the second reduction step, the
uncatalyzed hydroboration of HCOOBcat was investigated thoroughly. From HCOOBcat
(IM1), the reduction occurs through a classical four-membered ring transition state, TS2
+15.8 (+45.0) or TS2’ (+14.1 (+42.7) kcal.mol-1), yielding catBOCH2OBcat (IM2, -40.6 (11.6) kcal.mol-1 or formaldehyde (IM2’ -30.8 (-15.6) kcal.mol-1), respectively (Scheme 5-8).
Scheme 5-8: Calculated catalyst-free mechanism for the hydroboration of HCOO[B].
The transition state TS2 has been previously reported in the work of Wang et al., but the
authors have concluded that the energy barrier was too high for the reactions to occur at room
temperature.72 On the other hand, it has been reported that the reaction of catecholborane
with carboxylic acids of the type RCOOH (R=alkyl) at room temperature yields the
corresponding acyloxyboranes (RCOOBcat) as intermediates as well as H2.268 The addition
of two supplementary equivalents of HBcat results in the formation of RCH2OBcat, leading
to the corresponding alcohol after aqueous work-up. In order to verify that the reduction of
HCOOBcat by HBcat was indeed possible, the reaction between HBcat and formic acid
(HCOOH) was studied experimentally.
The addition of formic acid (1 equiv) to a slight excess of catecholborane (3.3 equiv) at room
temperature led to the rapid evolution of dihydrogen. As expected, monitoring of the reaction
using 1H NMR spectroscopy revealed the presence of HCOOBcat as an intermediate species,
but after 90 minutes, the signals attributed to HCOOH and HBcat disappeared completely,
resulting in total conversion to CH3OBcat and catBOBcat. The nature of the products was
confirmed by 11B{1H} NMR spectroscopy and literature precedents.139
106 Repeating the same experiment at 70°C yielded complete conversion after only 15 minutes.
With a computed barrier of +25.1 (+41.2) kcal.mol-1 for the hydroboration of HCOOBcat by
HBcat, it is clear that the reaction occurs much faster than what was previously assumed from
computational results and that HCOOBcat can be reduced without the implication of a
catalyst. (Scheme 5-9)
Scheme 5-9: Experimental verification for the hydroboration of formic acid by
catecholborane. [B] = Bcat.
However, in contrast to the other reported systems for the catalytic hydroboration of carbon
dioxide where formatoborate species were observed during catalysis,71,133 no trace of
HCOOBcat was observed during catalysis.199 Indeed, no HCOOBcat could be detected even
when monitoring at room temperature the reaction between 1 equiv of HBcat relative to
catalyst 1 under 1 atm of CO2. The only new species that was observed in this reaction
mixture was the formaldehyde adduct 2 (Scheme 5-10). This result suggests that contrarily
to all reported systems where the catalysts are of importance in the first reduction step,
catalyst 1 is playing an important role in the reduction of the formatoborate species. Such
result is in line with a previous report showing that 1 catalyzed the hydroboration of
methylformate.199
Scheme 5-10: Attempt to isolate HCOOBcat, leading to the exclusive formation of 2
(IM2C’). [B] = Bcat.
107 In order to reveal how this reduction step is catalyzed, the interaction of HCOOBcat (IM1)
with catalyst 1 was studied computationally. The most favored interaction (IM1C, -17.8
(+12.0) kcal.mol-1) being slightly exothermic by -6.8 kcal.mol-1 with respect to the free
reagents suggests that some of the HCOOBcat molecules will remain bound to the catalyst.
However, each isomer observed in Scheme 5-11 can still be considered as a potential
intermediate for the subsequent hydroboration reaction and as such, one must also take into
account the possible rearrangements of IM1C.
Scheme 5-11: Possible interactions and rearrangements of HCOOBcat with catalyst 1. [B]
= Bcat.
Interestingly, no suitable transition state was found directly from IM1D. This is in line with
the study of Musgrave et al. where it was demonstrated that even if the binding of a phosphine
center to CO2 was beneficial for the fixation of the CO2 molecule on a catalyst, a strong P-C
interaction may actually hinder hydride transfer since the electrophilic site on carbon is
occupied by the free electron pair of phosphorus.108 An interesting situation occurs in the
case of IM1D’, where the phosphine is no longer interacting with the electrophilic carbon
atom of the activated substrate. The activation of a HBcat molecule by the phosphorus
moiety, as previously observed for pathway C, leads to hydride transfer through the most
accessible TS for the reduction of HCOOBcat, TS2C (-13.6 (+32.2) kcal.mol-1), generating
IM2C (-48.0 (-2.2) kcal.mol-1) with a net energetic gain of 34.4 kcal.mol-1 (Scheme 5-12).
108 Scheme 5-12: Suggested pathway for the catalyzed reduction of HCOO[B] involving the
catalyst [B] = Bcat.
Other pathways are also less favored as the hydroboration through four-membered ring
transition states similar to pathway A, either from IM1B or IM1C, and leading to
catBOCH2OBcat type reduction products were found unlikely. Alternately, hydride transfer
through coordination of HBcat to the catechol oxygen atom of IM1D, similar to what was
observed in pathway B (Scheme 5-5) and leading to formaldehyde and catBOBcat, although
accessible, proved to be less favored than TS2C. These results underline the beneficial effect
of double Lewis acid activation while reinforcing the concept of hydride activation by the
Lewis basic center since the catalyzed reduction is 20.9 kcal.mol-1 more favored than the
catalyst free reduction. (Figure 5-4)
109 Figure 5-4: Relative energies of transition states and intermediates for the reduction of
HCOOBcat to CH2O or catBOCH2OBcat.
110 5.4.8
Third reduction step: reducing CH2O and derivatives to CH3OBcat
Although species 2 was previously characterized in solution, it was possible to observe in the
reduction process at 25 °C the formation of a crystalline solid that was identified as the
formaldehyde adduct, thus confirming the presence of this intermediate (Figure 5-5). The
various bond lengths in the crystal structure of 2 are in accordance to the computational data,
thus once more confirming the validity of the method.
Figure 5-5: ORTEP drawing of 2.*
*Anisotropic atomic displacement ellipsoids shown at 50% probability level. Selected bond
lengths [Å] and angles [°]: P(1)-C(1) 1.823(2), C(1)-O(1) 1.402(3), O(1)-B(1) 1.473(3),
C(13)-P(1)-C(1) 104.74(9), C(8)-C(13)-P(1) 117.28(14), C(13)-C(8)-B(1) 125.32(2), C(1)O(1)-B(1)113.67(2).
111 Therefore, IM2C must rearrange to this more stable intermediate. Having a closer look at
IM2C, it is better described as a simple adduct between 1 and catBOCH2OBcat where the
interactions occur through dative P-B and B-O bonds. However, the binding is favored by
only -7.4 (+9.4) kcal.mol-1. As can be observed in Scheme 5-13, catBOCH2OBcat may
rearrange to generate CH2O by releasing catBOBcat (IM2’ -30.8 (-15.6) kcal.mol-1). Such a
rearrangement was also assumed to happen by Wang et al. in their related theoretical study
of a catalytic CO2 hydroboration system.72 The system is then stabilized by the trapping of
formaldehyde by 1 to generate IM2C’ (2). The entropic stabilization due to the release of a
catBOBcat molecule is thought to be the driving force for the formation of this intermediate.
Scheme 5-13: Formation of IM2C’(2) through the rearrangement of catBOCH2OBcat to
CH2O. [B] = Bcat.
It is widely known that aldehydes are readily reduced by hydroboranes, but we were curious
to see if the trapping of formaldehyde by catalyst 1 would hinder or favor the reduction. Since
formaldehyde readily polymerizes to paraformaldehyde and the solubility of 2 in common
NMR solvents is very limited, 4-bromobenzaldehyde 3 was chosen as a model compound.
Monitoring of the reaction between 3 and 1.1 equiv of HBcat showed that the reduction takes
90 minutes to yield complete conversion to the corresponding alkoxyborane 4. Interestingly,
repeating the reaction in identical conditions but in the presence of 2 mol % of 1 led to the
complete conversion in less than 5 minutes, showing that 1 acts as a catalyst for the reduction
of aldehydes to alkoxyboranes.(Scheme 5-14).
112 O
H
CH2O[B]
+
1.1 H[B]
C6D6, r.t.
No cat. 90 min
2 mol% 1, <5 min
Br
Br
3
4
Scheme 5-14: Experimental verification for the catalytic role of 1 in the hydroboration of
4-bromobenzaldehyde by catecholborane. [B] = Bcat
This interesting result prompted us to investigate this final step computationally. From the
formaldehyde adduct IM2C’ (2), the activation of HBcat by the phosphine moiety (similar
to pathway C) leads to TS3C (-36.2 (7.3) kcal.mol-1), yielding the intermediate IM3C (-83.9
(-38.5) kcal.mol-1). Note that IM3C can easily rearrange to IM3 through TS3D (-72.6 (-25.0)
kcal.mol-1), regenerating catalyst 1 and producing CH3OBcat (Scheme 5-15).
Scheme 5-15: Catalyzed reduction of formaldehyde to CH3OBcat, regenerating the
catalyst. [B] = Bcat
113 The final reduction step represents an energetic gain of 25.8 kcal.mol-1. The catalyzed
reduction is 10.2 kcal.mol-1 more favored than the catalyst free reduction. All other pathways
(similar to pathways A or B) are less favorable. (Figure 5-6)
Figure 5-6: Relative energies of transition states and intermediates for the reduction of
CH2O to CH3OBcat.
114 5.4.9
Discussion
As discussed above, the catalyst is essential to lower the energy gap for the reduction of CO2
to HCOOBcat to occur, but also plays a significant role in enhancing the rates of the
subsequent reduction steps. The most favorable species for the reduction in the first step is
the possibility of having the Lewis acidic site of the catalyst binding CO2 while the phosphine
activates the borane to deliver a hydride to the activated electrophilic carbon of carbon
dioxide. Together, these factors lead to a lowering of 23.4 kcal.mol-1 of the energy barrier
when compared to the catalyst free reduction. This pathway puts emphasis on the fact that
the role of the catalyst is to simultaneously activate both of the reagents and not CO2 alone.
The reduction of both HCOOBcat and CH2O was shown to be possible without any
implication from the catalyst and consequently, some of these reductions are expected to
occur catalyst-free in the presence of a large excess of HBcat. However, activation of the
HBcat moiety by the phosphorus center while the substrate is fixed and activated by the
Lewis acidic boron center results in lowering the transition state energies by 20.9 and 10.2
kcal.mol-1 for the hydroboration of HCOOBcat and CH2O respectively. The rapid reduction
of HCOOBcat by the catalyst and in the reaction medium explains why it could not be
observed experimentally. On the other hand, the 15.4 kcal.mol-1 bonding interaction of the
catalyst with formaldehyde rationalizes the fact that this particular adduct can be observed
by NMR spectroscopy during catalysis. As it was found out in this study, 2 even crystallizes
out of the reaction medium at ambient temperature, while everything is soluble at 70 °C. This
aspect might explain the lower activity of this system at room temperature and the high
enhancement of the catalytic turn-overs with a relatively slight increase in temperature. The
entire catalytic process is summarized in Scheme 5-16.
115 Scheme 5-16: Proposed mechanistic pathway including important transition states for the
reduction of CO2 to CH3OBcat by 1. [B] = Bcat.
116 Taking these results into account, the classical FLP approach of using very bulky substituents
may lead to a more difficult activation of substrates. While a strongly Lewis basic phosphine
might bind CO2 and other intermediates more strongly and hinder hydride transfer, it may
also activate the reductant more effectively and increase catalytic activity. However, the use
of a moderate Lewis acid allows the release of the various hydroboration products in the
reaction medium, allowing their liberation from the catalyst. A key aspect of the system is
the presence of both the Lewis acid and base in a single molecule, reducing the entropic cost
of every catalyzed step. Finally, the importance of the oxygen substituents on the boron
center cannot be overlooked as their dynamic nature allows the formation of a number of
isomers and intermediates for a very flexible catalyst.
5.4.10 Conclusion
In conclusion, the full mechanism for the first catalytic hydroboration of carbon dioxide by
a FLP based system 1 was determined. The catalyst was shown to catalyze every step of the
reaction. The findings reported herein offer important insight on the aspects that need to be
considered for the design of ambiphilic catalysts. Current work focuses on preparing new
ambiphilic catalysts by varying the functional groups on phosphorus and boron in order to
achieve maximal catalytic efficiency and broaden the scope of reducing agent to
hydrosilanes. We are hopeful that these findings will inspire unprecedented FLP chemistry
and novel catalytic applications.
117 5.5 Experimental Data:
5.5.1
General experimental
Catecholborane, pinacolborane, formic acid and 4-bromobenzaldehyde were purchased from
Sigma-Aldrich and used as received. 1 was prepared according to the reported procedure.199
Benzene-d6 was purchased from Cambridge Isotopes Laboratories and dried over Na0
followed by a vac transfer. NMR spectra were recorded on a Agilent Technologies NMR
spectrometer at 500 MHz (1H), 125.758 MHz (13C), 160.46 MHz (11B), on a Varian Inova
NMR AS400 spectrometer, at 400.0 MHz (1H), 100.580 MHz (13C), 1H NMR and 13C {1H}
NMR chemical shifts are referenced to residual protons or carbons in deuterated solvent. 11B
{1H} was calibrated using an external reference of B(OMe)3. Multiplicities are reported as
singlet (s), broad singlet (s, br) doublet (d), triplet (t), multiplet (m). Chemical shifts are
reported in ppm. Coupling constants are reported in Hz.
5.5.2
Synthetic methodology
Typical procedure for the reduction of HCOOH
Catecholborane (41.7 mg, 0.35 mmol) and a known quantity (2.0 to 4.0 mg) of
hexamethylbenzene (HMB) (internal standard) were dissolved in benzene-d6 and transferred
to a J-Young NMR tube. Formic acid (4 µL 0,11 mmol) was then added via microsyringe.
Hydrogen evolution was observed and the reaction was then left to react at r.t. or 70 °C for a
set period of time. For the 70 °C reactions, the J-Young NMR tubes were vented after two
minutes of reaction to avoid pressure build up.
Procedure for the stoichiometric reaction of 1 and HBcat with CO2
Catecholborane (3.1 mg, 0.026 mmol) and a known quantity (2.0 to 4.0 mg) of
hexamethylbenzene (HMB) (internal standard) as well as 10.0 mg (0.026 mmol) of 1 were
dissolved in benzene-d6 and transferred to a J-Young NMR tube. The J-Young was frozen in
a -60 °C bath and put under vacuum after which 1 atm of CO2 was added. The reaction was
then followed by NMR spectroscopy.
The experiment was then repeated with 6.2 mg (0.052 mmol) (2 equivalents) of
catecholborane, showing the same results.
118 Procedure for the reduction of 4-bromobenzaldehyde
4-Bromobenzaldehyde (48.7 mg, 0.26 mmol) and a known quantity (2.0 to 5.0 mg) of
hexamethylbenzene (HMB) (internal standard) were dissolved in benzene-d6 and transferred
to a J-Young NMR tube. Catecholborane (34.7 mg, 0.29 mmol) was added at once (t=0) and
the reaction was followed by NMR spectroscopy. For the experiment with catalytic loading,
2.0 mg, 0.053 mmol of 1 was added to the solution with 4-bromobenzaldehyde.
5.5.3
Additional computational information regarding the catalyzed reduction of
CO2 to HCOOBcat
Here we can see that the initial catalyst rearrangement cannot be achieved through activation
of HBcat by the phosphine moiety (TS0D’). However, it is shown that IM0D’’ may
rearrange to the more stable IM0D. On the other hand, this does not change the fact that the
only accessible pathway goes through TS1D which was reported in the manuscript. Also,
transferring the hydride from IM0 through TS0D’’ may be possible, but it only leads to
IM0D’ which more easily accessed through the route presented in the manuscript. (Scheme
5-17)
Scheme 5-17: Calculated pathways for the reduction of CO2 to HCOOBcat. [B] = Bcat.
119 5.5.4
Additional computational information regarding the catalyzed reduction of
HCOOBcat to CH2O and derivatives
Firstly, no accessible transition state was located starting from IM1A, mostly due to the steric
hindrance of the extra Bcat group that rendered the coordination of the second equivalent of
HBcat difficult. From IM2B, two very similar transition states were found through
coordination of catecholborane to the free oxygen atom of the formate moiety TS2A and
TS2A’ both leading to IM2A. From IM1C, a single similar transition state TS2A’’ which
also led to IM2A was found. Once again, we see that the coordination of the substrate to the
catalyst does not significantly reduce the energetic gap for the classical 4-membered
hydroboration but actually hinders it due to entropic considerations. In fact, from IM1 the
gap was found higher for TS2A and TS2A’’ than the calculated gap for the uncatalyzed
reduction of HCOOBcat by HBcat.
Scheme 5-18: Calculated pathways for the reduction of HCOOBcat to CH2O and
derivatives. [B] = Bcat.
120 An interesting feature of IM1E also arises from the coordination of the formate group’s
oxygen to the boron atom. Indeed, as was seen before, the coordination increases the
electronic density on the adjacent catechol oxygen atoms, thus favoring the coordination of
a HBcat molecule. Taken into account together, these factors lead to the isolation of TS2B
generating IM2E which was not included in the manuscript due to lower energy barrier of
TS2C. (Scheme 5-18)
5.5.5
Additional computational information regarding the catalyzed reduction of
CH2O and derivatives to CH3OBcat
Once again, reduction through a 4 membered transition state TS3A to generate IM3 proved
to be energetically difficult. However, even though no TS associated to coordination to site
2 could be found due to geometric constraints, the coordination of formaldehyde allows site
3 of the catalyst to bind the HBcat IM2C’’Reduction can then occur through TS3B this time
generating IM3C (Scheme 5-19).
Scheme 5-19: Calculated pathways for the uncatalyzed reduction of CO2 to CH2O. [B] =
Bcat.
121 5.5.6
Additional computational information regarding the possible involvement of
IM3C in catalysis
Since all transition states were found to be much higher (higher energy gap) than those
reported in the manuscript, they have not been investigated any further. (Scheme 5-20).
Supporting this hypothesis, heating CH3OBcat in the presence of 1 did not result in any
observable reaction by NMR spectroscopy. Once again, the formation of IM3 from IM3C is
due to the entropic stabilization associated with the loss of a CH3OBcat molecule.
Scheme 5-20: Calculated pathways for the catalyzed reduction of CO2 to HCOOBcat from
IM3C. [B] = Bcat.
122 5.5.7
Additional computational information regarding the uncatalyzed reduction of
CO2 to CH3OBcat
Through the rearrangement of HCOOBcat IM1, to a cyclic IM1’’, an unprecendented
transition state TS2’’ occurs through coordination of one HBcat on the oxygen atom of the
catechol group present on HCOOBcat, generating a 6 membered transition state and leading
to IM2’. Two more intuitive and more favorable transition states were also located starting
from HCOOBcat, depending on which oxygen HBcat precoordinated prior to hydride
transfer. TS2 was discussed in the manuscript while the less favored TS2’ produced CH2O
as well as catBOBcat IM2’, (Scheme 5-21).
Scheme 5-21: Proposed mechanism for the catalyst free reduction of carbon dioxide to
CH3OBcat. [B] = Bcat.
No transition state leading to further reduction of IM2’ could be located. From
BcatOCH2OBcat (IM2), the high sp3 character of the CH2 fragment leaves difficult options
for hydride delivery and only one transition state TS3’, with an inaccessible energy barrier
was found.
123 Taking a closer look at the generated formaldehyde and derivatives reveals that
rearrangements are possible in solution. For instance, IM2’ can easily rearrange to the more
stable BcatOCH2OBcat through TSR2. Another transition state TSR, shows that
BcatOCH2OBcat can also rearrange in an intramolecular way to generate BcatOBcat and
formaldehyde. In fact, BcatOCH2OBcat can be viewed as a formaldehyde acetal. Even
though IM2 seems to be the most favorable intermediate, a small concentration of
formaldehyde is expected in solution.
More complete experimental data such as the optimized structured and free energies/
enthalpies of the reported compounds, further crystallographic information, NMR spectra
and additional reaction pathways, the effect of solvent on the atomic energies is available
through the ACS website at http://pubs.acs.org/.
124 Table 5-1: Crystal data and structural cefinements for compound 2.
Compound
2
Formula
C25H20BO3P
Mr (g mol−1)
410.19
Crystal size
0.32 x 0.26 x 0.20 mm3
Crystal color
Colorless
Crystal system
Orthorhombic
Space group
P 2 1 21 21
λ (Å)
0.71073
T (K)
150(2)
a (Å)
9.6270(7)
b (Å)
12.8045(9)
c (Å)
16.2628(1)
α (o)
90
β (o)
90
γ (o)
90
V (Å3)
2004.7(3)
Z
4
ρcal. (g cm−3)
1.359
F(000)
856
μ (mm−1)
0.162
θ for data collection (o)
2.024 to 28.356
Reflection collected
21474
GOOF
1.043
R1(wR2) (I > 2σ(I))
0.0333, (0.831)
R1(wR2) (all data)
0.0363, (0.0855)
125 5.6 Conclusion and perspectives
It is important to note that the principal interest and value of this work does not lie in the
determination of a specific mechanism for a specific transformation. The information
included in the aforementioned manuscript goes much beyond, exploring the fundamental
reactivity patterns that are inherent to all FLP systems. Firstly, the general belief that the
binding of both a Lewis acid and a Lewis base to CO2 would ‘‘activate’’ the molecule was
clearly shown to be a misconception. In fact, the binding of a Lewis base only fills the orbital
in which one aims to transfer the hydride, thus de-activating the CO2 molecule. On the other
hand, a Lewis acid can indeed activate CO2, by making its empty orbital more available to
accept the transfer of a hydride. Nonetheless, as it will be discussed in the next chapter in
more detail, a valid strategy is to use a Lewis base to activate a borane reagent in order to
facilitate hydride transfer.
The use of FLP systems such as a phosphine-borane catalyst combines both the strategies of
Lewis acid activation and reductant activation. The synergistic use of both methods allows
the employment of much less reactive active centers and opens up the way to new reactivity.
Of course, such a concept is not limited to CO2 reduction and hopefully, the conclusions that
can be drawn from this mechanistic study will inspire the design of many FLP catalysts
capable of a variety of transformations.
In parallel with the mechanistic investigation, we had already started optimizing the
substituents at boron and phosphorus in order to maximize efficiency. During the course of
the study, which was carried out as a collaborative effort with the team of Bourissou, we
noticed that the formaldehyde adduct was systematically found intact at the end of the
catalytic processes. This unusual result prompted us to prepare
adducts to gain more insight in the catalytic system. Using
13
13
C labeled formaldehyde
C labelling experiments, we
were able to prove that once the formaldehyde adduct was formed, the same formaldehyde
molecule remained on the catalyst throughout the entire catalytic process. (Scheme 5-22). 269
126 H
C O
H
R2P
CO2 + 3 HBCat
*
BCat
100% 13C labeling
recovered af ter reduction
70°C, C6D6
R = Ph, iPr
H
R2P
* CO2
+ 3 HBCat
H
C O
BCat
70°C, C6D6
H3C OBCat
+
catBOBcat
(1)
no 13C incorporation
no 13C incorporation
af ter reduction
H3C* OBCat
+
catBOBcat
(2)
R = Ph, iPr
Scheme 5-22: Labelling experiments with 13CH2O.*
*Figure reproduced with permission from Declercq, R.; Bouhadir, G.; Bourissou, D.; Légaré, M.-A.;
Courtemanche, M.-A.; Nahi, K. S.; Bouchard, N.; Fontaine, F.-G.; Maron, L. ACS Catal. 2015, 5,
2513–2520. © 2015 American Chemical Society
This very surprising result led us to question the proposed mechanism and thus, mechanistic
pathways starting from the formaldehyde adduct were investigated in more detail using DFT
computations. Three new transition states were located on the potential energy surface which
all involved simultaneous activation of CO2 and the reducing agent. The first transition state
involves activation of the borane by a disconnected catecholate moiety with simultaneous
activation of CO2 through hydrogen bonding. While uncommon, hydrogen bonding was
shown to be critical in CO2 reduction processes before.80,270 In another plausible transition
state, the borane is activated by a disconnected formaldehyde fragment while CO2 is activated
by the borane. Finally, a third transition state with a similar activation energy was located on
the potential energy surface. This time, simultaneous transfer of a boronium/hydridoborate
ion pair was found to decree the reactivity. These new transition states are summarized in
Figure 5-7.
127 Pathway A: H-Bond activati on CO2
Pathw ay B: Lew is acidic activation of CO2
O
C
H
[B]
O
H
O
B
H
[B]
Ph
P
O
CO2
C
Ph
O
IM1: -14.3 (1.6)
O
Ph
P
B
Ph
O
O
O
H
TS1a: -1.1 (28.6)
O
O
B
O
P
IM5: -1.7 (12.8)
Ph
Ph
Ph
O
CO2
C
O
O
CH 2
B
P
O
Ph
IM3: -6.3 (7.3)
Ph
TS1b: 1.9 (29.6)
C
O
CO2
H
B
O
[B]
H[B]
O
CH2
P
O
Ph
Ph
TS1c -1.8 (26.3)
[B]
O
O
CH 2
B
O
P
IM3: -6.3 (7.3)
Ph
Ph
H
O
CO2
B
O
CH 2
P
O
TS1d: 22.3 (38.8)
Figure 5-7: New transition states for CO2 reduction involving a formaldehyde adduct as
catalyst .*
*Reproduced with permission from Declercq, R.; Bouhadir, G.; Bourissou, D.; Légaré, M.-A.;
Courtemanche, M.-A.; Nahi, K. S.; Bouchard, N.; Fontaine, F.-G.; Maron, L. ACS Catal. 2015, 5,
2513–2520. © 2015 American Chemical Society
128 Ph
Pathway D: Competitive hydroboration of f ormaldehyde
H[B]
CH 2
P
O
Pathw ay C: Boronium/ hydridoborate transf er
O
CH 2
B
H [B]
O
H[B]
H
H
C
O
Ph
Ph
All of the transition states were found to be much lower in energy than the reduction of the
bound formaldehyde moiety, explaining why the formaldehyde remains coordinated
throughout catalysis. Furthermore, these transition states were found to be approximately 10
kcal/mol lower than the transition state for CO2 reduction reported in the original mechanistic
study.271 Although there is a good possibility that the initial reduction of CO2 occurs through
the previously proposed mechanism, these results suggest that oxygen bases are more potent
borane activators than phosphines. While counter-intuitive, the fact that only a very weak
Lewis acidic activation of CO2 is necessary to initiate reactivity is underlined by these
findings. Even though these unconventional reaction pathways were overlooked in the
original study, the same mode of action involving simultaneous activation of the reducing
agent and substrate are present, further illustrating the validity of the concepts developed
therein.
129 6 Metal-free catalytic hydrosilylation of CO2 to methanol
6.1 Base catalyzed hydroboration of CO2
It was first shown by Cantat and co-workers that guanidine and amidine superbases were
active catalysts for the hydroboration of CO2. The optimal results of 640 TON and 33 TOF
were obtained using a guanidine derivative (Figure 6-1a), but a major drawback of the system
was that only 9-BBN proved to be able to promote the reduction.239 Shortly after, Stephan
demonstrated that the use of organic superbases was not necessary for CO2 reduction using
a reactive borane such as 9-BBN. In fact, simple phosphines such as P(t-Bu)3 were shown to
be incredibly efficient catalysts, reaching a TON of 5500 and a TOF of 170 h-1 under 5 atm
of CO2 at 60 °C (Figure 6-1b) .240 Since strong hydrides were shown to be able to promote
CO2 reductions,83,272,273 one finds that these base catalyzed reductions most likely occur
through activation of 9-BBN by the Lewis bases, generating hydridoborate-like intermediates
and thus facilitating hydride transfer to CO2.
Figure 6-1: New catalysts for the hydroboration of CO2.
131 We hypothesized that BH3.SMe2 was not efficient in these transformations due to the ability
of the bases to form strong adducts with BH3. In an effort to expand the scope of these
reductions to the use of BH3.SMe2, we tested a number of commercially available bases in
catalytic conditions. Further investigation allowed us to demonstrate that non-nucleophilic
proton-sponge was capable of dissociating BH3 into a very reactive BH2+/BH4- ion pair that
promoted the reduction of CO2 to methoxyboranes in good yield (Figure 6-1c).274 Using a
similar strategy, it was later shown that using NaBH4 as a catalyst could promote the
reduction of CO2 using commercially available BH3.THF solutions,275 although some points
of the study were recently disputed by Cummins.273
6.2 Base catalyzed hydrosilylation of CO2
The development of base catalyzed hydroboration is evocative of the early work of Ying and
co-workers on the NHC catalyzed hydrosilylation of CO2 to methoxysilanes, species that can
be readily converted to methanol. Indeed, it is peculiar that such an important number of
hydroboration systems have been developed in a very brief time scale while 6 years later, no
other metal-free catalyst was reported to effect the hydrosilylation to methanol derivatives.
Still, a number of notable advances were made by Cantat and co-workers in the field of metalfree hydrosilylation of CO2. Even though the products that are formed are not methane,
methanol or any highly energetic material, their work must be underlined as it is intimately
related.
In 2012, the researchers introduced an interesting stratagem to gain access to new CO2
reduction products. The general concept remained very similar to the original report of Ying,
but with the difference of using guanidine superbases instead of NHCs. The twist was simple,
but ingenious: by adding a primary amine to the system, CO2 was converted to a carbamate
prior to reduction (Scheme 6-1).
Scheme 6-1: CO2 capture by dimethylamine.
132 By carrying out the base catalyzed hydrosilylation reaction in the presence of a primary
amine, carbamates were being reduced, generating formamides instead of the previously
reported methoxysilane products (Scheme 6-2a).276 The same strategy was later applied to
the formation of formamidines.277 Also, replacing the silane reagent for 9-BBN allowed the
researchers to promote the base catalyzed reduction all the way to methylamines (Scheme 62b).278
Scheme 6-2: Base catalzyed reduction of CO2 to methylamines.
However, the most interesting contribution was published only two months after the original
hydrosilylation to formamide report. In this work, the authors use the same procedure, but
using NHCs as catalyst instead of guanidines. The interesting change was that the reducing
agent polymethylhydrosiloxane (PMHS), an abundant nontoxic waste by-product of the
silicon industry was utilized as reducing agent (Figure Scheme 6-2c).279
133 While the use of boranes and silanes does not represent an economically viable way of storing
energy using CO2, the use of an industrial waste seemed like an enticing alternative. Even
though such waste would of course, be of limited availability as compared to the yearly
anthropogenic emissions of CO2, the development of a catalytic system to transform this
waste into accessible energy became our next objective.
In the past chapter, the advantages of ambiphilic systems for catalytic applications were
extensively demonstrated. The latest results were suggesting that an interaction as weak as a
simple hydrogen bond could help in the activation of CO2, suggesting that strong Lewis acids
were not essential to promote CO2 transformation. Meanwhile, the effectiveness of strong
organic bases in the activation of organosilane reagents had been conclusively established.
The basic idea that motivated the work in the next chapter was based on combining these
effects to develop an efficient system for the hydrosilylation of CO2 to methoxysilanes using
PMHS as a reducing agent.
Phosphazene superbases have been developed by Schewsinger280,281 and represent a very
peculiar class of organic super-bases. Indeed, unlike other bases, the coordination of a proton
or Lewis acid to the basic center results in the generation of a phosphonium center right next
to basic center, conferring a certain ambiphilic character to these molecules (Scheme 6-3).
R
R
P
R
R
H+
N
R
R
P
R
H
N
R
R= alkyl or aryl
Scheme 6-3: Protonation of a phosphazene superbase.
The original hypothesis that motivated this work was that upon coordination of the silane
reagent, the phosphonium center could act as a Lewis acid, serving to activate the carbon
dioxide fragment and orient it towards the activated hydride (Scheme 6-4).
134 Scheme 6-4: Proposed strategy for the hydrosilylation of CO2 using a phosphazene
catalyst.
In the following work, the reactivity of commercially available phosphazene bases in the
reductive hydrosilylation of CO2 was studied. It was found that the ambiphilic character did
promote reactions with carbon dioxide, but not quite as anticipated. Nonetheless, a thorough
reactivity study was carried out, leading to unexpected, yet consequential discoveries which
are detailed in the following publication.
135 6.3 Research article: Phosphazenes: Efficient Organocatalysts for the
Catalytic Hydrosilylation of Carbon Dioxide
6.3.1
Résumé
Les phosphazènes, super bases organiques, sont des organocatalyseurs efficaces pour
l’hydrosilylation du dioxyde de carbone sans métal de transition. Ils réagissent avec le CO2
pour former les oxydes de phosphine respectives, mais en présence d’hydrosilanes, le CO2
peut être sélectivement réduit en formates silylés, produits qui peuvent être à leur tour réduits
en méthoxysilanes par simple addition d’une charge supplémentaire de silane. Le système
peut générer jusqu’à 759 tours catalytiques à une fréquence maximale de 32 tours
catalytiques par heure. Il est aussi démontré que surprenamment, le solvant N,NDiméthylformamide peut réduire le CO2 en un mélange de formates silylés, d’acétals et de
méthoxysilanes en absence de catalyseur.
6.3.2
Abstract
Phosphazene superbases are efficient organocatalysts for the metal-free catalytic
hydrosilylation of carbon dioxide. They react with CO2 to form the respective phosphine
oxides, but in the presence of hydrosilanes, CO2 can be selectively reduced to silyl formates,
which can in turn be reduced to methoxysilanes by addition of an extra loading of silanes.
Activities reach a TOF of 32 h-1 with a TON of 759. It is also shown that unexpectedly, N,Ndimethylformamide can reduce CO2 to a mixture of silyl formates, acetals and methoxides in
the absence of any catalyst.
136 6.3.3
Introduction
The alarming rate at which the concentration of carbon dioxide is increasing in the
atmosphere is pushing the scientific community to find solutions in order to limit these
emissions and remove this greenhouse gas from the atmosphere.8 While many advances have
been made in the capture and storage of CO2, the use of carbon dioxide as a feedstock for the
synthesis of valuable products could provide an important incentive for CO2 recycling.14 Of
notorious importance is the large scale production of energy vectors such as methanol from
CO2, which would be one very efficient way to sequester considerable amounts of carbon
dioxide. This concept is promoted by Nobel laureate George A. Olah as the methanol
economy.20 While some heterogeneous systems have been reported for the CO2 reduction to
methanol, the use of transition metal catalyzed homogeneous systems have also been
reported for the hydroboration,71,133,229,233,282 the hydrosilylation70,90,130,131,173,235,246,283–287and
to a lesser extent the hydrogenation125,129,242,243,288,289 of carbon dioxide. However, the search
for new environmentally friendly and efficient systems remains a very active field of study.
During the past few years, metal-free catalysis has been shown to be a viable alternative to
the transition-metal catalyzed reduction of CO2. In fact, in some cases organocatalysts have
even been shown to exhibit catalytic activities surpassing that of transition-metal based
systems.250,290 The advent of Frustrated Lewis Pairs (FLP) chemistry,99,100 notably for CO2
capture,103,104,120 has led to the development of efficient organocatalytic systems for the
hydroboration of carbon dioxide to methoxide derivatives that can be readily hydrolysed to
methanol.106,238,271It was later shown that phosphines,240 strong nitrogen bases,239,274,278,291and
even BH4- could act as catalysts for the hydroboration of CO2.275 In a seminal report, Ying
and co-workers have reported that N-heterocyclic carbenes (NHCs) could act as catalysts for
the hydrosilylation of carbon dioxide to methoxysilanes with TOF and TON reaching 25.5 h1
and 1840, respectively.138
137 Since then, a number of systems have emerged for the reduction of CO2 using hydrosilanes
to generate value added products such as formamides,279 formamidines,276 methylamines,277
acetals132 and methane.104 While base-activation of hydrosilanes has been used for the
hydrosilylation of aldehydes, ketones and esters,292–296 to the best of our knowledge, Ying’s
report of NHC-catalyzed system remains the only metal-free system for the hydrosilylation
of CO2 to methanol or formic acid derivatives. Phosphazene derivatives of the general
structure (R2N)3P=N-R have been developed by Schwesinger as very strong organic
bases.280,281 They are well established as an alternative to NHC organocatalysts for the ring
opening polymerization of cyclic esters.297 Surprisingly, the use of those bases as catalysts
for CO2 reduction was never reported. In an effort to broaden the scope of organocatalyzed
reduction of CO2, we herein report that commercially available phosphazene superbases can
catalyze the reductive hydrosilylation of CO2.
138 6.3.4
Results and discussion
Firstly, as other strong Lewis bases such as NHCs (N-heterocyclic carbenes) are known to
bind carbon dioxide, the reaction between various commercially available phosphazene bases
(Scheme 1, 1-3) and carbon dioxide was investigated in order to verify if the isolation of a
CO2 adduct was possible. When 0.03 mmol of 1 was dissolved in benzene-d6 and was
exposed to 1 atm of CO2, a small amount of a new signal resonating at 21.3 ppm was observed
by 31P{1H} spectroscopy. Complete conversion to the new species was achieved after 8 hours
with only one notable change in the 1H NMR spectra. Indeed, the resonance attributed to the
t-Bu groups was shifted from 1.5 ppm to 0.9 ppm, with the loss of the 4JP-H coupling (1.2 Hz).
Upon evacuation of the volatiles in vacuo and re-dissolution in benzene-d6, the 31P{1H} NMR
spectra remained unchanged showing a singlet at 21.3 ppm while the signal attributed to the
t-Bu protons was no longer present in the 1H NMR spectrum.
Based on these experimental evidences, it was hypothesized that the 21.3 ppm species was
not a simple adduct between 1 and CO2, but was the product of a transformation of the
phosphazene with the elimination of the t-Bu group. Comparison of the 1H and
13
C {1H}
NMR signals for the new t-Bu containing species and that of a commercial sample confirmed
that the solution contained t-Bu isocyanate. Repeating the experiment at a larger scale
allowed the isolation of the phosphine oxide 4 in quantitative yield. As noted in Scheme 65, a plausible mechanism would first involve the formation of an adduct between 1 and CO2,
leading to the formation of a Lewis acidic phosphonium center that can interact with the
oxygen atom of the activated CO2 molecule. tBuNCO and 4 are then formed in a concerted
fashion, through a 2+2 rearrangement, reminiscent of the Wittig reaction mechanism.298 A
similar rearrangement has been observed previously in a titanium complex leading to the
generation of carbodiimides.299 The less bulky phosphazene 2 undergoes the same
rearrangement in the presence of carbon dioxide within 15 minutes, while the more bulky 3
also undergoes a similar rearrangement, releasing ethylisocyanate after 12 hours under 1 atm
of CO2 at 60 °C. It is interesting to note that in the case of 3, one phosphinimine moiety
remains intact.
139 Scheme 6-5: Rearrangement of phosphazenes in the presence of CO2 with proposed
intermediate.
Since the rearrangement with phosphazenes 1 and 3 is slow, we were curious to see if they
could be active catalysts in the hydrosilylation of carbon dioxide. In an initial experiment, 1
was dissolved in benzene-d6 in the presence of 5 equiv of Ph2SiH2. The sealed J-Young NMR
tube was then exposed to ca. 3 atm of CO2 and heated at 80 °C for 4 hours. Interestingly, 1H
NMR spectroscopy revealed that the hydrosilane was completely consumed, giving rise to
set of resonances in the δ ~ 3.8, ~ 5.8, and ~ 8.6 regions, indicative of CO2 reduction to
methoxy-, acetal- and formylsilanes. In order to ascertain that the reduction products did
originate from CO2 reduction, the reaction was carried out using
13
CO2 in N,N-
dimethylformamide since polar and coordinating solvents have been shown to accelerate the
reduction process.138 In a typical experiment, ca 3 atm of 13CO2 was condensed in a J-Young
NMR tube containing a 19.0 mM solution of 1 in presence of Ph2SiH2 (2.5 mol % based on
Si-H) in DMF-d7. After 30 minutes at room temperature, the signals belonging to silyl
formates (δ = 8.6, 1JC-H = 231.8 Hz), silyl acetals (δ = 5.8, 1JC-H = 171.8 Hz) and silyl
methoxides (δ = 3.8, 1JC-H = 144.2 Hz) could be identified using 1H NMR spectroscopy.138
The C-H correlations were confirmed using HSQC experiments.
Upon reaction completion, the sample was degassed by three subsequent freeze pump thaw
cycles to ensure complete removal of 13CO2 from the reaction medium after which a second
140 loading of Ph2SiH2 (2.5 mol % based on Si-H) was added. After 36 hours, the silyl formates
had completely converted to methoxysilane derivatives with 1H chemical shifts ranging
between 3.0-3.8 ppm.
Monitoring the same catalytic reaction using 31P{1H} NMR spectroscopy revealed that after
30 minutes, the signal of 1 (δ = -2.1) completely disappeared to form one new species
resonating at 22.9 as well as 4 in a 6:1 ratio. It is important to note that both signals remain
unchanged throughout the entire catalytic process. Interestingly, the signal at 22.9 ppm
slowly reverted back to the signal of 1 at -2.1 ppm upon additional loading of silane in the
absence of 13CO2. Based on the aforementioned spectroscopic observations, it seems that the
signal at 22.9 ppm arises from an interaction between 1 and silyl formates. The absence of
coupling with the 13C nucleus, points to an N-Si interaction rather than an N-C interaction.
In a blank experiment, heating 1 at 80 °C for 1 hour in the presence of an equimolar amount
of Ph2SiH2 in benzene-d6 yielded no change in the NMR spectra. Similarly, exposing 4 to ca
3 atm of CO2 in the presence of Ph2SiH2 in benzene-d6 did not yield any reaction, even in the
presence of t-Bu-isocyanate, suggesting that neither 4 nor the isocyanate has a role in the
reduction of CO2.
With a better understanding of the reaction, the reduction of 3 atm of CO2 with a 13 mM
solution of 1 in DMF using 40 equiv of Ph2SiH2 as a reductant was monitored using 1H NMR
spectroscopy. After only 60 minutes, 60% of the silane was consumed (TOF = 41 h-1, TON
= 48) while 90% was converted after 5 hours (TOF= 14.4 h-1, TON = 72). Interestingly, it
was observed that under these conditions, silyl formates were the predominant reduction
products, contrasting with the reported reactivity of NHC catalysts which formed a majority
of methoxysilanes.138
141 The preferential formation of silylformates over methoxy species suggests that the first
reduction of CO2 to generate formates occurs faster than the subsequent reduction steps which
will generate acetal and methoxysilane derivatives. Performing the catalytic reduction (1.25
mol% 1, Ph2SiH2) under 5 atm of CO2 led almost exclusively to the formation of
silylformates in a 97% ratio when compared to silylacetals (2%) and methoxysilanes (1%)
after only two hours. The remaining CO2 was then removed in vacuo and an additional
loading of Ph2SiH2 was added, leading to complete conversion of the silyl formates into
methoxysilane derivatives after 36 h (Scheme 6-6). Such selectivity in the formation of either
silylformates or silylmethoxydes is of notable interest.
Scheme 6-6: Product ratio after the catalytic reduction under 5 atm of CO2 and after
addition of an extra loading of silane in the absence of CO2. Yield based on Si-H.
The scope and efficiency of various silanes, solvents and phosphazene catalysts were probed
using a variety of conditions (Table 6-1). As can be observed in entry 3, changing the silane
from Ph2SiH2 to the less bulky PhSiH3 resulted in the reaction sustaining 41 turnovers over
4 h. On the other hand, no catalytic activity was observed using the bulkier Et3SiH, even after
24 hours under 3 atm of CO2 (entry 4). The alkoxysilane (EtO)3SiH did show modest activity,
completing 41 turnovers after 24h under 3 atm of CO2 (entry 5), while
PMHS
(polymethylhydrosiloxane) decomposed readily in the presence of 1 in DMF to generate a
gas which was assigned to be MeSiH3 as well as an insoluble white solid indicative of the
formation of polysiloxanes. Indeed, base catalyzed decomposition of PMHS is well
documented.300 Changing the solvent to another polar, coordinating solvent such as
acetonitrile gave activities comparable to those obtained in DMF, as can be seen in entry 6
with TOF of 16 h-1. On the other hand, the use of non-coordinating solvents such as toluene
or DCM resulted in a dramatic decrease in reactivity (entries 7-8).
142 Although 2 was shown to rapidly rearrange to 5 in the presence of carbon dioxide, it was
found to exhibit catalytic activity in the presence of Ph2SiH2, reaching 31 turnovers after 4
hours of reaction under 3 atm of CO2 (entry 9). The bulkier phosphazene 3 did not surpass
the activity of 1 but still managed 93 TON in 4h for a TOF of 23 h-1 (entry 10). Since a
phosphinimine moiety remains intact in the transformation of 3 to 6, the activity of 6 was
also investigated, but no significant reaction was observed (entry 11).
In order to mimic the activity presented in the reduction of CO2 with NHC catalysts, the use
of very low catalytic loading was carried out. With a catalytic loading of 0.01 mol%,
complete conversion of the silanes was achieved after 48 hours. These results would
represent a TOF of 206 h-1 and a TON of 9900, which is quite unexpected since one would
not expect such an increase in turn-over frequency upon dilution (entry 13). To our surprise,
repeating the experiment with the same parameters while omitting the catalyst showed that
21% of the silane was consumed after 4h and as much as 82% was consumed after 24h
(entries 14-15). It is important to note that no silane was consumed when the experiment was
repeated without CO2, confirming that the products do not originate from DMF reduction.
While it was previously observed that the choice of solvent was an important factor in silane
mediated CO2 reductions,138,301 it is surprising that such an important reduction of CO2 by
the solvent alone was not reported.302,iii When comparing the conversion with and without
catalyst (entries 1 and 14), it can be shown that the phosphazenes are nevertheless efficient
catalysts for the hydrosilylation of CO2. In acetonitrile or THF, the presence of a catalyst was
required to induce the reduction of CO2. In acetonitrile, the reduction catalyzed by 1 reached
a TON of 759 after 24 h, representing a TOF of 32 h-1. The latter result shows that
phosphazene remains robust catalyst over a long period of time. A large scale reaction was
carried out using 4.0 g of Ph2SiH2 in the presence of 1 (1.25 mol%) in CH3CN and from
which methanol was generated with 69 % yield upon hydrolysis.
iii
DMF was shown to affect the methylation of amines in the presence of Ph2SiH2 using CO2 as the carbon source: S. Das, F. D. Bobbink, G. Laurenczy and P. J. Dyson Angew. Chem. Int. Ed., 2014, 53, 12876‐12879. 143 The reduction mechanism for this system is most likely similar to the NHC catalyzed CO2
hydrosilylation mechanism which was studied computationally and was proposed to occur
via activation of the hydrosilane rather than through activation of CO2.92 While an
experimental/theoretical report by Ying contested the study, 303 the important solvent effects
were not studied in detail in any of the two reports. Another plausible pathway would involve
the formation of a 6-coordinate hyper-coordinate silicon center bearing a strongly
nucleophilic hydride. In fact, C=O bond reduction by hyper-coordinate silicon species is well
documented.304,305 It was even shown by Kobayashi that the use of DMF could promote the
formation of hyper-coordinate silicon species.306,307 Further investigation of the reaction
mechanism is currently ongoing in our research group.
Table 6-1: Catalytic hydrosilylation of CO2 by phosphazene bases.
#
Catalyst
([ ] μmol/L)
Silane
([ ] mmol/L ; equiv)c
Solvent.
P
T
Yield
(atm) (h) (% Si-H) TOFc TONc
1
2
3
4
5
6
7
8
9
10
11
12d
13d
14
15
16
1 (12.5)
1 (12.5)
1 (12.5)
1 (25)
1 (25)
1 (12.5)
1 (12.5)
1 (12.5)
2 (12.5)
3 (12.5)
6 (12.5)
1 (0.5)
1 (0.1)
1 (0.1)
Ph2SiH2 (1 ; 80)
Ph2SiH2 (1 ; 80)
PhSiH3 (1; 80)
Et3SiH (1 ; 40)
(EtO)3SiH (1 ; 40)
Ph2SiH2 (1 ; 80)
Ph2SiH2 (1 ; 80)
Ph2SiH2 (1 ; 80)
Ph2SiH2 (1 ; 80)
Ph2SiH2 (1 ; 80)
Ph2SiH2 (1 ; 80)
Ph2SiH2 (1 ; 2,000)
Ph2SiH2 (1; 10,000)
Ph2SiH2 (1; -)
Ph2SiH2 (1 ; -)
Ph2SiH2 (1; 1,000)
DMF
DMF
DMF
DMF
DMF
CH3CN
Toluene
DCM
DMF
DMF
DMF
DMF
DMF
DMF
DMF
CH3CN
1
5
1
3
3
1
1
1
3
3
1
1
1
1
1
1
4
2
4
24
24
4
4
4
4
4
1
24
48
4
24
24
96
95
42
0
41
79
14
1
39
93
1
93
98
21
82
76
19
38
8
0
1
16
3
0
8
19
0
N/A
N/A
N/A
N/A
32
77
76
33
0
16
63
11
0
31
74
0
N/A
N/A
N/A
N/A
759
a) The reactions were carried out at ambient temperature and the yields were determined using 1H
NMR integration with cyclohexane as an internal standard. b) Distribution of products observed. c)
Based on the number of Si-H transferred. d) Under these conditions, the DMF will be responsible
for the reduction
144 6.3.5
Conclusions
In summary, it was shown that although phosphazene superbases can rearrange to their
respective oxide in an unusual way, they represent an efficient new class of organocatalysts
for the reduction of carbon dioxide. One can easily control selectivity to methoxysilanes or
silyl formates by controlling the reaction conditions. Using a very simple protocol TOF
reaching 32 h-1 and TON reaching 759 were observed. Furthermore, it was shown that the
role of DMF in such a reduction process is of critical importance and its role as a catalyst
should be considered when evaluating the catalytic efficiency of organocatalysts for CO2
reduction. Current studies focus on the use of phosphazene bases for other transformations
involving the formation of value added products from carbon dioxide reductions and will be
reported in due course.
145 6.4 Experimental
6.4.1
General experimental:
While no particular precautions were made to avoid contact with air for most of the catalytic
reactions, the phosphazene bases and the preparation of the catalytic experiments have been
handled under an inert atmosphere of dinitrogen. Reactions were carried either in a sealed JYoung NMR tube, in which case NMR conversions are indicated, or in standard oven dried
vials or schlenk vessels. Benzene-d6 was purified by vacuum distillation from Na/K alloy,
bone dry CO2 was purchased from Praxair and used as received. 13CO2 (99% isotope label)
was purchased from Cambridge Isotope Laboratories and stored over CaCl2. Phosphazene
bases, solvents, DMF-d7 and silanes were purchased from Sigma Aldrich and used as
received without further purification.
NMR spectra were recorded on Agilent Technologies NMR spectrometer at 500 MHz (1H),
125.758 MHz (13C), 202.456 MHz (31P) 160.46 MHz, a Varian Inova NMR AS400
spectrometer, at 400.0 MHz (1H), 100.580 MHz (13C), 161.923 MHz (31P), or on a Bruker
NMR AC-300 at 300MHz (1H), 75.435 MHz (13C), 121.442 MHz (31P). 1H NMR and
13
C{1H} NMR chemical shifts are referenced to residual protons in deuterated solvent.
Multiplicities are reported as singlet (s), broad singlet (s, br) doublet (d), triplet (t), multiplet
(m). Chemical shifts are reported in ppm. Coupling constants are reported in Hz. gHSQC
experiments were performed in order to confirm C-H correlations. GC-MS characterization
was carried out using a Thermo Scientific trace GC ultra coupled with a ITQ 900 mass
spectrometer using electronic impact (EI) ionization. WARNING: Condensation of high
pressure of CO2 might lead to an explosion of the glassware. Care should be taken.
146 6.4.2
Initial test experiment:
7.8 mg (0.028 mmol) of 1 along with 26.2 mg (0.142 mmol) of Ph2SiH2 were dissolved in
benzene-d6 and introduced into a gas tight J-Young NMR tube. CO2 was added by freezing
the solution in a liquid nitrogen bath and leaving the sample under vacuum for ca 3 minutes.
The liquid nitrogen bath was removed and the sample was left to warm up for 20 seconds
after which CO2 was introduced into the tube. The sample was then heated at 80°C in an oil
bath after which it was analyzed by NMR spectroscopy.
6.4.3
Rearrangement of 1 to 4
200 mg of 1 (0.73 mmol) were placed in a sealable schlenk tube along with 1mL of toluene.
The solution was frozen in liquid nitrogen and CO2 was introduced. The reaction was heated
at 80 °C for 12 hours after which the volatiles were removed in vacuo, yielding pure 4 as
colorless oil, 150mg, Yield = 94% (crude) (Scheme 6-7).
Scheme 6-7: Preparation of 4 from 1.
Figure 6-2: Assignment of spectra for 4.
4:1H NMR 500MHZ: δ 3.07 (m, 4H, Hd); 2.9 (m, 2H, Hd or Hg); 2.48 (m, 2H, Hd or Hg); 2.39
(d, 3JP-H:11.0Hz, 6H, Hc); 1.60 (m, 1H, He or Hf);1.20 (m, 1H, He or Hf); 0.99 (t, 3JH-H=7.1
Hz, 6H, Ha).13C {1H} (126 MHz): δ 51.0 (s, 2C, C3); 39.7 (d, 2JP-C=5.3Hz, 2C, C2); 35.5 (s,
2C, C4); 26.3 (s, 1C, C5); 14.4 (s, 2C, C1). 31P {1H} (202MHz): δ 21.8 (s, 1P).
MS (EI):
147 6.4.4
Rearrangement of 2 to 5
Scheme 6-8: Preparation of 5 from 2.
6.5 mg (0.021 mmol) of 2 were dissolved in 0.4 mL of benzene-d6 and exposed to CO2 using
the same procedure that was described above. The compound was left to react for 15 minutes
after which the volatiles were removed in vacuo and the sample redissolved in benzene-d6.
Yield is quantitative (by NMR).
Scheme 6-9: Assignment of signals for 5.
5: 1H NMR 500MHZ: δ 3.11 (m, 12H, Ha); 1.5 (m, 12H, Hb).13C {1H} (126 MHz): δ 46.7
(d, 12C, 2JP-C=5.3Hz ); 26.7 (d, 12C, 3JP-C=7.6 Hz ). 31P {1H} (202MHz): δ 14.9 (s, 1P).
MS (EI):
N
P
N
N
O
calcd:187.10
found:187.24
148 calcd:70.11
found:70.12
6.4.5
Rearrangement of 3 to 6
Scheme 6-10: Preparation of 6 from 3.
Scheme 6-11: Assignment of signals for 6.
208 mg of 3 (0.61 mmol) was introduced in a vial (with no cap) which was deposited at the
bottom of a Schlenk tube. The Schlenk tube was put under a constant flow of carbon dioxide
in an oil bath maintained at 60 °C overnight. After 12 hours of reaction, the volatiles were
removed in vacuo, yielding 189 mg of crude 6 (Yield =99%).
6: 1H NMR 500MHZ: δ 2.8 (d, 3JP-H=10.8 Hz, 18H, Hb); 2.42 (d, 3JP-H=9.4 Hz, 12H, Hb).13C
{1H} (126 MHz): δ 38.2 (bs, 4C,C1); 37.1 (bs, 6C, C2 ). 31P{1H} (202MHz): δ 22.5 (d, 2JPP=43.0
Hz 1P, P1); 12.4 (d, 2JP-P=43.0 Hz 1P, P2).
MS (EI): (6 –NMe2)
149 6.4.6
Catalytic reduction using 13CO2.
2.1 mg of 1 (7.6 µmol) as well as 28.7 mg (0.157 mmol) of Ph2SiH2 were dissolved in 0.4 ml
of DMF-d7 after which 2 µL of C6H12 (internal standard) were added. The solution was
transferred to a J-Young NMR tube which was frozen in a liquid nitrogen bath. The J-Young
was left under vacuum for 30 minutes after which ~3 atm of 13CO2 were condensed in the
tube. The solution was left to warm at r.t. naturally and analyzed by 1H,
13
31
P{1H}, and
C{1H} NMR spectroscopy at regular intervals.
6.4.7
Following the reaction in DMF, with various loadings of silane:
Preparation of the mother liquor: a 12.5 μmol/ml solution has been prepared by dissolving
17.0 mg of 1 in 5.0 ml (4.74 g) of DMF. 0.4 ml of the mother liquor (5.2 µmol of 1) was
introduced in a J-Young NMR tube along with 38.4 mg (0.1 mmol) (Test A), 96.0 mg (0.52
mmol) (Test B) of Ph2SiH2. Then, CO2 was added following the procedure described in S1.
The samples were left to warm up at r.t. for 10 minutes before starting the acquisition at r.t.
The reaction was followed by 1H NMR spectroscopy and the intensity of the integration
plotted agains time for both reactions in the graphs presented below: (Figure 6-3 and 6-4)
100
90
80
Product ratio (%)
70
60
%Si‐H
50
%HCOO
40
%OCH2O
30
%OCH3
20
10
0
0
100
200
300
400
500
600
Time (min)
Figure 6-3 Product ratio over time for Test A
150 100
90
Product ratio (%)
80
70
60
%Si‐H
50
%HCOO
40
%OCH2O
30
%OCH3
20
10
0
0
100
200
300
400
500
600
700
Time (min)
Figure 6-4 Product ratio over time for Test B
In this test, the CO2 was depleted due to the presence of a large excess of silane, explaining
why the methoxide derivatives appear as the predominant product.
6.4.8
Reaction under 5 atm. of CO2
0.4 ml of the mother liquor was introduced in a 1.0 mL vial along with 38.4 mg of Ph2SiH2
(0.1 mmol). ca 0.5 mL of benzene-d6 was added to the solution as well as a magnetic stir bar.
The vial was placed inside a 170 mL Fischer porter vessel. The vessel was frozen in a liquid
nitrogen bath and put under vacuum for 60 minutes, after which 5 atm was condensed in the
fischer porter vase. The liquid nitrogen bath was removed and replaced with a water bath at
r.t. After 2 hours of reaction, the solution was analyzed by 1H NMR spectroscopy.
151 6.4.9
General procedures for catalytic tests:
Mother liquor preparation:
To simplify some manipulations, a 12.5 μmol/ml solution has been prepared by dissolving
17.0 mg of 1 in 5.0 ml (4.74 g) of DMF. It will be referred as mother liquor in the rest of this
section.
Various loading of Ph2SiH2 using 1 as catalyst (entry 1,3 and 4 of Table 6-1 in main text):
Entry 1: 17 mg (62.5 μmol) of 1 and 460 mg (2.5 mmol) of Ph2SiH2 have been weighed in
a vial and 5 ml of DMF have been added. The solution was then transferred to a Schlenk
flask. CO2 was added by freezing in a liquid nitrogen bath and leaving the sample under
vacuum for ca 3 minutes. The liquid nitrogen bath was removed and sample left to warm up
for ca 20 seconds after which CO2 was introduced in the Schlenk flask. Sample was then
allowed to warmup to r,t. in ca 5 min by stirring in a r.t. water bath and then left with stirring
at r.t. under CO2 flow. Aliquots were taken at different time interval, dissolved in benzened6, and analysed by 1H NMR spectroscopy. We also performed blanks, one without catalyst
and one without catalyst and CO2. The results are discussed in main text.
Entry 3: 400 μL of mother liquor (5 μmol of 1) and 922 mg (5 mmol) of Ph2SiH2 have been
diluted with 10 ml of DMF and the solution was then transferred to a Schlenk flask. CO2 has
been added in the same manner as for Entry 1. Aliquots were taken at different time interval,
dissolved in benzene-d6, and analysed by 1H NMR spectroscopy.
Entry 4: 80 μL of mother liquor (1 μmol of 1) and 922 mg (5 mmol) of Ph2SiH2 have been
diluted with 10 ml of DMF and the solution was then transferred to a Schlenk flask. CO2 has
been added in the same manner as for Entry 1. Aliquots were taken at different time interval,
dissolved in benzene-d6, and analysed by 1H NMR spectroscopy.
152 Solvent screening (entry 6-8 of Table 6-1 in main text):
3.4 mg (12.5 μmol) of 1 and 92.2 mg of Ph2SiH2 (0.5 mmol) have been weighted in vial and
1 ml of various solvents (DMF, CH3CN or Toluene) have been added. Then, CO2 was
bubbled through the solution for approximately 15 seconds after what the solutions were left
with stirring at r.t.under one atmosphere of CO2 using balloons. After 4 hours of reaction,
benzene-d6 was added and the solutions were analyzed by 1H NMR spectroscopy.
In the case of CH3CN good conversion was observed so more complete tests have been run
with the same method as for DMF (see Entry 1). The first one (80 equiv) using 17 mg (62.5
μmol) of 1, 460 mg (2.5 mmol) of Ph2SiH2 and 5 ml of CH3CN and a second one (1000 eq)
using 17 mg (62.5 μmol) of 1, 460 mg (2.5 mmol) of Ph2SiH2 and 5 ml of CH3CN. It is the
results of those tests that are reported in Table 6-1 in main text.
We also performed blanks, one without catalyst and one without catalyst and CO2, both
showing no conversion of Ph2SiH2 after 24h.
153 Silane screening (entry 3-5 of Table 6-1 in main text):
Silanes have been screened in a similar manner using the mother liquor. 0.1 mmol of Si-H
(80 eq) have been weighted in a vial and 1ml of mother liquor, (12.5 μmol of 1) have been
added. Then, CO2 was bubbled through the solution for approximately 15 seconds after what
the solutions were left with stirring under one atmosphere of CO2 using balloons. After 4
hours of reaction, benzene-d6 was added and the solutions were analyzed by 1HNMR
spectroscopy.
Since the conversion using (EtO)3SiH and Et3SiH were very low after only 4h, we perform
another experiment with higher 1 loading in a J-Young NMR tube. Briefly, 2.8 mg of 1 (10.2
μmol) and 400 μmol of silane (65.7 mg for (EtO)3SiH and 46.5 mg for Et3SiH) have been
weighted in vials. Then 0.4 ml of DMF and 5-6 drops of benzene-d6 have been added. The
solution was introduced into a gas tight J-Young NMR tube. CO2 was added by freezing in
a liquid nitrogen bath and leaving the sample under vacuum for ca 3 minutes. The liquid
nitrogen bath was removed and sample left to warm up for ca 20 seconds after which ca
3 atm of CO2 was introduced in the tube. The reactions were left at r.t. and monitored at
various time intervals by 1H NMR spectroscopy. The results obtained after 24h are reported
in Table 6-1 of the article main text.
Other phophazenes (entry 9-11 of Table 6-1 in main text):
Catalytic activities of other phosphazenes (structure 2, 3 and 6) have been tested in J-Young
NMR tube with Ph2SiH2. 5 μmol of phosphazene (1.6 mg for 2, 1.7 mg for 3 and, 1.6 mg for
6) and 400 μmol of Si-H (36.9 mg of Ph2SiH2) have been weighted in vials. Then 0.4 ml of
DMF and 5-6 drops of benzene-d6 have been added The solution was introduced into a gas
tight J-Young NMR tube. CO2 was added by freezing in a liquid nitrogen bath and leaving
the sample under vacuum for ca 3 minutes. The liquid nitrogen bath was removed and sample
left to warm up for ca 20 seconds after which ca 3 atm of CO2 was introduced in the tube.
The reactions were left at r.t. and monitored at various time intervals by 1HNMR
spectroscopy.
154 Big scale reaction in CH3CN:
50 mg (0.18 mmol) of 1 and 1.33g of Ph2SiH2 (7.2 mmol) have been weighed in vial and
transferred to a 100 mL schlenk tube. The compounds were dissolved in CH3CN and stirred
at r.t. for 24 hours under an atmosphere of CO2. Then, the sample was put under a constant
flow of dinitrogen after which 2,67 g of Ph2SiH2 (14.4 mmol) was added. The reaction was
left to stir at r.t. for 48 hours. A 5mL aliquot was taken from the reaction mixture to which 1
mL of a 10% solution of NaOH in water was added. The resulting mixture was filtered and
an internal standard of THF was added and the yield of methanol was determined by GCFID spectroscopy. GC spectra were recorded on a Hewlett Packard GC-FID 6890 Series with
an HP-5 (Crosslinked 5% PHME siloxane) column, using an isotherm at 40°C. Injection
volumes were 1 µL. A calibration plot was obtained using three aqueous solutions of THF
(1% v/v) containing methanol (0.5, 1 and 2 % v/v) to ensure linearity of the THF/MeOH
signal.Yield = 69%.
More detailed experimental data such as NMR spectra are available for free through the RSC
website: http://pubs.rsc.org/.
155 6.5 Conclusions and perspectives
Interestingly enough, the original hypothesis that phosphazene bases should react as
ambiphilic species was validated. Unfortunately, the proposed reactivity did not take place,
but an interesting rearrangement of the bases into their respective phosphine oxides was
reported instead. It came to our attention after the publication of the article that the
rearrangement of phosphazene bases into phosphine oxide in the presence of CO2 was
reported before.136,308 In fact, this degradation pathway was undesirable in our case and of
course the main point of the article remained that the super-bases did exhibit very good
catalytic activity despite this de-activation pathway. The profound influence of solvent on
reactivity led us to realize that DMF solutions promoted CO2 hydrosilylation. Such an
observation is very important for the field. Indeed, most if not all articles dealing with base
catalyzed CO2 hydrosilylation report that best results are obtained in the presence of polar
solvents, without giving any detailed explanation.138,276,301–303
These results also shed doubts on the true catalytic role of NHCs in Ying’s original report.
Indeed, all of the experimental work was carried out in DMF and the reported turnovers were
obtained from a multi-day reaction. It is perhaps a coincidence that the numbers are almost
identical to those we obtained during a blank reaction carried out in DMF. Interestingly, the
authors report methanol derivatives as the sole reaction products while the main products
observed in our study are formoxysilanes. Thus, it could be that factually, NHCs are
catalyzing the reduction of formoxysilanes to methoxysilanes.
On the other hand, it is clear that bases such as phosphazenes do promote the reduction of
carbamates, even in polar solvents such as THF. As such, it would not be shocking to say
that NHCs are also active catalysts, but care must be taken on how the turnovers are
calculated in the cases where DMF is used as solvent. Anyhow, the implications of this
discovery goes much beyond a debate of turnover numbers. Indeed, if coordinating solvents
can promote hydrosilylation of CO2, it is most likely that other such processes, notably the
reduction of aldehydes, ketones and esters could be promoted using a similar strategy. Such
mild reaction conditions could be of great interest in the late stage functionalization of
complex molecules since hydrosilanes are much more tolerant of functional groups as
opposed to the traditionally used boron or aluminum hydrides.
156 Similarly, if the hypothesis of an hyper-coordinate silicon species is valid, the power of this
strategy could be harnessed to develop many other reactions of interest. If hydrides can be
transferred readily, it might be that in the appropriate conditions, alkyl or even aryl groups
could be delivered via similar hyper-coordinate intermediates. Altogether, the unusual
solvent effect that was discovered opens up an exciting nexus of possibilities in main group
chemistry.
In hindsight, even though our original working hypothesis did not come through and even
though PMHS could not be used as a reductant, this report still represents an important result
in the field. In fact, the catalytic system can either generate formoxysilanes or
methoxysilanes, depending on the reaction conditions. With a TOF of 32 h-1 and a TON of
759, the commercially available 1 represents the best catalyst (if not the only, depending on
the true activity of NHCs) for the metal-free hydrosilylation of CO2 to methanol.
157 7 Metal-free hydrogenation of carbon dioxide
7.1 Introduction to FLP mediated hydrogenations
Having developed the most performant metal-free systems for the hydroboration and
hydrosilylation of CO2 to methanol derivatives, an ambitious final goal had yet to be reached:
the metal-free catalytic hydrogenation of CO2. In fact, even though CO2 reduction employing
highly energetic material such as boranes and silanes is of fundamental interest and allowed
the understanding of many important reactivity parameters, the only sustainable and
economically relevant source of energy that could be implemented in the process remains
hydrogen. Fortunately, apart from CO2 binding, FLPs also exhibit versatile hydrogen
splitting reactivity. In fact, shortly after Stephan and Welch’s initial discovery of metal-free
hydrogen activation, a tremendous amount of work has been oriented towards hydrogenation
reactions. For a complete discussion on these hydrogenation systems, the reader is directed
to recent books,309,310 and a series of excellent reviews.99–102,311 Notable recent examples
include the hydrogenation of alkynes by Repo312 as well as the very recent hydrogenation of
carbonyl compounds by Ashley63 and Stephan.313
This chapter presents a specific approach to the hydrogenation of carbon dioxide which was
refined over multiple years. Tackling a task as difficult and challenging as metal-free CO2
hydrogenation, we first attempted to identify and understand the fundamental concepts
governing FLP reactivity. In order to fully appreciate the underlying concepts, the
development of the ideas that ultimately led to the design of this system will be fully detailed
in the following section, leading in the end to the elaboration of three simple statements that
guided us through the design of FLPs suitable for the task at hand.
159 7.2 Understanding FLP hydrogenation
Most of the early FLP mediated catalysis reports involved polar substrates that were naturally
prone to hydrogenation such as imines, enamines, silylenol ethers and related
substrates.99,157,314,315 The use of these naturally reactive molecules allowed catalysis, but the
class of substrate remained quite restricted. Certainly, the metal-free aspect of FLP catalysts
attracted much attention, but most of the reactions operated in suboptimal conditions, often
requiring high catalytic loadings as well as relatively high temperatures and pressures. Of
course, strongly activated substrates could react under ambient conditions, but such
limitations in substrate scope and reaction conditions consistently prevented metal-free
hydrogenation to step out of the shadow of well-established transition-metal based systems.
During the early years of FLP chemistry, the most widespread technique used to identify
active systems was to add H2 to a solution containing the active components, followed by
NMR spectroscopy characterization. The active FLPs identified using this method
systematically led to undesirable thermodynamic stabilization upon H2 splitting. Undesirable
indeed, let us revisit a concept that was first put forward in chapter 1 (Figure 7-1).
Figure 7-1: Qualitative analysis of hydrogen splitting by FLP systems.
160 In analogy to CO2 activation, the activation of H2 by very reactive centers will lead to a
significant stabilization of the H2 splitting product, making it that much more difficult to
reach the desired transition state. This simple thermodynamic analysis allowed us to postulate
a first working hypothesis.
1. The splitting of H2 needs to be energetically neutral (ΔG≈0) to favor optimal transfer
to the substrate.
This statement is simply an expression of the most basic underlying concept of catalysis. In
catalysis, the kinetic barrier of the reaction must be lowered without affecting the overall
thermodynamics as to avoid ending up in a thermodynamic well. In the case of FLPs, the
splitting of H2 is kinetically mandatory in order to transform the unreactive hydrogen
molecule into a reactive proton/hydride pair. Nevertheless, any further stabilization of the
proton/hydride pair will inevitably hinder the transfer of the said reactive fragments.
Following this line of thought, the hydrogen molecule can be simply seen as the combination
of a proton and a hydride. Yet, the transfer of electrons in most FLP systems is uneven,
resulting is a disparity of the electronic charges on the proton and hydride. In a very eyeopening study, Papai and co-workers calculated the thermodynamics of H2 splitting for all
the reported FLP systems at the time.316 Closer analysis of the data that is included in the
supporting information allows us to give an approximate quantification of this charge
disparity. Taking a classical FLP system (Mes3P/BCF) for example, the hydride binding by
BCF was calculated to provide 72.5 kcal mol-1 of stabilization. On the other hand, binding of
a proton to Mes3P leads to 52.7 kcal/mol of stabilization, for a total of 125.2 kcal/mol. In this
particular system, the hydride stabilization accounts for 58% of the total thermodynamic
stabilization while the proton accounts for 42%. By controlling the charge of the reactive
fragments, one can control the inherent reactivity of the Lewis pair, leading to the second
working hypothesis.
2. The Lewis acidity/basicity in an FLP system will dictate the electron distribution of
the generated proton/hydride pair.
161 This statement seems obvious at first, but taking the discussion beyond simple theoretical
analysis will shed light on the repercussions of charge disparity. Perhaps the most striking
report illustrating both working hypotheses 1 and 2 is the publication by Stephan and coworkers that a combination of Et2O and BCF can readily split hydrogen.317 Indeed, the
combination of BCF, Et2O and H2 did not yield any observable change by NMR
spectroscopy. Yet, the use of HD led to the scrambling of the HD molecule into H2 and D2,
indicative of H2 activation. These results clearly show that even though it cannot be seen by
NMR spectroscopy (ΔG>0), H2 activation is occurring (Scheme 7-1).
Scheme 7-1: Hydrogen activation by a BCF/Et2O Lewis Pair.
Since the splitting of H2 cannot be seen by NMR and is therefore endothermic, in accordance
with hypothesis 1, the generated hydrogen fragments should be highly reactive. Secondly, as
Et2O is a very weak Lewis base, it is expected that most of the stabilization required for H2
splitting will be provided through hydride stabilization by the BCF moiety. As such, in line
with hypothesis 2, a very reactive proton should be generated, along with a relatively stable
hydride. Just so, the tentative hydrogenation of styrene derivatives resulted mainly in the
Friedel-Crafts dimers, a reactivity that is typical of acidic catalysis.318 On the other hand, the
use of substrates capable of generating stable tertiary carbocations such as 1,1diphenylethylene were catalytically hydrogenated (Scheme 7-2).
Scheme 7-2: Catalytic hydrogenation by a BCF/ET2O Lewis Pair.
Kinetic experiments showed that the splitting of hydrogen is not the rate-limiting step of the
reaction, leaving room for only one explanation: the system generates a very reactive proton
162 which reacts with the alkene moiety, generating a stable carbocation. The need to heat at 50
°C for 4 days to achieve complete hydrogenation is most certainly associated to the ratelimiting step of transferring an unreactive hydride from BCF to the stable carbocation.
This very interesting contribution, along with our report of CO2 reduction using apparently
unreactive FLP catalysts convincingly confirmed that even though some reaction
intermediates are slightly thermodynamically uphill, and thus invisible by NMR
spectroscopy, the FLP systems may still present important activity.
The general consensus in the field of FLP chemistry has always been that the use of strong
Lewis acids is essential for H2 activation, thus most of the efforts so far have been oriented
towards systems based on BCF moieties or other highly Lewis acidic fragments. All these
system will inevitably have in common the strong stabilization of a hydride fragment,
limiting the scope of hydrogenation reactivity to activated substrates capable of accepting a
weakly nucleophilic hydride. Higher catalyst loadings, higher H2 pressures or higher
temperature have proven to be efficient ways to promote this hydride delivery in a number
of systems, but the use of weaker Lewis acids could unleash the full potential of an
unexplored facet of FLP chemistry, leading to our third working hypothesis.
3. The electron density of the generated proton/hydride pair must be complementary to
that of the substrate.
Admittedly, the use of a weak Lewis acid requires the use of a stronger Lewis base. There
are only two reported examples of this “inverse’’ FLP approach that combines a strong Lewis
base with a weak Lewis acid. The first one was reported by Bercaw and Labinger who
accidentally discovered that a combination of a trialkylborane with a strong phosphazene
base led to dihydrogen splitting and subsequently, to hydrogenation of the CO fragment via
delivery of a reactive hydride (Scheme 7-3).170 Using an even stronger carbanion base,
Krempner and co-workers were able to split hydrogen with a Lewis acid as weak as BEt3.319
Unfortunately, the reactivity of these promising systems was never pushed any further.
Nonetheless, those reports illustrate that it is possible to split H2 using weak Lewis acids.
163 Scheme 7-3: FLP mediated reduction of CO at a rhenium center using a strong
phosphazene base in combination with a weak Lewis acid.
While the previous examples represent rather extreme cases, precise control of the charges
on the proton and hydride fragments opens endless possibilities. For instance, let’s apply the
concept to a very challenging reaction: the hydrogenation of CO2. From a very simplistic
perspective, the reaction can be seen as the hydrogenation of very stable C=O bonds. Further
inspiration can be found from one of the best metal-based catalyst for C=O bond
hydrogenation. Indeed, the Noyori catalyst operates efficiently by simultaneously delivering
a proton/hydride pair to reduce the C=O bond (Figure 7-2). 320
Figure 7-2: Schematic representation of an hydrogenation reaction mediated by the Noyori
catalyst.
164 As it was stated many times, the ambiphilic nature of CO2 advocates the use of an ambiphilic
activation process. Quite simply, the Lewis acidic fragment can be a proton, while the Lewis
basic fragment can be a hydride. Obviously, an intramolecular system would alleviate the
entropic challenge of combining a catalyst, hydrogen and CO2 in a single transition state and
would also take full advantage of a simultaneous proton/hydride delivery mechanism (Figure
7-3).
Figure 7-3: General strategy for the hydrogenation of CO2 by FLP systems.
The question that remains is: should the hydride be more reactive than the proton or viceversa? While it is generally accepted that the carbon dioxide molecules possesses an
ambiphilic character, what is most often overlooked is the fact that CO2 has two partial
negative charges. As such, the positive charge on carbon is twice as important as the negative
charge located on a single oxygen (Figure 6-4).
Figure 7-4: A CO2 molecule with the partial charges.
It is thus logical that the most important part of an electron transfer to CO2 should occur
from the hydride to the carbon atom. In fact, it was illustrated regularly in the present thesis
that hydride delivery was crucial in CO2 reduction systems. Similar conclusions can be drawn
from a number of reactive transition-metal based catalysts capable of CO2 hydrogenation to
formate with impressive activity.77,243,282
The concept was elegantly demonstrated by Musgrave and Zimmerman in a theoretical study
of a model reaction: hydrogen transfer to CO2 from ammonia borane (NH3BH3). From Figure
7-5,321 it is very evident that in the transition state, most of the electronic density is transferred
165 from the hydride fragment to the antibonding orbital of carbon. Even though most of the
charge is transferred through the hydride, simultaneous interaction of the proton remains a
key aspect in order to facilitate the hydrogenation process by avoiding high energy anionic
intermediates (Figure7-5).
H
H
H N B H
H
H
+
O=C=O
H
H
H N B H
H
H
O C
H
H
H N B H
H
O
+
H
O C
O
Figure 7-5: Hydrogen transfer from ammonia-borane to CO2.*
*Figure reprinted with permission from (Zimmerman, P. M.; Zhang, Z.; Musgrave, C. B. Inorg.
Chem. 2010, 49, 8724-8728.) © 2010 American Chemical Society
166 Taking the preceding arguments into account, the strategy becomes much clearer. An
intramolecular FLP which will split hydrogen at ΔG≈0 and generate an electron rich hydride
fragment must be designed. But a final issue remains: FLPs are known to bind both H2 and
CO2. In order to avoid a thermodynamic pit, CO2 binding must be prevented, meaning that
on top of the previous requirements, the system must bind H2 selectively over CO2. In an
effort to identify the factors that govern this selectivity, the thermodynamics of H2 splitting
and CO2 binding for a variety of arene-bridged FLP models were computed at the ωB97xD/631++G** level of theory (Figure 7-6). It should be specified that the proven versatility and
stability of the aryl-bridge motivated the choice of this particular framework.
Figure 7-6: DFT study of the thermodynamics of H2 splitting and CO2 binding by a variety
of aryl bridged FLP systems.*
*Level of theory:WB97XD/6-31++G**, solvent=benzene (SMD). Energies are in kcal.mol-1.
167 With an energetic difference of 25.1 kcal.mol-1, an aluminum-phosphorus system (1, red)
shows a significant preference for CO2 binding over hydrogen binding, most likely due to
the strong oxophilicity of organometallic aluminum compounds. However, changing the
aluminum atom for a boron atom (2, black) results in a significant decrease of the CO2
binding energy. But still, the H2 binding energy remains higher than CO2 binding by 8.4 kcal
mol-1. Interestingly, a 4.8 kcal mol-1 weaker CO2 binding is observed when changing from a
phosphorus atom (1, red) to a nitrogen center (3, pink). Presumably, this difference may be
due to a better overlap of the phosphorus orbital which is very rich in s character with the
antibonding orbital of CO2. Following the trend, the use of an N-B based FLP results in the
weakest carbon dioxide binding energy (4, blue). Remarkably, in sharp contrast with FLPs
1-3, the N-B based FLP 4 stabilizes the splitting of H2 preferentially to CO2 binding.
Finally having identified all the key characteristics of an ideal FLP, we set out to prepare a
first generation of derivatives. First attempts were directed towards the preparation of very
simple FLP systems containing methyl substituents on nitrogen and mesityl groups on boron.
The bulky aryl groups on boron were chosen to prevent possible hydrolysis in the presence
of reduction products as well as to favor release of the reduction products. The following
chapter deals with the development of these N-B based FLPs for the hydrogenation of CO2.
168 7.3 Research article: Intramolecular B/N Frustrated Lewis Pairs and the
Hydrogenation of Carbon Dioxide
7.3.1
Résumé
Les espèces PLF 1-BR2-2-NMe2-C6H4 (R = 2,4,6-Me3C6H2 1, 2,4,5-Me3C6H2 2) réagissent
avec H2 dans une séquence de réactions d’activation de l’hydrogène et de protodéborylation
pour former (1-BH2-2-NMe2-C6H4)2 3. Bien que 1 puisse réagir avec un mélange
d’hydrogène et de dioxyde de carbone pour former un mélange de formates, d’acétals et de
méthoxydes,
le
composé
2
donne
un
seul
produit
de
réduction,
l’acétal
(C6H4(NMe2)(B(2,4,5-Me3C6H2)O)2CH2 4. Le mécanisme de réduction est considéré.
7.3.2
Abstract
The FLP species 1-BR2-2-NMe2-C6H4 (R = 2,4,6-Me3C6H2 1, 2,4,5-Me3C6H2 2) reacts with
H2 in sequential hydrogen activation and protodeborylation reactions to give (1-BH2-2NMe2-C6H4)2 3. While 1 reacts with H2/CO2 to give formyl, acetal and methoxy-derivatives,
2 reacts with H2/CO2 to give (C6H4(NMe2)(B(2,4,5-Me3C6H2)O)2CH2 4. The mechanism of
CO2 reduction is considered.
169 7.3.3
Introduction
General concerns regarding global warming, climate changes, and the need for renewable
fuels have prompted researchers from around the world to target methodologies to utilize
CO2 as a C1 source.15,75,322,323 Transition metal catalysts have been uncovered that either
hydrogenate70,80,125,244,288, hydrosilylate,70,90,130,131,283 or hydroborate CO2 to formic acid,
methanol, methane, CO and methoxide derivatives.71,72,133,282,324,325 An alternative strategy
for the reduction of CO2 which is gaining attention is based on non-metal catalysts. While
strong
Lewis
bases
can
reduce
CO2
using
either
hydrosilanes
or
hydroboranes,92,138,274,279,284,326 our research groups have been exploring the utility of
Frustrated Lewis Pairs (FLPs) for the capture and the reduction of CO2. Since the original
report by Stephan, Erker and co-workers on the capture of CO2 by FLPs,103 a number of interor intramolecular FLP variants have been employed to sequester CO2 and much of this
chemistry has been recently reviewed.101,250,311,327 Beyond capture, FLP mediated CO2
reductions have been probed. The reaction of Al/P FLPs with CO2 and ammonia-borane was
shown to give methanol106 while an alternative reaction pathway affords CO.209,328 In a
related study, Piers and coworkers used Et3SiH as a reductant to catalytically generate CH4
and (Et3Si)2O.104 While Stephan and coworkers have also reported the catalytic reduction of
CO2 using phosphine/CH2I2 and ZnBr2 to give CO and phosphine oxide,160 Fontaine and
coworkers described one of the most efficient systems to date for the reduction of CO2 using
ambiphilic FLP Ph2PC6H4B(O2C6H4), generating methoxyboranes with TOF exceeding 900
h-1 at 70 °C.139,269 In related work, Stephan and co-workers have also described the use of
C3H2(NPR2)2BC8H14238 and phosphines240 to catalyze the hydroboration of CO2 affording
mixtures of HCO2(B(C8H14)), H2C(OB(C8H14))2 and MeOB(C8H14).
170 Although hydroboration and hydrosilylation of CO2 to methanol are academically
interesting, only the hydrogenation of CO2 could be industrially viable. Ashley and O’Hare63
have reported the only metal-free system in which CO2 is hydrogenated. Employing the FLP
TMP/B(C6F5)3 (TMP = tetramethylpiperidine) , CH3OH was generated after 6 days at 160
°C under CO2 and H2. While this precedent establishes the concept, the development of an
efficient FLP catalyst requires attention to the entropic challenge associated with bringing all
reagents together. In addition, since the transformation of CO2 to methanol is a 6-electron
process generating formic acid and formaldehyde as intermediates, thus involving three very
distinct reduction steps, the Lewis acidity of the electrophilic boron center must be
judiciously designed to facilitate hydride delivery. To address these issues, we are exploring
intramolecular FLP systems which incorporate tri-aryl boron centers that are significantly
less Lewis acidic than the ubiquitous B(C6F5)3.97,253 In this fashion, the proximity of the
Lewis acid and base reduces the entropic barrier, while the reduced Lewis acidity at B is
expected to promote hydride delivery. In this manuscript, we describe the reactivity of these
B/N FLPs with H2 and the subsequent hydrogenation of CO2 at ambient temperature.
7.3.4
Results and discussion
The fluorescence properties related to compounds 1-BR2-2-NMe2-C6H4 (R = 2,4,6-Me3C6H2
(1), 2,4,5-Me3C6H2 (2)) have been previously studied, although in our hands the reported
synthetic route proved problematic.329 Nonetheless, 1 was prepared in 72% yield by the
stoichiometric reaction of 1-Li-2-NMe2-C6H4 with (2,4,6-Me3C6H2)2BF in toluene. In a
similar fashion, the corresponding reaction with (2,4,5-Me3C6H2)2BCl yielded 2 in 64%
yield, following crystallization from a saturated solution in cold hexanes (Scheme 7-4).
Scheme 7-4: Preparation of 1-2.
171 The reactivity of these bright green compounds with both H2 and CO2 was investigated.
When a benzene-d6 solution of 1 was exposed to either 1 atm of CO2 or 4 atm of H2, no
change was evidenced by 1H NMR spectroscopy. However, heating for 24 h a solution of 1
at 80 °C under 1 atm of HD led to isotopic scrambling as evidenced by the observation of H2
and HD by 1H NMR spectroscopy. In addition, new signals at 6.72 and 2.16 ppm were
observed and assigned to free mesitylene, suggesting that protodeborylation occurred after
the activation of H2. Indeed, protodeborylation reactions have been shown to occur before in
related systems.193,314,330 Monitoring of this protodeborylation with the use of cyclohexane
as an internal standard, revealed that 1 releases both of its mesityl groups after 72 hours at
80 °C affording (1-BH2-2-NMe2-C6H4)2 (3). The nature of the aryl group impacts the facility
of protodeborylation as the species 2, with one less methyl in ortho position than 1, was
converted to 3 after 72 hours at room temperature or after 6 hours at 80 °C. Compound 3 was
prepared on a larger scale from 1 at 80 °C under 4 atm of H2 for 48 hours and was ultimately
isolated in 54 % yield. The broad signal at 3.55 ppm in the 1H NMR spectrum was attributed
to B-H protons, which became sharper with 11B decoupling. The presence of the B-H bonds
was further confirmed by the broad 11B NMR signal at 2.5 ppm. The HRMS data suggest that
compound 3 is dimeric (m/z: 265 = [MH]), which is further supported by the observation
of inequivalent methyl groups on nitrogen. This view was further supported by computational
studies, in which a number of isomeric forms of 3 were considered and where the dimeric
form which adopts a ‘’boat’’ shaped 8-membered ring was computed to be 9.2 kcal.mol-1
more stable than the monomeric form (Scheme 7-5).
Scheme 7-5: DFT study of possible isomers of 3.*
*Level of theory:WB97XD/6-31++G**, solvent=benzene (SMD). Energies are in kcal.mol-1.
172 It is noteworthy that Repo and coworkers have recently described 1-BH2-2-TMP-C6H4,
which is also a dimer; however in this case, structural characterization confirmed that the
steric congestion favors dimerization via the B-H bonds.331
DFT calculations were also employed to shed light on the mechanism of this transformation
(Scheme 7-6). The activation of H2 by 1 or 2 proceeds through TS1 to generate A’ and A”,
respectively, in a slightly endothermic process. Subsequent protodeborylation can occur
through TS2, eliminating the B-bound aryl substituent to give the ambiphilic hydroboranes
B’ and B”, respectively.20 Further activation of H2 via TS3 to give C’ and C”, prompts a
second protodeborylation reaction pathway via TS4 affording the primary amino-borane
product 3. While the computed energies for these reactions of 1 and 2 follow the same trends,
the reduced steric demands of 2 leads to significant lowering of the activation barriers.
protodeborylation
TS2': 28.0(15.5)
TS2": 22.0(9.8)
H 2 split
TS1':21.7(9.2)
TS1":16.9(4.2)
H R H
Me2 N
B R
Me2N
H H R
B R
A
TS2
Me2N
A': 12.2(0.4)
A": 7.0(-4.0)
protod eborylation
TS4': 11.9(3.1)
H 2 split
TS4":
11.6(5.1)
TS3': 9.4(0.9)
TS3": 10.1(4.3)
Me2N
C': -1.4(-9.1)
C": 0.4(-7.5)
1: 0(0)
2: 0(0)
B' : -9.0(-6.7)
B":-7.4(-5.6)
B R
B
H H H
B R
C
Me2N
3': -10.5(-5.4)
3": -9.7(-2.0)
H
H
B H
3
Scheme 7-6: DFT study of H2 activation and protodeborylation events.
*Level of theory:WB97XD/6-31++G**, solvent=benzene (SMD). Energies are in kcal.mol-1. X’
refers to R = 2,4,6-Me3C6H2, X” refers to 2,4,5-Me3C6H2.
173 The hydrogenation of CO2 with 1 and 2 was investigated and in general was found to produce
several boron bound formates, acetals and methoxides (Table 7-1). Heating a benzene-d6
solution of 1 to 80 °C for 24 hours under 4 atm of H2 and 1 atm of 13CO2, resulted in the
appearance of doublets arising from the coupling with the 13C atom for the formate (HCOO
at ca. 8.5 ppm (JCH ~ 210 Hz)) and acetal derivatives in 1H NMR spectrum (ca. 5.2 ppm (JCH
~ 165 Hz)). It was found that CO2 was transformed into 0.89 equivalents of boron bound
formates relative to the amount of 1 at the start of the reaction, and 0.31 equivalents of boron
bound acetals. Repeating the experiment with a reduced CO2 pressure (0.5 atm) led to similar
conversions to formates and acetals, but in addition 0.1 equivalent boron bound methoxides
were formed (ca. 3.5 ppm, JCH ~ 140 Hz). Further reduction of the CO2 pressure to 0.1 atm
resulted in the formation of methoxides and traces of 13CH4. An experiment under 1 atm of
CO2 and 4 atm H2 in bromobenzene-d5 yielded 0.75, 0.21 and 0.07 equivalents of formate,
acetal and methoxide species after only 24 hours at 130 °C. In contrast, 3 did not react in the
presence of H2 and CO2, even after prolonged heating at 80 °C. This lack of reactivity is
consistent with its dimeric nature that provides a stabilization of 13.4 kcal.mol-1.
174 Table 7-1: Hydrogenation of carbon dioxide by 1 and 2
#
FLP
CO2
T (h)
Equivalentsa of
T (°C)
(atm)
HCOO
OCH2O
H2
CH3O
consumed
1
1b
1
216
80
0.89
0.31
0
1.5
2
1b
0.5
216
80
0.84
0.34
0.1
1.8
3
1b
0.1
216
80
0
0
0.08
0.25
4
1b,c
1
24
130
0.75
0.21
0.07
1.4
5
2d,e
1
72
23
0
0.37
0
0.74
6
2
1
3
80
0.21
0.30
0
0.81
Conditions: 0.014 mmol 1 or 2, 0.4 mL benzene-d6, 4 atm H2. Yields were determined by NMR
integration with respect to an internal standard (cyclohexane). a. Equivalents of the indicated
hydrogenation moiety relative to the amount of starting aminoborane. b. A white precipitate crashed
out of the solution so 0.1 mL of CD3CN was added before taking the spectra. c. Reaction was carried
out in bromobenzene-d5 d. Reaction was carried out under 1 atm of H2 e. Compound 4 was exclusively
formed.
175 Interestingly, the analogous reactions of 2 gave a single acetal species after 72 hours at room
temperature in the presence of 1 atm H2 and 1 atm CO2 as evidenced by 1H NMR
spectroscopy. On the other hand, higher temperature gave some additional formate species.
When carried out on a larger scale, product 4 was isolated in 60% yield. The NMR data and
the crystallographic structure (Figure 7-7) supported the formulation of 4 as
(C6H4(NMe2))(B(2,4,5-Me3C6H2)O)2CH2.
Based
on
these
observations,
the
first
protodeborylation step is believed to be required prior to CO2 reduction while complete
protodeborylation inhibits the reduction processes due to dimerization of 3.
Figure 7-7: ORTEP depiction of 4.*
*50% thermal ellipsoids are shown, N: blue, C: black, O: red, B: orange. H-atoms are omitted for
clarity.
The initial steps in reaction of 1 and 2 with H2/CO2 were probed using DFT computations.
The reactions of the products of H2 activation A-C (Scheme 7-6) with CO2 were considered.
The barriers to reduce CO2 with A and B were computed to range between 27.2 and 34.7
kcal.mol-1 whereas the transition state with C was found to be only 24.4 kcal.mol-1 for R =
2,4,6-Me3C6H2 and 22.1 kcal.mol-1 for R = 2,4,5-Me3C6H2. The transition state of interest
(Figure 7-8) illustrates a concerted interaction of the proton on N with one of the O of CO2
176 with the simultaneous interaction of the boron-bound hydride with the C atom, thus directing
the hydride delivery to the carbon atom. This TS is reminiscent of that proposed for the
bifunctional Noyori-type catalysts for metal-based ketone reduction320 and a similar
transition state was proposed by Musgrave, Zhang and Zimmerman321 for CO2 reduction
using ammonia borane as a model reductant. Subsequent reductions of formic acid are
thought to proceed either via similar hydride delivery to formate or by simple hydroboration,
generating acetal derivatives. It is also interesting that the minor variation in the steric
demands of the substituent on B provide a mixture of reduction products in the case of
reactions of 1 yet allow the isolation of 4 at room temperature in the reaction of 2.
Figure 7-8: Geometry of TS for reaction of C with CO2.*
*Level of theory:WB97XD/6-31++G**, solvent=benzene (SMD).
While previous reports have described conceptually important metal-free catalytic
hydrosilylation or hydroboration of CO2, the present report is a rare example of direct FLP
hydrogenation of CO2 as only the earlier report by O’Hare and Ashley63 described the use of
H2 in the metal-free reduction of CO2. Nonetheless, the present intramolecular FLPs effect
this reduction under much milder conditions (ambient temperature).
177 7.3.5
Conclusions
The reactions of the present N/B intramolecular FLPs with H2 demonstrate a rare
case where weakly Lewis acidic boron centres participate in H2 activation. Such
systems offer increased facility for hydride delivery and thus provide an avenue to
CO2 reduction. Moreover, the reaction with CO2 is facilitated by the concurrent
interaction of NH and BH fragments with CO2 affording formate, acetal and methoxyderivatives. While the present systems are generated by protodeborylation, the
reactivity suggests that judicious substituent selection could provide an avenue to the
design of intramolecular FLPs catalysts for H2/CO2 chemistry. Efforts towards such
metal-free catalysts for CO2 hydrogenation are the subject of current work in our
laboratories.
178 7.4 Experimental
7.4.1
General experimental:
Unless otherwise specified, all the manipulations were conducted under an inert atmosphere
of dinitrogen, using standard Schlenk and glovebox techniques. Reactions were carried either
in a sealed J-Young NMR tube, in which case NMR conversions are indicated, or in standard
oven dried Schlenk vessels. Benzene-d6 was purified by vacuum distillation from Na/K alloy,
or by degassing by three subsequent freeze pump thaw cycles followed by standing over
activated 3 Å molecular sieves. Acetonitrile-d3 and bromobenzene-d5 were dried over 3 Å
molecular sieves. Dry CO2 was purchased from Praxair and used as received. 13CO2 (99%
isotope label) was purchased from Cambridge Isotope Laboratories or Aldrich. Ultra high
purity hydrogen (5.0 grade) was purchased from Praxair and used as received. 5Bromo,1,2,4-trimethylbenzene,
2-bromo-1,3,5-trimethylbenzene
and
BF3.Et2O
were
purchased from Aldrich or TCI and used as received. FBMes2- was synthesized according to
a
literature
procedure
and
further
purified
by
sublimation.332
(2-
(Dimethylamino)phenyl)lithium was synthesized according to a literature procedure.97
Benzaldehyde was purchased from Aldrich and distilled under reduced pressure before use.
NMR spectra were recorded on Agilent DD2-600 at 600 MHz (1H), 150.84 MHz (13C),
192.45 MHz (11B), Agilent Technologies NMR spectrometer at 500 MHz (1H), 125.758 MHz
(13C), 160.462 MHz (11B), a Varian Inova NMR AS400 spectrometer, at 400.0 MHz (1H),
100.580 MHz (13C), 128.378 MHz (11B). 1H NMR and
13
C{1H} NMR chemical shifts are
referenced to residual protons in deuterated solvent. Multiplicities are reported as singlet (s),
broad singlet (s, br) doublet (d), triplet (t), multiplet (m). Chemical shifts are reported in ppm.
Coupling constants are reported in Hz. gHSQC experiments were performed in order to
confirm C-H correlations. Mass spectroscopy analysis were performed on a JMS T100-LC
using AccuTOF DART.
WARNING: Condensation of high pressure of H2 and CO2 might lead to an explosion of the
glassware. Care should be taken.
179 7.4.2
Synthesis of compounds 1, 2 and precursors
Dimethylbis(2,4,5-trimethylphenyl)stannane
This compound was synthesized with a slightly modified approach from a known
procedure.333
Ca. 50 mL of THF and 3.48 g (143mmol, 4.0 equivalents) of magnesium turnings were
combined in a Schlenk flask. To the resulting mixture, 25.9 g (130 mmol, 3.7 equivalents) of
1-bromo-2,4,5-trimethylbenzene were added and heated at 40 °C for 2 hours. Me2SnCl2
(7.7g, 35 mmol, 1.0 equivalent) was dissolved in ca. 20 mL of THF in a separate Schlenk
flask. The solution containing the tin compound was then cannulated into the Schlenk flask
containing the Grignard reagent. The resulting mixture was heated at 50°C for 4 hours.
Work-up: The solution was cooled down in an ice bath and 20 mL of a saturated solution of
NH4Cl was added. The contents were transferred to a separatory funnel and extracted three
times with a mixture of Et2O/Hexanes (50/50). The organic layer was then washed with water
and with brine. The organic layer was dried with Na2SO4 and the solvents removed by rotary
evaporation. The compound was then passed through a silica plug using hexanes as an eluent
which was then removed by rotary evaporation yielding 11.3 g of white crystals. Yield =
83%.
1
H NMR 400 Mhz : δ 7.20 (s, 2H, m-Ar); 7.05 (s, 2H, o-Ar); 2.35 (s, 6H, m-Me); 2.27 (s, 6H,
p-Me); 2.25 (s, 6H, o-Me); 0.55 (s, 6h, Sn-Me). 13C {1H} (101 MHz): δ 142.3 (m-Ar); 137.8
(m-Arquat); 137.2 (p-Ar); 136.7 (o-Ar); 133.3 (o-Arquat); 130.7 (Sn-Ar); 24.4 (m-Me); 19.8
(p-Me); 19.3 (o-Me); -8.03 (Sn-Me).
180 Chlorobis(2,4,5-trimethylphenyl)borane
This compound was synthesized with a slightly modified approach from a known
procedure.334
In a Schlenk flask, 11.1 g of dimethylbis(2,4,5-trimethylphenyl)stannane (0.029 mol, 1
equivalent) was dissolved in ca 10mL of hexanes in a teflon capped Schlenk flask with a
magnetic stir bar. 29 mL (0.029 mol, 1 equivalent) of a 1.0 molar solution of BCl3 in heptane
was added at once. The resulting mixture was heated for 40 hours at 100 °C with vigorous
stirring. Upon reaction completion, the Schlenk flask was allowed to cool to r.t. naturally
after which it was placed in a cold bath. The solution containing the title compound was
filtered from the precipitated Me2SnCl2. The resulting solution was then evaporated until the
title compound started precipitating out. The flask was then stored at -35 °C for 4 hours after
which the hexanes were filtered by keeping the solution at -35 °C. The compound was then
further purified by sublimation at 80 °C at 0.1 Tor. 6.43 g of white crystals, yield = 80%.
Recrystallization from hexanes yielded monocrystals suitable for X-ray crystallography.
(The structure is reported in S9).
1
H NMR 400MHZ: δ 7.62 (s, 2H, o-Ar); 6.8 (s, 2H, m-Ar); 2.28 (s, 6H, o-Me); 1.98 (s, 6H,
p-Me); 1.97 (s, 6H, m-Me); 13C{1H} (101 MHz): δ 140.8 (m-Arquat); 140.5 (p-Arquat); 139.3(s,
br, Ar-B); 136.8 (m-Ar); 133.5 (o-Arquat); 132.3 (o-Ar); 22.7 (o-Me); 19.8 (p-Me); 19.1 (mMe).
181 1-(Dimesitylboryl)-2-NMe2-C6H4 (1)
This compound was synthesized with a slightly modified approach from a known
procedure.97
1.5 g (5.5 mmol, 1.0 equivalent) of dimesitylboron fluoride was dissolved in toluene and
cannulated to a -78°C solution of 700 mg (5.5 mmol, 1.0 equivalent) of (2(dimethylamino)phenyl)lithium in 10mL of toluene. The resulting mixture was then left to
warm to r.t. naturally and to stir for 16 hours. Upon reaction completion the solution was
bright, fluorescent green.
Work-up: The salts were left to separate without agitation and the solution was filtered via
cannula. The residue was washed once with toluene (10mL). The volatiles were then removed
in vacuo. Upon cooling, 1.8 g of a green solid was recovered: yield= 91%. The compound
can be further purified by recrystallization from a saturated hexane solution at -35°C. Using
this method, 1.46 g of pure compound was recovered. Yield = 72%
The characterization of this compound is conform to that previously reported.329
1-(Bis(2,4,5-trimethylphenyl)boryl)-2-NMe2-C6H4 (2)
This compound was synthesized with a slightly modified approach from a known
procedure.97
967 mg (3.4 mmol, 1.0 equivalent) of chlorobis(2,4,5-trimethylphenyl)borane was dissolved
in toluene and cannulated to a -78°C solution of 432 mg (3.4 mmol, 1.0 equivalent) of (2(dimethylamino)phenyl)lithium in 10 mL of toluene. The resulting mixture was then left to
warm to r.t. naturally and to stir for 16 hours.
Work-up: The salts were left to separate without agitation and the solution was filtered via
cannula. The residue was washed once with toluene (10 mL). The volatiles were then
removed in vacuo and left under vacuum at 110 °C for 2 hours. Upon cooling, 1.1 g of a
sticky green solid was recovered: yield= 88 %. The compound can be further purified by
recrystallization from a saturated hexane solution at -35 °C. Using this method, 807 mg of
pure compound was recovered. Yield = 64 % The characterizations of this compound is
conform to those that were previously reported.329
182 7.4.3
HD scrambling with 1
In a glove box, 5.2 mg (0.014 mmol) of 1 was dissolved in benzene-d6 (0.5 ml) and the green
solution transferred to a J Young NMR tube. The solution was degassed via three freezepump-thaw cycles in liquid nitrogen. The tube was charged with HD (4 atm) and 1H NMR
recorded. The reaction was heated at 80 °C for 14 h, cooled and 1H-NMR recorded. (Scheme
7-7)
7.4.4
Synthesis of 3 from 1
Scheme 7-7: Synthesis of 3 from 1.
In a glove box, 1 (46 mg, 0.13 mmol) was dissolved in benzene (4 ml) and the green solution
transferred to a J Young flask (bomb) equipped with a teflon coated magnetic stir bar. The
solution was degassed via three freeze-pump-thaw cycles in liquid nitrogen. The flask was
charged with H2 (4 atm) and the mixture was stirred and heated at 80 °C for 48 h. The volatiles
were removed in vacuo to give a pale green solid. The product was crystallised from hot nhexane (2 ml), which after decantation of the supernatant and washing the solid with nhexane (0.5 ml) gave 3 as white feathery crystals (9 mg, 54%).
183 *The H atoms are numbered based on the carbon atom to which they are attached
1
H NMR (600 MHz, benzene-d6): δ 7.76 (dd, J = 7.3, 1.5 Hz, 1H, H7), 7.16 (td, J = 7.3, 1.0
Hz, 1H, H5), 6.99 (ddd, J = 8.0, 7.3, 1.5 Hz, 1H, H6), 6.62 (dd, J = 8.0, 1.0 Hz, 1H, H4),
3.54 (app. br. d, J = 96 Hz, 1H, BH), 2.77 (s, 3H, C2), 2.55 (s, 3H, C1).
13
C{1H} NMR (151 MHz, benzene-d6): δ 158.1 (C3), 137.3 (C7), 126.6 (H5), 126.5 (H6),
116.9 (C4), 59.3 (C1), 48.2 (C2).
11
B{1H} NMR (193 MHz, benzene-d6): δ +2.5.
HRMS (DART-TOF+): mass [MH] calc’d for C16H23B2N2 265.20473 Da, measured
265.20539 Da.
7.4.5
Hydrogenation of CO2 - J Young NMR tube experiments
In a glove box, the solvent (0.5 ml) and cyclohexane (~1 l) was added to 0.014 mmol of
aminoborane. The bright green solution was transferred to a J Young NMR tube and the
mixture was degassed via three freeze-pump-thaw cycles in liquid nitrogen. The tube was
charge with the stated pressure of CO2 or 13C labelled CO2 (99% 13C). The mixture was frozen
in liquid nitrogen and the tube charged with H2 (4 atm). The 1H NMR was recorded and the
tube heated at the stated temperature. The reaction was periodically monitored by 1H-, 13Cand 11B-NMR. When stated, at the end of the reaction, the reaction solution was colourless
and contained a white precipitate. The mixture was homogenised by cooling to 20 °C and
then adding acetonitrile-d3 (0.15 ml) followed by shaking of the tube. The final 1H-, 13C- and
11
B-NMRs were then recorded. For the experiments involving 2, the solution remained
homogeneous throughout the entire process.
184 Table 7-2: Hydrogenation of CO2 with 1 and 2
#
Cat.
T(°C)
1a
2a
3a
t (h)
Solv.
24
1
80
72
216
C6D6
Number of CO2 equiv. hydrogenated
H2
11
B shifts / ppm
HCOO
H2CO
CH3O
CH4
equiv.
* = <10%
0.24
0.02
0
0
0.28
+72*, +44*, +7.2
0.31
0.02
Trace
0
0.36
+44*, +7.2, 5.2
0.30
0.00
Trace
0
0.32
+7.2*, +5.2, +2.5
4a,b
216
0.89
0.00
Trace
0
1.51
+5.2, +2.5
5c
24
0.21
0.04
Trace
0
0.29
+72*, +44*, +7.2
0.29
0.03
Trace
0
0.36
+44*, +7.2
0.29
0
0.01
0
0.33
+7.2*, +5.2, +2.5*
6c
7c
1
80
72
216
C6D6
8b,c
216
0.84
0.34
0.01
0
1.54
+7.2*, +5.2, +2.5
9d
24
0
-
0.065
-
0.21
+72, +6.7*, +2.5*
0
0
0.075
0.002
0.24
+6.7, +2.5
0
0
0.083
0.004
0.26
+6.7, +2.5
0
0
0.083
-
0.25
+6.7, +2.5
0.41
0
0.079
0
0.64
+5.2*, +2.2
0.75
0.21
0.079
0
1.39
+5.2, +2.2
0.19
0.28
0
0
0.74
+46*, +7.0
0.22
0.31
0
0
0.84
+46*, +7.0
0.37
0
0
0
0.74
+46*
10d
11d
1
80
12b,d
13a
14a,b
15a
16a,b
17e
72
216
C6D6
216
1
130
2
80
2
23
24
24
6
6
72
BrC6D5
C6D6
C6D6
Conditions: 4 atm. H2 a.1 atm CO2 b. reaction mixture homogenized with 0.1 mL of CD3CN c. 0.5 atm CO2
d.0.1 atm CO2 e. 1 atm H2 1 atm CO2
185 7.4.6
Synthesis of 4 from 2 (Bigger scale hydrogenation)
Scheme 7-8: Synthesis of 4 from 2 (Bigger scale hydrogenation).
In a glove box, 2 (200 mg, 0.54 mmol) was dissolved in toluene (4 ml) and the green solution
transferred to a teflon capped Schlenk flask equipped with a teflon coated magnetic stir bar.
The solution was degassed via three freeze-pump-thaw cycles in liquid nitrogen. The flask
was charged with CO2 (1atm) and H2 (1 atm) and the mixture was stirred at r.t. for 72 h after
which the bright green coloration had almost completely disappeared and white precipitate
had formed. The volatiles were removed in vacuo and left under vacuum at 60°C for one
hour to ensure complete removal of 1,2,4-trimethylbenzene. The residue was dissolved in hot
hexanes (ca 4mL) and filtered hot. The hexane solution was stored at -35°C for 72 hours,
yielding the title compound as colorless crystals (90 mg, 60%). Note that the crystals readily
re-dissolve upon warming of the hexanes solution. The crystals were isolated by cold
filtration at -35°C followed by rapid evacuation of the remaining traces of hexanes in vacuo.
(Scheme 7-9)
186 Scheme 7-9: Assignment of the NMR spectra for 4.
*The H atoms are numbered based on the carbon atom to which they are attached
4: 1H NMR (500 MHz, benzene-d6) δ 7.69 (s, 2H, H13), 7.29 (d, J = 7.2 Hz, 2H, H6), 7.19
(t, J = 7.8 Hz, 2H, H4), 6.99 (s, 2H, H10), 6.80 (t, J = 7.3 Hz, 2H, H5), 6.66 (m, 2H,H3),
5.72 (s, 2H, H17), 2.82 (s, 6H, H14), 2.48 (s, 12H, H1), 2.04 (s, 6H, H15), 1.97 (s, 6H, H16).
13
C {1H} NMR (126 MHz, benzene-d6) δ (155.7, C8), (143.3, C7), (140.5, C9), (139.6, C13),
(133.8, C6), (132.7, C11) (132.6, C10) (130.0, C4), (128.4, C12), (119.3,C5), (114.8, C3),
(90.9, C17), (43.4, C1), (23.2, C14), (19.7,C15),(19.2, C16).
11
B {1H} NMR (193 MHz, benzene-d6): δ +46.1.
187 7.4.7
Computational details:
All the calculations were performed on the full structures of the reported compounds.
Calculations were performed with the GAUSSIAN 03 and GAUSSIAN 09 suite of
programs.218,251 The ωB97XD functional252 was qualified as promising by Grimme159 and
was used to accurately describe the mechanism of FLP mediated hydrogenation of alkynes312
and was thus used in combination with the 6-31G** basis set for all atoms220,221 The transition
states were located and confirmed by frequency calculations (single imaginary frequency).
The stationary points were characterized as minima by full vibration frequencies calculations
(no imaginary frequency). All geometry optimizations were carried out without any
symmetry constraints. The energies were then refined by single point calculations to include
solvent effects using the SMD solvation model255 with the experimental solvent (benzene)
at the ωB97XD /6-31++G** level of theory.254 All structures with their associated free
enthalpy and Gibbs free energies as well as their cartesian coordinates are fully detailed in
section S8.
188 7.4.8
DFT calculation for the protodeborylation of 1 and 2
Figure 7-9: First protodeborylation step for 1.
189 Figure 7-10: Second protodeborylation step for 1.
190 Figure 7-11: First protodeborylation step for 2.
191 Figure 7-12: Second protodeborylation step for 2.
*Because of the asymmetry in Mes’, many conformers can be possible for most of these
structures. However, it was assumed that energetic differences between them were negligible
and they weren’t all computed. Nevertheless, uncertainty on the values reported might be
slightly higher than for NMe2BMes2, but still we are confident that the conclusions based on
those calculations should not be altered significantly.
192 7.4.9
DFT study of the possible isomers of 3
Figure 7-13: DFT optimized structures of the possible isomers for compound 3.
193 7.4.10 DFT study of the hydrogenation of CO2
Figure 7-14: Hydrogenation of CO2 by 1.
194 Figure 7-15: Hydrogenation of CO2 by 1 (continued).
195 Figure 7-16: Hydrogenation of CO2 by 2.
196 Figure 7-17: Hydrogenation of CO2 by 2 (continued).
197 7.5 Conclusions and perspectives
A first generation of N-B based intramolecular have been designed and tested in the target
reaction of CO2 hydrogenation. While this first attempt did not result in considerable catalytic
activity, many of the proposed hypotheses have been confirmed, supporting the validity of
the proposed strategy. Firstly, it was identified that in contrast to phosphorus or aluminum
based Lewis pairs, the combination of nitrogen and boron favors hydrogen splitting over CO2
binding.
More interestingly, the fact that no H2 splitting adduct is observed whilst the
protodeborylation reactions occurs without difficulty is consistent with a system where H2
splitting is slightly endergonic (ΔG>0). Regardless, a slight lowering of the required free
energy for H2 splitting could be highly beneficial in increasing the concentration of activated
hydrogen in solution. Also, the use of mesityl substituents at the boron center proved to be a
poor choice due to an unanticipated decomposition pathway. Furthermore, the lack of
conformational freedom induced by the bulky groups conveyed additional entropic stress to
the overall system.
Nonetheless, the proposition that poor Lewis acids could promote hydride delivery to CO2
did prove to be a reasonable element of design. Moreover, computations support the
hypothesis that simultaneous hydride/proton transfer contributes to a seamless hydrogenation
process. Unfortunately, decomposition of the system prevents any catalytic turnovers. Still,
the metal-free hydrogenation of CO2 was achieved under ambient conditions for the first
time. Despite the drawbacks, this first iteration of an intramolecular amino-borane FLP
catalyst represents the most efficient reported to date for metal-free hydrogen transfer to CO2.
Perhaps the most important advantage is that the parameters of the system are controlled and
understood. Indeed, by fine tuning the active centers, it is expected that the energy required
for H2 activation can be slightly optimized while simultaneously preventing the undesirable
protodeborylation reaction. The first results towards this goal will be presented in the next
chapter.
198 8 Conclusion and perspectives
8.1 General conclusions
This thesis has presented the development and application of various new transition-metal
free systems for the catalytic reduction of CO2 to methanol derivatives. Altogether, a unique
and unconventional perspective of Frustrated Lewis Pair chemistry inspired the design of the
most active metal-free catalysts for the hydroboration, hydrosilylation and hydrogenation of
CO2. In every chapter, some unsuspected underlying reactivity was discovered, allowing us
to expand the fundamental knowledge surrounding the transformations of the carbon dioxide
molecule.
Firstly, the ability of phosphorus-aluminum Lewis pairs to activate CO2 was unambiguously
demonstrated. While methylene bridged FLPs underwent undesirable rearrangements upon
CO2 binding, the use of a phenyl bridge successfully mitigated this decomposition pathway.
Already, this first incursion in the world of FLP chemistry from the Fontaine group
challenged one of the fundamental aspects of this young research field. Indeed, it was shown
that steric bulk is not always a prerequisite to Lewis pair reactivity. In fact, not only did a
simple aryl-bridged aluminum phosphorus Lewis pair exhibit reversible CO2 binding, it also
proved to be a very effective hydroboration catalyst, or so it seemed.
In truth, the compound was shown to decompose under the reaction conditions. Careful
identification of the decomposition products led to the isolation of a very simple phosphineborane species. Unsurprisingly, this simple molecule did not show any evident spectroscopic
change upon exposure to either the reducing agent or CO2. Then again, pushing the
investigation allowed us to discover that this species was a very active catalyst, challenging
the misconception that highly reactive Lewis pairs were required for CO2 transformation.
Not only was this the first report of a metal-free catalyst for the hydroboration of CO2, but
the activity surpassed that of the best transition-metal based systems. Moreover, this was
accomplished using the cheap, high hydrogen content borane BH3•SMe2 in combination with
a cheap, easily prepared and air stable catalyst.
199 The theoretical investigation of a number of reaction pathways allowed us to establish a solid
understanding of the basic mechanics of FLP chemistry, revealing that the original idea of
simultaneous CO2/substrate activation proved more potent than simple CO2 binding.
Interestingly enough, this realization enabled the use of much less reactive centers to promote
CO2 transformations. Still, we did not envision up to which point the Lewis acidity could be
subsided without affecting catalytic performances. Indeed, it was later found that the first
reported formaldehyde trapping product originating from CO2 reduction could unexpectedly
promote catalysis through interactions as weak as hydrogen bonding.
In an effort to exploit this novel mild activation concept and employ it for the transformation
of waste into energetic CO2 reduction products, we discovered that commercially available
phosphazene bases were very efficient CO2 hydrosilylation catalysts. Depending on the
reaction conditions, either silylformates or methoxysilanes could be formed selectively for
the first time. This time, thorough investigation of the unusual solvent effects allowed us to
determine that DMF alone was able to promote CO2 hydrosilylation and we thus proposed
an alternative, plausible mechanism for such transformations.
In order to put things into perspective, the key discoveries in catalytic CO2 reduction
chemistry are summarized on a timeline in Figure 8-1. In 2006, the record for CO2 reduction
stood at 226 turnovers for the reduction of CO2 to methane by a zirconium catalyst. In the
years that followed, the same strategy of using highly Lewis acidic metal centers to activate
the reducing agent instead of CO2 was developed abundantly, allowing impressive activities
to be reached. Interestingly, the use of the alternate strategy of Lewis base activation led to
the formation of methanol derivatives instead. Simultaneously, novel systems were emerging
for the hydrogenation of CO2, but the high pressures and temperature that were required were
still far from ideal. It is really in 2013, with our original report of metal-free hydroboration
that the field of metal-free catalysis started evolving rapidly. Today, it is not surprising to
find systems that can sustain thousands of turnovers, even in the absence of transition-metal
catalysts.
200 Figure 8-1: Timeline of the important systems for CO2 reduction to energy rich material.
201 This rapid evolution led us to understand the key parameters involved in CO2 reduction, and
by assembling the knowledge that was obtained through careful study of these reduction
systems, three basic hypotheses were developed in order to fully exploit the potential of FLPs
in hydrogenation reactions. 1) The splitting of H2 needs to be energetically neutral (ΔG≈0)
to favor optimal transfer to the substrate. 2) The Lewis acidity/basicity in an FLP system
will dictate the electron distribution of the generated proton/hydride pair. 3) The electron
distribution of the generated proton/hydride pair must be complementary to that of the
substrate. In the case of CO2, a strong hydride needed to be generated.
By following these hypotheses and judiciously choosing the Lewis pair as to avoid CO2
coordination, a first generation of FLPs for CO2 hydrogenation were prepared. The
unexpected decomposition of the new molecules by protodeborylation interrupted catalysis,
resulting in stoichiometric hydrogenation of CO2. Nonetheless, the ability of the system to
promote such a difficult reaction at room temperature validated our approach. In recent work,
we have been able to circumvent the protodeborylation reaction and identify a promising
catalyst design which will be discussed in more details in the following section.
8.2 Ongoing and future work
8.2.1
Non-CO2 related chemistry
Starting with the hydrosilylation, the ability of DMF to promote CO2 reduction raises a
number of interesting questions that merit further investigation. In order to answer some of
these, a serious mechanistic investigations including a kinetic analysis of the reaction
components at various concentrations of DMF as well as a study of the effect of various
coordinating solvents (HMPA, DMSO, DMAC) and silanes on the reduction process should
be carried out. Further insight could evidently be obtained by simple DFT calculations.
As it was stated before, the scope of hydrosilylations in highly polar and coordinating
solvents could be expanded to aldehydes, ketones, esters and perhaps even further. An acute
understanding of the system would allow one to manipulate the reaction conditions to achieve
the desired reactivity. For instance, the transfer of more than simple hydrides could be
possible, opening new avenues for the formation of C-C bonds.
202 While we are currently developing ways to prevent the protodeborylation reaction in N-B
based FLPs, this unexpected decomposition pathway revealed a very interesting feature of
these particular FLPs. Closer inspection of the protodeborylation transition state unveils that
it is nothing but the reverse process of the C-H bond activation of an arene. Taking advantage
of this, we were able to tune the active groups of the FLP framework to generate the first
metal-free system for the catalytic C-H borylation of heteroarenes. This finding illustrates
that the concepts developed in this thesis go far beyond hydrogenation reactions and it is
most likely that many other metal-free processes can be engineered using simple, but
ingeniously designed FLP catalysts.
8.2.2
CO2 related chemistry
Moving on to the FLP mediated hydrogenation reactions, the work reported in chapter 5 is
but a first iteration of many future optimisations. Taking one step back, we directed our
efforts at identifying suitable substituents for the boron center by developing an aminoborane that would not decompose in the presence of the first CO2 reduction product: formic
acid. A closer analysis of the protodeborylation reaction revealed that the rate of this reaction
is directly proportional to the hybridization of the carbon substituent on boron, following the
trend sp>sp2>sp3, most likely due to an increasingly localized bonding orbital resulting from
a better overlap.
Using previously reported procedures, we have easily prepared a new FLP consisting of a
TMP nitrogen center bridged by an aryl group to a BBN moiety. This FLP shows practically
no sign of degradation upon heating at 80 °C in the presence of excess formic acid for several
hours (Scheme 8-1).
Scheme 8-1: Reversible splitting of formic acid by a novel, robust FLP system.
203 These results suggest that not only would the BBN framework remains intact in the presence
of the CO2 reduction products, but its low Lewis acidity would allow the delivery of a very
reactive hydride. Following the same logic, the lower Lewis acidity should also favor the
release of the reaction products: a win-win situation. Admittedly, the lower acidity of the
BBN framework as opposed to the reported aryl groups would mandate the use of a stronger
Lewis base to ensure efficient H2 splitting.
Just so, Bercaw and Labinger have already proven that the combination of BBN with a strong
phosphazene base results in the generation of a very reactive hydride after H2 splitting.170
Current work in collaboration with the Stephan group is thus aimed at the preparation of a
series of intramolecular phosphinimine/BBN systems with the goal of generating an active
catalyst for CO2 hydrogenation (Figure 8-2).
Figure 8-2: New FLP framework for C=O bond hydrogenation.
Simply put, the reduction of CO2 can be seen as the reduction of a rather unreactive C=O
bond. Therefore, the application of these metal-free hydrogenation systems are not limited to
carbon dioxide, but the reactivity could easily be extended to the hydrogenation of other
carbonyl compounds of interest such as aldehydes, ketones esters and perhaps even amides.
In point of fact, the developed intramolecular FLPs mimic the mode of action of the Noyori
asymmetric hydrogenation catalyst.320 Thus, it would not be surprising that upon
identification of an efficient ketone hydrogenation catalyst, the method could be extrapolated
to asymmetric reactions. In truth, the design of chiral Lewis acids is well documented and
with the help of DFT calculations, the energetic, steric and asymmetric parameters of such
systems could be fine-tuned rather effortlessly.
204 8.3 Philosophical discussion
8.3.1
Carbon dioxide
In order to fully appreciate the progress that has been made in CO2 reduction in the past few
years, let us take a few step backs and have a look at the evolution of the field. It was not
until the late 1970s that a transition metal complex of CO2 was characterized for the first
time. Despite the countless CO2 complexes that were reported between then and 2010, the
reports involving catalytic reduction of the small molecule remained incredibly scarce and
the activities were mediocre at best. Perhaps that this discrepancy between number of
reported complexes and number of catalytic systems could be somewhat rationalized by the
mindset that dominated the field at the time.
Indeed, in countless presentations, research articles, dissertation, books and publications, we
find an introduction that states more or less the following: CO2 is thermodynamically stable
and it is kinetically inactive. In truth, this exact introduction was the opening lines of my predoctoral examination. It seems, to me at least, that this general belief, that CO2 is such an
unreactive molecule, might have had an important role in hampering the development of
transition-metal based systems for the reduction of CO2. This obstruction may also have
carried onto the first years of FLP mediated CO2 activation, limiting the early development
of the field to stoichiometric transformations. In fact, the use of incredibly active Lewis acid
is not so different from the very reactive metal centers that were used at the time to trap this
‘‘inert’’ molecule.
Fortunately, the last five years have been witness to a tremendous progress in CO2 chemistry.
Reflecting on the past, I feel like I started on the wrong foot and so I would suppose that this
somewhat inaccurate vision of CO2 may have misled others before me. A saying from my
undergraduate organic teacher might be the best way to explain what I’m trying to say here
and I quote ‘‘It’s like trying to kill a fly with a bazooka!’’. What we’ve learned in the past
few years and throughout this thesis is that CO2 is not all that unreactive after all. Actually,
CO2 can easily be reduced with the help of very simple catalysts, or even with solvent! In
point of fact, it would seem we were simply not using tools that were adapted to this very
unique molecule.
205 The work presented herein represents a drafted version of one of many tools that will need
to be developed to generate efficient catalysts for future generations. If there is one thing that
I would like any reader to remember from this thesis is the idea that when it comes to CO2
reduction, the hydride dictates the reactivity. This simple concept was underlined from the
beginning, with the Walsh diagram presented in chapter 1: if one wants to transform the CO2
molecule, it must be bent and the easiest way to achieve this is by filling the π* orbital of the
carbon atom. The beauty of this concept lies in the fact that the stronger the hydride, the less
reactive the Lewis acid. This implies that not only a highly reactive reduction system would
necessarily release the reduction products readily, it would be highly robust and tolerant to a
wide range of additives and reaction conditions.
While Lewis acid catalyzed reductions are fundamentally interesting, the same inherent
problem plagues most of the systems: in a CO2 reduction process, the presence of the oxygen
atoms represents an issue since the use of highly oxophilic species will inevitably lead to
quenching of the Lewis acidity. Evidently, this problem can be circumvented by using
stoichiometric amounts of boranes or silanes that trap oxygen, but in order to generate
industrially relevant systems, we will have to move away from the stoichiometric use of these
reagents. Due to these limitations, it is my opinion that the use of oxophilic systems most
often involving strong Lewis acids have a limited applicability potential. Part of the solution
may lie in the use of carbon based Lewis acids which are not oxophilic despite being highly
Lewis acidic and witnessing the development of this emerging chemistry in the years to come
will doubtless be fascinating.
Another emerging tool which, in my opinion, holds great promise is the use of additive
reagents to trap intermediate carbon dioxide reduction products. As was discussed in chapter
1, one of the biggest problems associated with CO2 reduction is that formic acid is generated
after the first hydrogenation step. The problem lies in the fact that the formation of formic
acid by CO2 hydrogenation is thermodynamically uphill and thus the molecule has to be
trapped unless there is a very important driving force pushing the reaction towards the
products. In most systems, a base is used, but not only does this consume stoichiometric
amounts of base, it shuts down the reactivity towards further reduction.
206 The use of amines to trap formic acid is an interesting approach as the condensation of the
amine will ensure that the reaction is downhill and will also prevent formation of a reactive
formic acid molecule. While amines will inhibit reactivity in the presence of strong Lewis
acids, the use of weak Lewis acids which are good hydride generators should not be
problematic. Perhaps that a combination of both these tools could provide interesting avenues
towards the reduction of CO2. Amines are definitely not the only option as formic acid is a
very reactive and versatile molecule. For instance, a diol could be used to form an
intermediate acetal species or by dehydration, one could transform formic acid into formic
anhydride temporarily. Other options could include the stabilization of formic acid by
hydrogen bonding or extracting it using phase separation technologies.
All in all, many strategies are largely unexplored and others remain to be uncovered. The last
five years of CO2 chemistry have seen more progress than the 40 preceding years combined.
New strategies and tools are being developed as our understanding of the greenhouse gas is
evolving. While this thesis is only a piece of the puzzle, I am convinced that the years to
come will be as fruitful, if not more, as the past few years in terms of discoveries and
scientific advancements in the field of CO2 reduction.
207 8.3.2
Frustrated Lewis Pairs
The concepts developed herein are not limited to the reactivity of CO2, far from that. With
proper proton/hydride reactivity tuning, and with optimization of the bridging backbone, the
relative position of the proton and hydride can be controlled very precisely in order to
accommodate a wide range of substrates for hydrogenation. As such, we expect that alkynes,
alkenes, ketones, aldehydes, imines, nitriles and much more could by hydrogenated
effectively by intramolecular FLP systems.
Indeed, FLP chemistry can be seen as pure ligand design, where fine-tuning of the active
centers allows perfect complementarity to the substrate. The beauty of intramolecular FLP
systems is that one has absolute control over the reactivity of multiple active centers acting
synergistically. One has also complete control over the spatial distribution of the active
centers through the bridging framework and over the steric environment through the
substituents at the active centers. In a way, such precise control is somewhat reminiscent of
enzymatic chemical processes: simply put, one must construct a cavity containing cooperating active centers that can specifically accommodate the reagents, but not the products.
The design of such systems would have been incredibly time-consuming if they were being
developed a few decades ago, through trial and error. But with the advent of computational
chemistry, the precision of modern DFT calculations and the simplicity of FLP systems,
quick iterations between computations and experiments allows one to optimize the
parameters of a particular reaction in a very short timespan. As we have been able to show
with the catalytic borylation of C-H bonds, the reactivity is not limited to hydrogenation
reactions. In fact, the possibilities are limited to one’s imagination.
208 Metal-free catalysts are highly desirable in the food, pharmaceutical and polymer industries
in which the presence of trace metals is highly regulated by international instances. Other
advantages of metal-free systems, especially FLPs, is the low production cost and the
abundance of the elements that are used, making it a sustainable alternative. Being less than
10 years old, the field of FLP chemistry is still in at an infant stage. While most of the basic
reactivity has been explored using highly active Lewis pairs and stoichiometric reactions,
recent years have been witness to tremendous advances in FLP catalysis. In the last three
years, many new catalytic applications such as the hydrogenation of carbonyl groups and
alkynes, the catalytic reduction of CO2 and the catalytic borylation of heteroarenes have been
put forward for the first time.
With a deeper fundamental understanding of this chemistry at hand, it is doubtless that the
next ten years will see unprecedented advances in metal-free catalysis. I would even go as
far as to say that in analogy with organometallic chemistry which stemmed from scientific
curiosity only to become one of the most useful tools in chemistry a few decades later, the
golden age of metal-free catalysis may be close at hand.
209 8.4 Personal bibliogaphy
While some articles have been included entirely and discussed in detail, other contributions
were only briefly overviewed. As such, a list of publications and patents to which I
contributed during my Ph.D studies is listed below to provide the reader with more in-depth
discussions on the articles which were not fully included in this thesis.
1) Boudreau, J.; Courtemanche, M-A.; Fontaine, F-G. Reactivity of Lewis pairs
(R2PCH2AlMe2) with carbon dioxide. Chem. Commun. 2011, 47, 11131-11133.
2) Boudreau, J.; Courtemanche, M-A.; Marx,V.; Burnell, J.; Fontaine, F-G. Ambiphilic
molecules for trapping reactive intermediates: interrupted Nazarov reaction of allenyl vinyl
ketones with Me2PCH2AlMe2. Chem. Commun. 2012, 48, 11250 – 11252.
3) Courtemanche, M.-A.; Légaré, M.-A.; Maron, L.; Fontaine, F-G. A highly active
phosphine–borane organocatalyst for the Reduction of CO2 to methanol using hydroboranes
J. Am. Chem. Soc. 2013, 135, 9326-9329.
4) Courtemanche, M.-A.; Larouche, J.; Légaré, M.-A.; Wenhua, B.; Maron, L.; Fontaine, FG A tris(triphenylphosphine)aluminum ambiphilic precatalyst for the reduction of carbon
dioxide with catecholborane. Organometallics. 2013, 32 6804-6811. Invited contribution
5) Fontaine, F-G.; Courtemanche, M.-A.; Légaré, M.-A. Transition-metal-free catalytic
reduction of carbon dioxide. Chem. Eur. J. 2014, 20, 2990-2996. Invited contribution
6) Courtemanche, M.-A.; Légaré, M.-A.; Maron, L.; Fontaine, F-G. Reducing CO2 to
methanol using frustrated Lewis pairs: on the mechanism of phosphine-borane mediated
hydroboration of CO2. J. Am. Chem. Soc. 2014, 136, 10708-10717.
7) Légaré, M.-A.; Courtemanche, M.-A.; Fontaine, F-G. Lewis base activation of boranedimethylsulfide into strongly reducing ion pairs for the transformation of carbon dioxide to
methoxyboranes. Chem. Commun. 2014, 50, 11362-11365
8) Courtemanche, M.-A.; Légaré, M.-A.; Rochette, E.; Fontaine, F.-G. Phosphazenes:
efficient organocatalysts for the catalytic hydrosilylation of carbon dioxide Chem. Commun.
2015, 51, 6858-6861.
210 9) Declercq, R.; Bouhadir, G.; Bourissou, D.; Légaré, M.-A.; Courtemanche, M.-A.; Nahi,
K. S.; Bouchard, N.; Fontaine, F.G. and Maron, L. Hydroboration of carbon dioxide using
ambiphilic phosphine–borane catalysts: on the role of the formaldehyde adduct ACS
Catal. 2015, 5, 2513–2520.
10) Courtemanche, M.-A.; Pulis, A. P.; Rochette, R.; Légaré, M-A.; Stephan, D. W.;
Fontaine, F.-G. Intramolecular B/N frustrated Lewis pairs and the hydrogenation of carbon
dioxide. Chem. Commun. 2015, 51, 6858-6861
11) Rochette, E.; Courtemanche, M.-A.; Pulis, A. P.; Bi, W.; Fontaine, F.-G. Ambiphilic
frustrated Lewis pairs exhibiting high robustness and reversible water Activation : towards
the metal-free hydrogenation of carbon dioxide. Molecules, 2015, 20, 11902-11914.
12) Légaré, M.-A.; Courtemanche, M.-A.; Rochette, E.; Fontaine, F.-G. Metal-free catalytic
C-H bond activation and borylation of heteroarenes, Science, 2015, 349, 513-516.
Patents
1. Courtemanche, M.-A.; Fontaine, F.G. Catalysts for the Reduction of Carbon Dioxide. U.S.
Patent No. 14/259,733
2. Légaré M. A.; Fontaine, F.G.; Courtemanche, M.-A.; Process for the functionalization of
sp2 carbons. Patent pending
211 9 Bibliography
(1)
Ashton, T. S.; Hudson, P. The Industrial Revolution, 1760-1830, Oxford pap.; London, New
York, Toronto, 1997.
(2)
Partington, J. R. A Short History of Chemistry: Third Edition, Dover.; New York, 2011.
(3)
Kumar, K. modernization :: Population change | Encyclopedia Britannica
http://www.britannica.com/EBchecked/topic/387301/modernization/12022/Populationchange (accessed Apr 29, 2015).
(4)
Annual Energy Outlook 2015 - Energy Information Administration
http://www.eia.gov/forecasts/aeo/ (accessed May 5, 2015).
(5)
U.S. Census Bureau, Energy and Utilities; Washington, 2012.
(6)
Olah, G. A.; Molnár, Á. Hydrocarbon Chemistry; John Wiley & Sons, 2003.
(7)
Bengtsson, L. Geosphere-Biosphere Interactions and Climate; Cambridge University Press,
2011.
(8)
Monastersky, R. Nature 2013, 497, 13–14.
(9)
Canadell, J. G.; Le Quéré, C.; Raupach, M. R.; Field, C. B.; Buitenhuis, E. T.; Ciais, P.;
Conway, T. J.; Gillett, N. P.; Houghton, R. A.; Marland, G. Proc. Natl. Acad. Sci. U. S. A.
2007, 104, 18866–18870.
(10)
Hartmann, D.L., A.M.G. Klein Tank, M. Rusticucci, L.V. Alexander, S. Brönnimann, Y.
Charabi, F.J. Dentener, E.J. Dlugokencky, D.R. Easterling, A. Kaplan, B.J. Soden, P.W.
Thorne, M. W. and P. M. Z. In Observations: Atmosphere and Surface. In: Climate Change
2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment
Report of the Intergovernmental Panel on Climate Change; Stocker, T.F., D. Qin, G.-K.
Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. B. and P. M. M., Ed.;
Cambridge University Press: Cambridge, United Kingdom and New York, NY, USA, 2013;
pp 159–254.
(11)
White, C. M.; Smith, D. H.; Jones, K. L.; Goodman, A. L.; Jikich, S. A.; LaCount, R. B.;
DuBose, S. B.; Ozdemir, E.; Morsi, B. I.; Schroeder, K. T. Energy & Fuels 2005, 19, 659–
724.
(12)
Bottoms, R. R. Process for separating acidic gases. US1783901 A, December 1930.
(13)
Rochelle, G. T. Science 2009, 325, 1652–1654.
(14)
D’Alessandro, D. M.; Smit, B.; Long, J. R. Angew. Chem., Int. Ed. 2010, 49, 6058–6082.
(15)
Aresta, M. Carbon Dioxide Recovery and Utilization; Springer Science & Business Media,
2003.
(16)
Meessen, J. H. In Ullmann’s Encyclopedia of Industrial Chemistry; Wiley-VCH Verlag
GmbH & Co. KGaA, 2000.
(17)
Taherimehr, M.; Pescarmona, P. P. J. Appl. Polym. Sci. 2014, 131, 41141.
213 (18)
Chunshan Song; Anne F. Gaffney; Kaoru Fujimoto. CO2 Conversion and Utilization; Song,
C., Gaffney, A. F., Fujimoto, K., Eds.; ACS Symposium Series; American Chemical
Society: Washington, DC, 2002; Vol. 809.
(19)
Aresta, M.; Dibenedetto, A. Dalton Trans. 2007, 38, 2975–2992.
(20)
George A. Olah, Alain Goeppert, G. K. S. P. Beyond Oil and Gas: The Methanol Economy,
2nd, Updated and Enlarged Edition; Wiley-VCH, 2011.
(21)
Programme, U. N. E. The Hydrogen Economy: A Non-technical Review; UNEP/Earthprint,
2006.
(22)
Olah, G. A. Angew. Chem., Int. Ed. 2005, 44, 2636–2639.
(23)
George A. Olah: Hydrocarbons for the 21st Century - The Work of the Loker Hydrocarbon
Research Institute http://www.nobelprize.org/nobel_prizes/chemistry/laureates/1994/olaharticle.html (accessed May 5, 2015).
(24)
Olah, G. A. Chem. Eng. News Arch. 2003, 81, 5.
(25)
Hutchings, G. J.; Hunter, R. Catal. Today 1990, 6, 279–306.
(26)
Jacobson, M. Z. Energy Environ. Sci. 2009, 2, 148–173.
(27)
Palo, D. R.; Dagle, R. A.; Holladay, J. D. Chem. Rev. 2007, 107, 3992–4021.
(28)
Aresta, M.; Nobile, C. F.; Albano, V. G.; Forni, E.; Manassero, M. J. Chem. Soc. Chem.
Commun. 1975, 635–636.
(29)
Vol’pin, M. E. Pure Appl. Chem. 1972, 30, 607–626.
(30)
Aleksandrov, G. G.; Struchkov, Y. T. J. Struct. Chem. 1972, 12, 605–612.
(31)
Floriani, C.; Fachinetti, G. J. Chem. Soc. Chem. Commun. 1974, 615–616.
(32)
Castro-Rodriguez, I.; Meyer, K. J. Am. Chem. Soc. 2005, 127, 11242–11243.
(33)
Calabrese, J. C.; Herskovitz, T.; Kinney, J. B. J. Am. Chem. Soc. 1983, 105, 5914–5915.
(34)
Herskovitz, T. J. Am. Chem. Soc. 1977, 99, 2391–2392.
(35)
Aresta, M.; Nobile, C. F. J. Chem. Soc. Dalt. Trans. 1977, 708–710.
(36)
Mason, M. G.; Ibers, J. A. J. Am. Chem. Soc. 1982, 104, 5153–5157.
(37)
Dohring, a; Jolly, P. W.; Kruger, C.; Romao, M. J. Zeitschrift für Naturforsch. B 1985, 40,
484–488.
(38)
Aresta, M.; Nobile, C. F. Inorg.Chim. Acta 1977, 24, L49–L50.
(39)
Komiya, S.; Akita, M.; Kasuga, N.; Hirano, M.; Fukuoka, A. J. Chem. Soc. Chem. Commun.
1994, 1115–1116.
(40)
Sakamoto, M.; Shimizu, I.; Yamamoto, A. Organometallics 1994, 13, 407–409.
(41)
Tetrick, S. M.; Xu, C.; Pinkes, J. R.; Cutler, A. R. Organometallics 1998, 17, 1861–1867.
214 (42)
Tetrick, S. M.; Tham, F. S.; Cutler, A. R. J. Am. Chem. Soc. 1997, 119, 6193–6194.
(43)
Hanna, T. A.; Baranger, A. M.; Walsh, P. J.; Bergman, R. G. J. Am. Chem. Soc. 1995, 117,
3292–3293.
(44)
Audett, J. D.; Collins, T. J.; Santarsiero, B. D.; Spies, G. H. J. Am. Chem. Soc. 1982, 104,
7352–7353.
(45)
Kubiak, C. P.; Woodcock, C.; Eisenberg, R. Inorg. Chem. 1982, 21, 2119–2129.
(46)
Field, J. S.; Haines, R. J.; Sundermeyer, J.; Woollam, S. F. J. Chem. Soc. Chem. Commun.
1990, 984–985.
(47)
Field, J. S.; Haines, R. J.; Sundermeyer, J.; Woollam, S. F. J. Chem. Soc. Dalt. Trans. 1993,
2735–2748.
(48)
Tso, C. C.; Cutler, A. R. J. Am. Chem. Soc. 1986, 108, 6069–6071.
(49)
Vites, J. C.; Steffey, B. D.; Giuseppetti-Dery, M. E.; Cutler, A. R. Organometallics 1991,
10, 2827–2834.
(50)
Pinkes, J. R.; Steffey, B. D.; Vites, J. C.; Cutler, A. R. Organometallics 1994, 13, 21–23.
(51)
Senn, D. R.; Emerson, K.; Larsen, R. D.; Gladysz, J. A. Inorg. Chem. 1987, 26, 2737–2739.
(52)
Gibson, D. H.; Ye, M.; Sleadd, B. A.; Mehta, J. M.; Mbadike, O. P.; Richardson, J. F.;
Mashuta, M. S. Organometallics 1995, 14, 1242–1255.
(53)
Eady, C. R.; Guy, J. J.; Johnson, B. F. G.; Lewis, J.; Malatesta, M. C.; Sheldrick, G. M. J.
Chem. Soc. Chem. Commun. 1976, 601–602.
(54)
John, G. R.; Johnson, B. F. G.; Lewis, J.; Wong, K. C. J. Organomet. Chem. 1979, 169,
C23–C26.
(55)
Balbach, B. K.; Helus, F.; Oberdorfer, F.; Ziegler, M. L. Angew. Chem., Int. Ed. 1981, 93,
479–480.
(56)
Gambarotta, S.; Arena, F.; Floriani, C.; Zanazzi, P. F. J. Am. Chem. Soc. 1982, 104, 5082–
5092.
(57)
Fachinetti, G.; Floriani, C.; Zanazzi, P. F. J. Am. Chem. Soc. 1978, 100, 7405–7407.
(58)
Gibson, D. H. Coord. Chem. Rev. 1999, 185-186, 335–355.
(59)
Lundquist, E. G.; Huffman, J. C.; Folting, K.; Mann, B. E.; Caulton, K. G. Inorg. Chem.
1990, 29, 128–134.
(60)
Cokoja, M.; Bruckmeier, C.; Rieger, B.; Herrmann, W. A.; Kühn, F. E. Angew. Chem., Int.
Ed. 2011, 50, 8510–8537.
(61)
Mikkelsen, M.; Jørgensen, M.; Krebs, F. C. Energy Environ. Sci. 2010, 3, 43–81.
(62)
Miguel, C. V.; Soria, M. A.; Mendes, A.; Madeira, L. M. J. Nat. Gas Sci. Eng. 2015, 22, 1–
8.
(63)
Ashley, A. E.; Thompson, A. L.; O’Hare, D. Angew. Chem., Int. Ed. 2009, 48, 9839–9843.
215 (64)
Rusching, U.; Muller, U.; Willnow, P.; Hopner, T. Eur. J. Biochem. 1976, 70, 325–330.
(65)
Obert, R.; Dave, B. C. J. Am. Chem. Soc. 1999, 121, 12192–12193.
(66)
Aresta, M.; Dibenedetto, A. Greenhouse Gas Sinks, CABI Books.; Reay, D.; Hewitt, N.;
Grace, J.; Smith, K. A., Ed.; 2007.
(67)
Jeoung, J.-H.; Dobbek, H. Science 2007, 318, 1461–1464.
(68)
Huff, C. A.; Sanford, M. S. J. Am. Chem. Soc. 2011, 133, 18122–18125.
(69)
Eisenschmid, T. C.; Eisenberg, R. Organometallics 1989, 8, 1822–1824.
(70)
Matsuo, T.; Kawaguchi, H. J. Am. Chem. Soc. 2006, 128, 12362–12363.
(71)
Chakraborty, S.; Zhang, J.; Krause, J. A.; Guan, H. J. Am. Chem. Soc. 2010, 132, 8872–
8873.
(72)
Huang, F.; Zhang, C.; Jiang, J.; Wang, Z.-X.; Guan, H. Inorg. Chem. 2011, 50, 3816–3825.
(73)
Jessop, P. G.; Ikariya, T.; Noyori, R. Chem. Rev. 1995, 95, 259–272.
(74)
Jessop, P. G.; Joó, F.; Tai, C.-C. Coord. Chem. Rev. 2004, 248, 2425–2442.
(75)
Federsel, C.; Jackstell, R.; Beller, M. Angew. Chem., Int. Ed. 2010, 49, 6254–6257.
(76)
Himeda, Y.; Onozawa-Komatsuzaki, N.; Sugihara, H.; Kasuga, K. Organometallics 2007,
26, 702–712.
(77)
Tanaka, R.; Yamashita, M.; Nozaki, K. J. Am. Chem. Soc. 2009, 131, 14168–14169.
(78)
Onishi, N.; Xu, S.; Manaka, Y.; Suna, Y.; Wang, W.-H.; Muckerman, J. T.; Fujita, E.;
Himeda, Y. Inorg. Chem. 2015, 54, 5114–5123.
(79)
Fong, H.; Peters, J. C. Inorg. Chem. 2015, 54, 5124–5135.
(80)
Schmeier, T. J.; Dobereiner, G. E.; Crabtree, R. H.; Hazari, N. J. Am. Chem. Soc. 2011, 133,
9274–9277.
(81)
Kumar, B.; Llorente, M.; Froehlich, J.; Dang, T.; Sathrum, A.; Kubiak, C. P. Annu. Rev.
Phys. Chem. 2012, 63, 541–569.
(82)
Smestad, G. P.; Steinfeld, A. Ind. Eng. Chem. Res. 2012, 51, 11828–11840.
(83)
Burr Jr., J. G.; Brown, W. G.; Heller, H. E. J. Am. Chem. Soc. 1950, 72, 2560–2562.
(84)
Wartik, T.; Pearson, R. K. J. Am. Chem. Soc. 1955, 77, 1075–1075.
(85)
Pearson, Richard K. Wartik, T. Reaction products of NaBh4 and CO2. US2872474 A,
February 3, 1959.
(86)
Eisenberg, F.; Bolden, A. H. Carbohydr. Res. 1967, 5, 349–350.
(87)
Grice, K. A.; Groenenboom, M. C.; Manuel, J. D. A.; Sovereign, M. A.; Keith, J. A. Fuel
2015, 150, 139–145.
(88)
Zhao, Y.; Zhang, Z.; Qian, X.; Han, Y. Fuel 2015, 142, 1–8.
216 (89)
Schäfer, A.; Saak, W.; Haase, D.; Müller, T. Angew. Chem., Int. Ed. 2012, 51, 2981–2984.
(90)
Khandelwal, M.; Wehmschulte, R. J. Angew. Chem., Int. Ed. 2012, 51, 7323–7326.
(91)
Wehmschulte, R. J.; Saleh, M.; Powell, D. R. Organometallics 2013, 32, 6812–6819.
(92)
Huang, F.; Lu, G.; Zhao, L.; Li, H.; Wang, Z.-X. J. Am. Chem. Soc. 2010, 132, 12388–
12396.
(93)
N., L. G. Valence and the structure of atoms and molecules, Chemical C.; New York, 1923.
(94)
Brown, H. C.; Schlesinger, H. I.; Cardon, S. Z. J. Am. Chem. Soc. 1942, 64, 325–329.
(95)
Wittig, G.; Benz, E. Chem. Ber. 1959, 92, 1999–2013.
(96)
Roesler, R.; Piers, W. E.; Parvez, M. J. Organomet. Chem. 2003, 680, 218–222.
(97)
Chernichenko, K.; Nieger, M.; Leskelä, M.; Repo, T. Dalton Trans. 2012, 41, 9029–9032.
(98)
Welch, G. C.; San Juan, R. R.; Masuda, J. D.; Stephan, D. W. Science 2006, 314, 1124–
1126.
(99)
Stephan, D. W.; Erker, G. Angew. Chem., Int. Ed. 2010, 49, 46–76.
(100) Stephan, D. W.; Erker, G. Chem. Sci. 2014, 5, 2625–2641.
(101) Stephan, D. W. Acc. Chem. Res. 2014, 48, 306–316.
(102) Paradies, J. Angew. Chem., Int. Ed. 2014, 53, 3552–3557.
(103) Mömming, C. M.; Otten, E.; Kehr, G.; Fröhlich, R.; Grimme, S.; Stephan, D. W.; Erker, G.
Angew. Chem., Int. Ed. 2009, 48, 6643–6646.
(104) Berkefeld, A.; Piers, W. E.; Parvez, M. J. Am. Chem. Soc. 2010, 132, 10660–10661.
(105) Wen, M.; Huang, F.; Lu, G.; Wang, Z.-X. Inorg. Chem. 2013, 52, 12098–12107.
(106) Ménard, G.; Stephan, D. W. J. Am. Chem. Soc. 2010, 132, 1796–1797.
(107) Roy, L.; Zimmerman, P. M.; Paul, A. Chemistry 2011, 17, 435–439.
(108) Lim, C.-H.; Holder, A. M.; Hynes, J. T.; Musgrave, C. B. Inorg. Chem. 2013, 52, 10062–
10066.
(109) Ménard, G.; Stephan, D. W. Dalton Trans. 2013, 42, 5447–5453.
(110) Ménard, G.; Gilbert, T.; Hatnean, J. Organometallics 2013, 4416–4422.
(111) Fontaine, F.-G.; Zargarian, D. J. Am. Chem. Soc. 2004, 126, 8786–8794.
(112) Amgoune, A.; Bourissou, D. Chem. Commun. 2011, 47, 859–871.
(113) Kameo, H.; Nakazawa, H. Chem. Asian J. 2013, 8, 1720–1734.
(114) Kuzu, I.; Krummenacher, I.; Meyer, J.; Armbruster, F.; Breher, F. Dalton Trans. 2008,
5836–5865.
(115) Fontaine, F.-G.; Boudreau, J.; Thibault, M.-H. Eur. J. Inorg. Chem. 2008, 2008, 5439–5454.
217 (116) Dureen, M. A.; Stephan, D. W. J. Am. Chem. Soc. 2010, 132, 13559–13568.
(117) Appelt, C.; Westenberg, H.; Bertini, F.; Ehlers, A. W.; Slootweg, J. C.; Lammertsma, K.;
Uhl, W. Angew. Chemie Int. Ed. 2011, 50, 3925–3928.
(118) Zhao, X.; Stephan, D. W. Chem. Commun. 2011, 47, 1833–1835.
(119) Peuser, I.; Neu, R. C.; Zhao, X.; Ulrich, M.; Schirmer, B.; Tannert, J. A.; Kehr, G.; Fröhlich,
R.; Grimme, S.; Erker, G.; Stephan, D. W. Chem. Eur. J. 2011, 17, 9640–9650.
(120) Boudreau, J.; Courtemanche, M.-A.; Fontaine, F.-G. Chem. Commun. 2011, 47, 11131–
11133.
(121) Boudreau, J.; Fontaine, F.-G. Organometallics 2011, 30, 511–519.
(122) Thibault, M.-H.; Boudreau, J.; Mathiotte, S.; Drouin, F.; Sigouin, O.; Michaud, A.; Fontaine,
F.-G. Organometallics 2007, 26, 3807–3815.
(123) Karsch, H. H.; Appelt, A.; Koehler, F. H.; Mueller, G. Organometallics 1985, 4, 231–238.
(124) Goerigk, L.; Grimme, S. Phys. Chem. Chem. Phys. 2011, 13, 6670–6688.
(125) Wesselbaum, S.; Vom Stein, T.; Klankermayer, J.; Leitner, W. Angew. Chem., Int. Ed. 2012,
51, 7499–7502.
(126) Wesselbaum, S.; vom Stein, T.; Hintermair, U.; Klankermayer, J.; Leitner, W. Chemie Ing.
Tech. 2014, 86, 1428–1428.
(127) Wesselbaum, S.; Moha, V.; Meuresch, M.; Brosinski, S.; Thenert, K. M.; Kothe, J.; Stein, T.
vom; Englert, U.; Hölscher, M.; Klankermayer, J.; Leitner, W. Chem. Sci. 2015, 6, 693–704.
(128) Beydoun, K.; Ghattas, G.; Thenert, K.; Klankermayer, J.; Leitner, W. Angew. Chemie Int.
Ed. 2014, 53, 11010–11014.
(129) Beydoun, K.; vom Stein, T.; Klankermayer, J.; Leitner, W. Angew. Chem., Int. Ed. 2013, 52,
9554–9557.
(130) Park, S.; Bézier, D.; Brookhart, M. J. Am. Chem. Soc. 2012, 134, 11404–11407.
(131) Mitton, S. J.; Turculet, L. Chemistry 2012, 15258–15262.
(132) Berkefeld, A.; Piers, W. E.; Parvez, M.; Castro, L.; Maron, L.; Eisenstein, O. Chem. Sci.
2013, 4, 2152–2162.
(133) Bontemps, S.; Vendier, L.; Sabo-Etienne, S. Angew. Chem., Int. Ed. 2012, 51, 1671–1674.
(134) Neu, R. C.; Ménard, G.; Stephan, D. W. Dalton Trans. 2012, 41, 9016–9018.
(135) Harhausen, M.; Fröhlich, R.; Kehr, G.; Erker, G. Organometallics 2012, 31, 2801–2809.
(136) Jiang, C.; Stephan, D. W. Dalton Trans. 2013, 42, 630–637.
(137) Barry, B. M.; Dickie, D. A.; Murphy, L. J.; Clyburne, J. A. C.; Kemp, R. A. Inorg. Chem.
2013, 52, 8312–8314.
(138) Riduan, S. N.; Zhang, Y.; Ying, J. Y. Angew. Chem., Int. Ed. 2009, 48, 3322–3325.
218 (139) Courtemanche, M. A.; Larouche, J.; Légaré, M. A.; Bi, W.; Maron, L.; Fontaine, F. G.
Organometallics 2013, 32, 6804–6811.
(140) Rowlands, G. J. Tetrahedron 2001, 57, 1865–1882.
(141) Gröger, H. Chemistry 2001, 7, 5246–5251.
(142) Shibasaki, M.; Kanai, M.; Funabashi, K. Chem. Commun. 2002, 1989–1999.
(143) Ma, J.-A.; Cahard, D. Angew. Chem., Int. Ed. 2004, 43, 4566–4583.
(144) Yuan, Z.; Taylor, N. J.; Sun, Y.; Marder, T. B.; Williams, I. D.; Lap-Tak, C. J. Organomet.
Chem. 1993, 449, 27–37.
(145) James, T. D.; Sandanayake, K. R. A. S.; Shinkai, S. Angew. Chemie Int. Ed. 1996, 35, 1910–
1922.
(146) Agou, T.; Kobayashi, J.; Kawashima, T. Org. Lett. 2005, 7, 4373–4376.
(147) Zhu, L.; Zhong, Z.; Anslyn, E. V. J. Am. Chem. Soc. 2005, 127, 4260–4269.
(148) Agou, T.; Kobayashi, J.; Kawashima, T. Inorg. Chem. 2006, 45, 9137–9144.
(149) Lee, M. H.; Agou, T.; Kobayashi, J.; Kawashima, T.; Gabbaï, F. P. Chem. Commun. 2007,
1133–1135.
(150) Hudnall, T. W.; Kim, Y.-M.; Bebbington, M. W. P.; Bourissou, D.; Gabbaï, F. P. J. Am.
Chem. Soc. 2008, 130, 10890–10891.
(151) Sun, Y.; Wang, S. Inorg. Chem. 2009, 48, 3755–3767.
(152) Hudson, Z. M.; Wang, S. Acc. Chem. Res. 2009, 42, 1584–1596.
(153) Kim, Y.; Hudnall, T. W.; Bouhadir, G.; Bourissou, D.; Gabbaï, F. P. Chem. Commun. 2009,
3729–3731.
(154) Wade, C. R.; Gabbaï, F. P. Dalton Trans. 2009, 9169–9175.
(155) Rao, Y.-L.; Schoenmakers, D.; Chang, Y.-L.; Lu, J.-S.; Lu, Z.-H.; Kang, Y.; Wang, S.
Chemistry 2012, 18, 11306–11316.
(156) Boudreau, J.; Courtemanche, M.-A.; Marx, V. M.; Jean Burnell, D.; Fontaine, F.-G. Chem.
Commun. 2012, 48, 11250–11252.
(157) Chase, P. A.; Welch, G. C.; Jurca, T.; Stephan, D. W. Angew. Chem., Int. Ed. 2007, 46,
8050–8053.
(158) Stephan, D. W. Dalton Trans. 2009, 3129–3136.
(159) Grimme, S.; Kruse, H.; Goerigk, L.; Erker, G. Angew. Chem., Int. Ed. 2010, 49, 1402–1405.
(160) Dobrovetsky, R.; Stephan, D. W. J. Am. Chem. Soc. 2013, 135, 4974–4977.
(161) Grimmett, D. L.; Labinger, J. A.; Bonfiglio, J. N.; Masuo, S. T.; Shearin, E.; Miller, J. S. J.
Am. Chem. Soc. 1982, 104, 6858–6859.
(162) Hill, A.; Owen, G.; White, A.; Williams, D. Angew. Chem., Int. Ed. 1999, 38, 2759–2761.
219 (163) Parkin, G. Organometallics 2006, 25, 4744–4747.
(164) Hill, A. F. Organometallics 2006, 25, 4741–4743.
(165) Emslie, D. J. H.; Blackwell, J. M.; Britten, J. F.; Harrington, L. E. Organometallics 2006,
25, 2412–2414.
(166) Oakley, S. R.; Parker, K. D.; Emslie, D. J. H.; Vargas-Baca, I.; Robertson, C. M.;
Harrington, L. E.; Britten, J. F. Organometallics 2006, 25, 5835–5838.
(167) Miller, A. J. M.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc. 2008, 130, 11874–11875.
(168) Emslie, D. J. H.; Harrington, L. E.; Jenkins, H. A.; Robertson, C. M.; Britten, J. F.
Organometallics 2008, 27, 5317–5325.
(169) Owen, G. R.; Gould, P. H.; Hamilton, A.; Tsoureas, N. Dalton Trans. 2010, 39, 49–52.
(170) Miller, A. J. M.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc. 2010, 132, 3301–3303.
(171) Cowie, B. E.; Emslie, D. J. H.; Jenkins, H. A.; Britten, J. F. Inorg. Chem. 2010, 49, 4060–
4072.
(172) Zech, A.; Haddow, M. F.; Othman, H.; Owen, G. R. Organometallics 2012, 31, 6753–6760.
(173) Jana, R.; Blacque, O.; Jiang, Y.; Berke, H. Eur. J. Inorg. Chem. 2013, 2013, 3155–3166.
(174) Moret, M.-E.; Zhang, L.; Peters, J. C. J. Am. Chem. Soc. 2013, 135, 3792–3795.
(175) Anderson, J. S.; Moret, M.-E.; Peters, J. C. J. Am. Chem. Soc. 2013, 135, 534–537.
(176) Granville, S. L.; Welch, G. C.; Stephan, D. W. Inorg. Chem. 2012, 51, 4711–4721.
(177) Wagler, J.; Brendler, E. Angew. Chem., Int. Ed. 2010, 49, 624–627.
(178) Bontemps, S.; Gornitzka, H.; Bouhadir, G.; Miqueu, K.; Bourissou, D. Angew. Chem., Int.
Ed. 2006, 45, 1611–1614.
(179) Bontemps, S.; Bouhadir, G.; Miqueu, K.; Bourissou, D. J. Am. Chem. Soc. 2006, 128,
12056–12057.
(180) Sircoglou, M.; Bontemps, S.; Mercy, M.; Saffon, N.; Takahashi, M.; Bouhadir, G.; Maron,
L.; Bourissou, D. Angew. Chem., Int. Ed. 2007, 46, 8583–8586.
(181) Bontemps, S.; Sircoglou, M.; Bouhadir, G.; Puschmann, H.; Howard, J. a K.; Dyer, P. W.;
Miqueu, K.; Bourissou, D. Chemistry 2008, 14, 731–740.
(182) Sircoglou, M.; Bontemps, S.; Bouhadir, G.; Saffon, N.; Miqueu, K.; Gu, W.; Mercy, M.;
Chen, C.-H.; Foxman, B. M.; Maron, L.; Ozerov, O. V.; Bourissou, D. J. Am. Chem. Soc.
2008, 130, 16729–16738.
(183) Vergnaud, J.; Grellier, M.; Bouhadir, G.; Vendier, L.; Sabo-Etienne, S.; Bourissou, D.
Organometallics 2008, 27, 1140–1146.
(184) Bontemps, S.; Bouhadir, G.; Apperley, D. C.; Dyer, P. W.; Miqueu, K.; Bourissou, D. Chem.
Asian J. 2009, 4, 428–435.
220 (185) Sircoglou, M.; Mercy, M.; Saffon, N.; Coppel, Y.; Bouhadir, G.; Maron, L.; Bourissou, D.
Angew. Chem., Int. Ed. 2009, 48, 3454–3457.
(186) Sircoglou, M.; Bontemps, S.; Mercy, M.; Miqueu, K.; Ladeira, S.; Saffon, N.; Maron, L.;
Bouhadir, G.; Bourissou, D. Inorg. Chem. 2010, 49, 3983–3990.
(187) Malacea, R.; Saffon, N.; Gómez, M.; Bourissou, D. Chem. Commun. 2011, 47, 8163–8165.
(188) Derrah, E. J.; Sircoglou, M.; Mercy, M.; Ladeira, S.; Bouhadir, G.; Miqueu, K.; Maron, L.;
Bourissou, D. Organometallics 2011, 30, 657–660.
(189) Boone, M. P.; Stephan, D. W. J. Am. Chem. Soc. 2013, 135, 8508–8511.
(190) Porcel, S.; Bouhadir, G.; Saffon, N.; Maron, L.; Bourissou, D. Angew. Chem., Int. Ed. 2010,
49, 6186–6189.
(191) Baslé, O.; Porcel, S.; Ladeira, S.; Bouhadir, G.; Bourissou, D. Chem. Commun. 2012, 48,
4495–4497.
(192) Bebbington, M. W. P.; Bontemps, S.; Bouhadir, G.; Bourissou, D. Angew. Chem., Int. Ed.
2007, 46, 3333–3336.
(193) Moebs-Sanchez, S.; Saffon, N.; Bouhadir, G.; Maron, L.; Bourissou, D. Dalton Trans. 2010,
39, 4417–4420.
(194) Moebs-Sanchez, S.; Bouhadir, G.; Saffon, N.; Maron, L.; Bourissou, D. Chem. Commun.
2008, 3435–3437.
(195) Bontemps, S.; Bouhadir, G.; Dyer, P. W.; Miqueu, K.; Bourissou, D. Inorg. Chem. 2007, 46,
5149–5151.
(196) Sircoglou, M.; Bouhadir, G.; Saffon, N.; Miqueu, K.; Bourissou, D. Organometallics 2008,
27, 1675–1678.
(197) Appelt, C.; Slootweg, J. C.; Lammertsma, K.; Uhl, W. Angew. Chem., Int. Ed. 2012, 51,
5911–5914.
(198) Appelt, C.; Slootweg, J. C.; Lammertsma, K.; Uhl, W. Angew. Chem., Int. Ed. 2013, 52,
4256–4259.
(199) Courtemanche, M. A.; Légaré, M. A.; Maron, L.; Fontaine, F. G. J. Am. Chem. Soc. 2013,
135, 9326–9329.
(200) Erdmann, M.; Rösener, C.; Holtrichter-Rößmann, T.; Daniliuc, C. G.; Fröhlich, R.; Uhl, W.;
Würthwein, E.-U.; Kehr, G.; Erker, G. Dalton Trans. 2013, 42, 709–718.
(201) Robinson, G. H.; Lee, B.; Pennington, W. T.; Sangokoya, S. A. J. Am. Chem. Soc. 1988,
110, 6260–6261.
(202) Lee, B.; Sangokoya, S. A.; Pennington, W. T.; Robinson, G. H. J. Coord. Chem. 1990, 21,
99–105.
(203) Hitchcock, P. B.; Lappert, M. F.; Merle, P. G. Dalton Trans. 2007, 585–594.
221 (204) Chivers, T.; Eisler, D. J.; Fedorchuk, C.; Schatte, G.; Tuononen, H. M.; Boeré, R. T. Chem.
Commun. 2005, 3930–3932.
(205) Breit, N. C.; Ancelet, T.; Quail, J. W.; Schatte, G.; Müller, J. Organometallics 2011, 30,
6150–6158.
(206) Jones, C.; Junk, P. C.; Leary, S. G.; Smithies, N. A. Main Group Met. Chem. 2001, 24, 801–
802.
(207) Engelhardt, L. M.; Kynast, U.; Raston, C. L.; White, A. H. Angew. Chemie Int. Ed. 1987, 26,
681–682.
(208) Ashenhurst, J.; Brancaleon, L.; Gao, S.; Liu, W.; Schmider, H.; Wang, S.; Wu, G.; Wu, Q.
G. Organometallics 1998, 17, 5334–5341.
(209) Ménard, G.; Gilbert, T. M.; Hatnean, J. A.; Kraft, A.; Krossing, I.; Stephan, D. W.
Organometallics 2013, 32, 4416–4422.
(210) Westcott, S. A.; Blom, H. P.; Marder, T. B.; Baker, R. T.; Calabrese, J. C. Inorg. Chem.
1993, 32, 2175–2182.
(211) Atwood, D. A.; Jegier, J. A.; Liu, S.; Rutherford, D.; Wei, P.; Tucker, R. C. Organometallics
1999, 18, 976–981.
(212) Harder, S.; Brandsma, L.; Kanters, J. a.; Duisenberg, A.; van Lenthe, J. H. J. Organomet.
Chem. 1991, 420, 143–154.
(213) Bruker AXS Inc: Madison, WI 2007.
(214) Saint. Bruker AXS Inc: Madison, WI 2007.
(215) Sheldrick, G. M. Bruker AXS Inc: Madison, WI 2007.
(216) Sheldrick, G. M. Acta Crystallogr. A. 2008, 64, 112–122.
(217) Spek, A. L. Acta Crystallogr. Sect. D Biol. Crystallogr. 2009, 65, 148–155.
(218) Gaussian03; Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.;
Cheeseman, J. R.; Montgomery Jr., J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J.
M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.;
Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.;
Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.;
Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.;
Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V.
G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Ortiz, J. V; Cui, Q.; Baboul, A. G.; Clifford, S.;
Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin,
R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe,
M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A.
Gaussian Inc: Wellington, CT 2004.
(219) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
222 (220) Francl, M. M. J. Chem. Phys. 1982, 77, 3654–3665.
(221) Hehre, W. J.; Ditchfield, R.; Pople, J. a. J. Chem. Phys. 1972, 56, 2257–2261.
(222) Andrae, D.; Häußermann, U.; Dolg, M.; Stoll, H.; Preuß, H. Theor. Chim. Acta 1990, 77,
123–141.
(223) Bergner, A.; Dolg, M.; Küchle, W.; Stoll, H.; Preuß, H. Mol. Phys. 2006, 80, 1431–1441.
(224) Huang, K.; Sun, C.-L.; Shi, Z.-J. Chem. Soc. Rev. 2011, 40, 2435–2452.
(225) Sakakura, T.; Choi, J.-C.; Yasuda, H. Chem. Rev. 2007, 107, 2365–2387.
(226) Olah, G. A.; Goeppert, A.; Prakash, G. K. S. J. Org. Chem. 2009, 74, 487–498.
(227) Balaraman, E.; Gunanathan, C.; Zhang, J.; Shimon, L. J. W.; Milstein, D. Nat. Chem. 2011,
3, 609–614.
(228) Kleeberg, C.; Cheung, M. S.; Lin, Z.; Marder, T. B. J. Am. Chem. Soc. 2011, 133, 19060–
19063.
(229) Zhang, L.; Cheng, J.; Hou, Z. Chem. Commun. 2013, 49, 4782–4784.
(230) Pereira, S.; Srebnik, M. Organometallics 1995, 14, 3127–3128.
(231) Lang, A.; Knizek, J.; Noth, H.; Schur, S.; Thomann, M. Z. Anorg. Allg. Chem. 1997, 623,
901–907.
(232) Bontemps, S.; Sabo-Etienne, S. Angew. Chem., Int. Ed. 2013, 52, 10253–10255.
(233) Bontemps, S.; Vendier, L.; Sabo-Etienne, S. J. Am. Chem. Soc. 2014, 136, 4419–4425.
(234) Anker, M. D.; Arrowsmith, M.; Bellham, P.; Hill, M. S.; Kociok-Köhn, G.; Liptrot, D. J.;
Mahon, M. F.; Weetman, C. Chem. Sci. 2014, 5, 2826–2830.
(235) Leblanc, F. a.; Piers, W. E.; Parvez, M. Angew. Chem., Int. Ed. 2014, 53, 789–792.
(236) Barnett, B. R.; Moore, C. E.; Rheingold, A. L.; Figueroa, J. S. Chem. Commun. 2015, 51,
541–544.
(237) Lu, Z.; Wang, Y.; Liu, J.; Lin, Y.; Li, Z. H.; Wang, H. Organometallics 2013, 32, 6753–
6758.
(238) Wang, T.; Stephan, D. W. Chem. Eur. J. 2014, 20, 3036–3039.
(239) Das Neves Gomes, C.; Blondiaux, E.; Thuéry, P.; Cantat, T. Chem. Eur. J. 2014, 20, 7098–
7106.
(240) Wang, T.; Stephan, D. W. Chem. Commun. 2014, 50, 7007–7010.
(241) Schaub, T.; Paciello, R. A. Angew. Chem., Int. Ed. 2011, 50, 7278–7282.
(242) Federsel, C.; Boddien, A.; Jackstell, R.; Jennerjahn, R.; Dyson, P. J.; Scopelliti, R.;
Laurenczy, G.; Beller, M. Angew. Chem., Int. Ed. 2010, 49, 9777–9780.
(243) Jeletic, M. S.; Mock, M. T.; Appel, A. M.; Linehan, J. C. J. Am. Chem. Soc. 2013, 31,
11533–11536.
223 (244) Langer, R.; Diskin-Posner, Y.; Leitus, G.; Shimon, L. J. W.; Ben-David, Y.; Milstein, D.
Angew. Chem., Int. Ed. 2011, 50, 9948–9952.
(245) Huff, C. A.; Sanford, M. S. ACS Catal. 2013, 3, 2412–2416.
(246) Wehmschulte, R. J.; Saleh, M.; Powell, D. R. 2013, 2–9.
(247) Hounjet, L. J.; Caputo, C. B.; Stephan, D. W. Angew. Chem., Int. Ed. 2012, 51, 4714–4717.
(248) Takeuchi, K.; Stephan, D. W. Chem. Commun. 2012, 11304–11306.
(249) Theuergarten, E.; Schlösser, J.; Schlüns, D.; Freytag, M.; Daniliuc, C. G.; Jones, P. G.;
Tamm, M. Dalton Trans. 2012, 41, 9101–9110.
(250) Fontaine, F.-G.; Courtemanche, M.-A.; Légaré, M.-A. Chem. Eur. J. 2014, 20, 2990–2996.
(251) Gaussian09; Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.;
Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.;
Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J.
L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.;
Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery Jr., J. A.; Peralta, J. E.; Ogliaro,
F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.;
Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi,
M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.;
Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.;
Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.;
Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.;
Ortiz, J. V; Cioslowski, J.; Fox, D. J. Gaussian Inc: Wallingford, CT 2009.
(252) Chai, J.-D.; Head-Gordon, M. Phys. Chem. Chem. Phys. 2008, 10, 6615–6620.
(253) Chernichenko, K.; Madarász, A.; Pápai, I.; Nieger, M.; Leskelä, M.; Repo, T. Nat. Chem.
2013, 5, 718–723.
(254) Grimme, S. J. Comput. Chem. 2006, 27, 1787–1799.
(255) Marenich, A. V; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B 2009, 113, 6378–6396.
(256) Huang, D.; Makhlynets, O. V; Tan, L. L.; Lee, S. C.; Rybak-Akimova, E. V; Holm, R. H.
Inorg. Chem. 2011, 50, 10070–10081.
(257) Huang, D.; Makhlynets, O. V; Tan, L. L.; Lee, S. C.; Rybak-Akimova, E. V; Holm, R. H.
Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 1222–1227.
(258) Zhang, C.-G.; Zhang, R.; Wang, Z.-X.; Zhou, Z.; Zhang, S. B.; Chen, Z. Chemistry 2009,
15, 5910–5919.
(259) Liang, Y.; Liu, S.; Xia, Y.; Li, Y.; Yu, Z.-X. Chemistry 2008, 14, 4361–4373.
(260) Chen, Y.; Ye, S.; Jiao, L.; Liang, Y.; Sinha-Mahapatra, D. K.; Herndon, J. W.; Yu, Z.-X. J.
Am. Chem. Soc. 2007, 129, 10773–10784.
(261) Wang, H.; Liu, F.; Helgeson, R. C.; Houk, K. N. Angew. Chem., Int. Ed. 2012, 655–659.
224 (262) Miyashi, T.; Nishizawa, Y.; Fujii, Y.; Yamakawa, K.; Kamata, M.; Akao, S.; Mukai, T. J.
Am. Chem. Soc. 1986, 108, 1617–1632.
(263) Martin, R. L.; Hay, P. J.; Pratt, L. R. J. Phys. Chem. A 1998, 102, 3565–3573.
(264) DiMare, M. J. Org. Chem. 1996, 61, 8378–8385.
(265) Corey, E. J.; Link, J. O. Tetrahedron Lett. 1989, 30, 6275–6278.
(266) Corey, E. .; Bakshi, R. K. Tetrahedron Lett. 1990, 31, 611–614.
(267) Eisenberger, P.; Bailey, A. M.; Crudden, C. M. J. Am. Chem. Soc. 2012, 134, 17384–17387.
(268) Kabalka, G. W.; Baker, J. D.; Neal, G. W. J. Org. Chem. 1977, 42, 512–517.
(269) Declercq, R.; Bouhadir, G.; Bourissou, D.; Légaré, M.-A.; Courtemanche, M.-A.; Nahi, K.
S.; Bouchard, N.; Fontaine, F.-G.; Maron, L. ACS Catal. 2015, 5, 2513–2520.
(270) Beley, M.; Collin, J. P.; Ruppert, R.; Sauvage, J. P. J. Am. Chem. Soc. 1986, 108, 7461–
7467.
(271) Courtemanche, M.-A.; Légaré, M.-A.; Maron, L.; Fontaine, F.-G. J. Am. Chem. Soc. 2014,
136, 10708–10717.
(272) Wartik, T.; Pearson, R. K. J. Inorg. Nucl. Chem. 1958, 7, 404–411.
(273) Knopf, I.; Cummins, C. C. Organometallics 2015, 34, 1601–1603.
(274) Légaré, M.-A.; Courtemanche, M.-A.; Fontaine, F.-G. Chem. Commun. 2014, 50, 11362–
11365.
(275) Fujiwara, K.; Yasuda, S.; Mizuta, T. Organometallics 2014, 33, 6692–6695.
(276) Das Neves Gomes, C.; Jacquet, O.; Villiers, C.; Thuéry, P.; Ephritikhine, M.; Cantat, T.
Angew. Chem., Int. Ed. 2012, 51, 187–190.
(277) Jacquet, O.; Das Neves Gomes, C.; Ephritikhine, M.; Cantat, T. ChemCatChem 2013, 5,
117–120.
(278) Blondiaux, E.; Pouessel, J.; Cantat, T. Angew. Chem., Int. Ed. 2014, 53, 12186–12190.
(279) Jacquet, O.; Das Neves Gomes, C.; Ephritikhine, M.; Cantat, T. J. Am. Chem. Soc. 2012,
134, 2934–2937.
(280) Schwesinger, R.; Schlemper, H. Angew. Chemie Int. Ed. 1987, 26, 1167–1169.
(281) Schwesinger, R.; Willaredt, J.; Schlemper, H.; Keller, M.; Schmitt, D.; Fritz, H. Chem. Ber.
1994, 127, 2435–2454.
(282) Shintani, R.; Nozaki, K. Organometallics 2013, 32, 2459–2462.
(283) Motokura, K.; Takahashi, N.; Kashiwame, D.; Yamaguchi, S.; Miyaji, A.; Baba, T. Catal.
Sci. Technol. 2013, 3, 2392–2396.
(284) Jacquet, O.; Frogneux, X.; Das Neves Gomes, C.; Cantat, T. Chem. Sci. 2013, 4, 2127–2131.
(285) Cui, X.; Dai, X.; Zhang, Y.; Deng, Y.; Shi, F. Chem. Sci. 2014, 5, 649–655.
225 (286) Frogneux, X.; Jacquet, O.; Cantat, T. Catal. Sci. Technol. 2014, 4, 1529–1533.
(287) Li, Y.; Fang, X.; Junge, K.; Beller, M. Angew. Chem., Int. Ed. 2013, 52, 9568–9571.
(288) Huff, C. A.; Sanford, M. S. J. Am. Chem. Soc. 2011, 133, 18122–18125.
(289) Rezayee, N. M.; Huff, C. A.; Sanford, M. S. J. Am. Chem. Soc. 2015, 137, 1028–1031.
(290) Fiorani, G.; Guo, W.; Kleij, A. W. Green Chem. 2015, 17, 1375–1389.
(291) Ho, S. Y.-F.; So, C.-W.; Saffon-Merceron, N.; Mézailles, N. Chem. Commun. 2015, 51,
2107–2110.
(292) Fernández-Salas, J. A.; Manzini, S.; Nolan, S. P. Chem. Commun. 2013, 49, 9758–9760.
(293) Manas, M. G.; Sharninghausen, L. S.; Balcells, D.; Crabtree, R. H. New J. Chem. 2014, 38,
1694–1700.
(294) Volkov, A.; Tinnis, F.; Adolfsson, H. Org. Lett. 2014, 16, 680–683.
(295) Xie, W.; Zhao, M.; Cui, C. Organometallics 2013, 32, 7440–7444.
(296) Li, Y.; Molina de La Torre, J. A.; Grabow, K.; Bentrup, U.; Junge, K.; Zhou, S.; Brückner,
A.; Beller, M. Angew. Chem., Int. Ed. 2013, 52, 11577–11580.
(297) Zhang, L.; Nederberg, F.; Pratt, R. C.; Waymouth, R. M.; Hedrick, J. L.; Wade, C. G.
Macromolecules 2007, 40, 4154–4158.
(298) Maryanoff, B. E.; Reitz, A. B.; Mutter, M. S.; Whittle, R. R.; Olofson, R. A. J. Am. Chem.
Soc. 1986, 108, 7664–7678.
(299) Kilgore, U. J.; Basuli, F.; Huffman, J. C.; Mindiola, D. J. Inorg. Chem. 2006, 45, 487–489.
(300) Revunova, K.; Nikonov, G. I. Chemistry 2014, 20, 839–845.
(301) Lescot, C.; Nielsen, D. U.; Makarov, I. S.; Lindhardt, A. T.; Daasbjerg, K.; Skrydstrup, T. J.
Am. Chem. Soc. 2014, 136, 6142–6147.
(302) Das, S.; Bobbink, F. D.; Laurenczy, G.; Dyson, P. J. Angew. Chem., Int. Ed. 2014, 53,
12876–12879.
(303) Riduan, S. N.; Ying, J. Y.; Zhang, Y. ChemCatChem 2013, 5, 1490–1496.
(304) Boyer, J.; Corriu, R. J. P.; Perz, R.; Reye, C. Tetrahedron 1981, 37, 2165–2171.
(305) Corriu, R. J. P.; Perz, R.; Reye, C. Tetrahedron 1983, 39, 999–1009.
(306) Kobayashi, S.; Nishio, K. Tetrahedron Lett. 1993, 34, 3453–3456.
(307) Oshitari, T.; Tomita, M.; Kobayashi, S. Tetrahedron Lett. 1994, 35, 6493–6494.
(308) Molina, P.; Alajarin, M.; Arques, A. Synthesis (Stuttg). 1982, 1982, 596–597.
(309) Frustrated Lewis Pairs II; Erker, G., Stephan, D. W., Eds.; Topics in Current Chemistry;
Springer Berlin Heidelberg: Berlin, Heidelberg, 2013; Vol. 334.
226 (310) Frustrated Lewis Pairs I; Erker, G., Stephan, D. W., Eds.; Topics in Current Chemistry;
Springer Berlin Heidelberg: Berlin, Heidelberg, 2013; Vol. 332.
(311) Stephan, D. W.; Erker, G. Angew. Chemie Int. Ed. 2015, 6400–6441.
(312) Chernichenko, K.; Madarász, A.; Pápai, I.; Nieger, M.; Leskelä, M.; Repo, T. Nat. Chem.
2013, 5, 718–723.
(313) Mahdi, T.; Stephan, D. W. J. Am. Chem. Soc. 2014, 136, 15809–15812.
(314) Stephan, D. W.; Greenberg, S.; Graham, T. W.; Chase, P.; Hastie, J. J.; Geier, S. J.; Farrell,
J. M.; Brown, C. C.; Heiden, Z. M.; Welch, G. C.; Ullrich, M. Inorg. Chem. 2011, 50,
12338–12348.
(315) Xu, B.-H.; Kehr, G.; Fröhlich, R.; Wibbeling, B.; Schirmer, B.; Grimme, S.; Erker, G.
Angew. Chem., Int. Ed. 2011, 50, 7183–7186.
(316) Rokob, T. A.; Hamza, A.; Pápai, I. J. Am. Chem. Soc. 2009, 131, 10701–10710.
(317) Hounjet, L. J.; Bannwarth, C.; Garon, C. N.; Caputo, C. B.; Grimme, S.; Stephan, D. W.
Angew. Chem., Int. Ed. 2013, 52, 7492–7495.
(318) Moreau, M.; Matyjaszewski, K.; Sigwalt, P. Macromolecules 1987, 20, 1456–1464.
(319) Li, H.; Aquino, A. J. A.; Cordes, D. B.; Hung-Low, F.; Hase, W. L.; Krempner, C. J. Am.
Chem. Soc. 2013, 135, 16066–16069.
(320) Noyori, R. Angew. Chemie Int. Ed. 2002, 41, 2008–2022.
(321) Zimmerman, P. M.; Zhang, Z.; Musgrave, C. B. Inorg. Chem. 2010, 49, 8724–8728.
(322) Correa, A.; Martín, R. Angew. Chem., Int. Ed. 2009, 48, 6201–6204.
(323) Dalton, D. M.; Rovis, T. Nat. Chem. 2010, 2, 710–711.
(324) Sgro, M. J.; Stephan, D. W. Angew. Chem., Int. Ed. 2012, 11343–11345.
(325) Chakraborty, S.; Zhang, J.; Patel, Y. J.; Krause, J. A.; Guan, H. Inorg. Chem. 2013, 52, 37–
47.
(326) Courtemanche, M.-A.; Légaré, M.-A.; Rochette, É.; Fontaine, F.-G. Chem. Commun. 2015,
51, 6858–6861.
(327) Scott, D. J.; Fuchter, M. J.; Ashley, A. E. J. Am. Chem. Soc. 2014, 136, 15813–15816.
(328) Ménard, G.; Stephan, D. W. Angew. Chem., Int. Ed. 2011, 50, 8396–8399.
(329) Albrecht, K.; Kaiser, V.; Boese, R.; Adams, J.; Kaufmann, D. E. J. Chem. Soc. Perkin
Trans. 2 2000, 2153–2157.
(330) Devillard, M.; Brousses, R.; Miqueu, K.; Bouhadir, G.; Bourissou, D. Angew. Chemie Int.
Ed. 2015, 54, 5722–5726.
(331) Chernichenko, K.; Kótai, B.; Pápai, I.; Zhivonitko, V.; Nieger, M.; Leskelä, M.; Repo, T.
Angew. Chem., Int. Ed. 2015, 54, 1749–1753.
227 (332) Pelter, A.; Singaram, B.; Warren, L.; Wilson, J. W. Tetrahedron 1993, 49, 2965–2978.
(333) Reich, H. J.; Holladay, J. E.; Mason, J. D.; Sikorski, W. H. J. Am. Chem. Soc. 1995, 117,
12137–12150.
(334) Panda, T. K.; Gamer, M. T.; Roesky, P. W. Organometallics 2003, 22, 877–878.
228