Dinucleating Ligand Platforms Supporting Zinc and Indium

DINUCLEATING LIGAND PLATFORMS SUPPORTING ZINC AND INDIUM

CATALYSTS FOR STEREOSELECTIVE LACTIDE POLYMERIZATION

Alexandre Bertrand Kremer

B.Sc., Ecole Supérieur de Chimie, Physique et Electronique de Lyon, 2012

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF

THE REQUIREMENTS FOR THE DEGREE OF

MASTER OF SCIENCE

THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES

(Chemistry)

THE UNIVERSITY OF BRITISH COLUMBIA

(Vancouver)

July, 2016

Abstract

With the recent advances in cyclic ester polymerization to access biodegradable and bio-based plastic, an increasing number of metal-based initiators have been reported to mediate ring opening polymerization with high yield and stereoselectivity. However, the origin of the stereoselectivity and the factors affecting it are yet little understood. The role of the metal nuclearity for the polymerization of lactide has especially been subject to various speculations. In our group, we developed unique asymmetrically bridged indium initiators, explored their ring opening polymerization mechanism and found evidence for a type of tandem catalysis. In order to have a better understanding of the role of the two indium metal centres during polymerization, we aimed to synthesize dinucleating analogues.

The syntheses of the first alkoxide bridged indium complex supported by a chiral dinucleating ligand platform, along with its zinc analogue, are reported. Both complexes were synthesized in a one pot reaction starting from a chiral dinucleating bis(diamino)phenolate ligand platform, NaOEt, and respective metal salts. The dinucleating indium species bearing an achiral ligand backbone previously synthesized in our group was also further investigated. Both indium complexes catalyze the ring opening polymerization of racemic lactide to afford highly heterotactic PLA (Pr > 0.85). The indium complex bearing the achiral ligand backbone affords atactic PLA (Pr = 0.46) from meso-LA. The role of the dinucleating ligand structure in catalyst synthesis and polymerization activity is discussed.

Preface

The first chapter of this thesis was written entirely by myself with minor corrections from my supervisor Dr. Parisa Mehrkhodavandi. The second chapter was also written by myself with major contributions from my supervisor. This work was published in collaboration with Dr. Kim Osten and

Dr. Insun Yu who initiated this project and first isolated the indium complexes bearing achiral proligands. The synthesis, optimization and purification of the chiral proligands as well as the related complexes was developed entirely by myself. I completed all polymerization studies presented in this work. The single crystal structures presented in this work were obtained by Dr. Insun Yu, Dr. Kim Osten and myself. Dr. Insun Yu, Dr. Dinesh Aluthge and Tannaz Ebrahimi contributed to the solving of X-ray crystal structures. The second chapter of this thesis has been previously published in the American

Chemical Society Journal Inorganic Chemistry (DOI: 10.1021/acs.inorgchem.6b00358).

iii

Table of Content

Abstract ...... ii

Preface ...... iii

Table of Content ...... iv

List of Tables ...... vi

List of Figures ...... vii

List of Abbreviations and Symbols...... xiii

Acknowledgements...... xv

Chapter 1: General Introduction...... 1

1.1. Introduction to poly(lactic acid) ...... 1

1.2. Metal mediated ring opening polymerization mechanism of lactide ...... 5

1.3. Tacticity in PLA ...... 8

1.4. ‘PLA toward sustainability’ or ‘Towards sustainable PLA’? ...... 11

1.5. Defined bimetallic catalysts for the ROP of lactide ...... 15

1.5.1. Dimeric initiators ...... 15

1.5.2. Tethered initiators ...... 42

1.5.3. Dinucleating initiators ...... 50

1.5.4. Heterobimetallic complexes ...... 56

1.5.5. Conclusion ...... 59

1.6. Scope of the thesis ...... 61

Chapter 2: Dinucleating ligand platforms supporting zinc and indium catalysts for stereoselective lactide polymerization ...... 63

2.1. Introduction ...... 63

2.2. Synthesis and characterization of chiral proligands ...... 66

2.3. Synthesis of indium and zinc alkoxide complexes bearing chiral ligand backbones...... 68

2.4. Synthesis indium alkoxide complexes bearing achiral ligand backbone...... 78

2.5. Polymerization of lactide...... 81

2.6. Conclusions and perspectives ...... 83

2.7. Experimental procedures...... 85

References ...... 93

Appendices ...... 100

List of Tables

Table 2-1: Polymerization of lactide by initiators 2, 4, and 5...... 82 Table A-1: Bernoullian equations of probability obtained for the different tetrad present in PLA. .... 100 Table E-1: Selected bond lengths and angles for complexes 7a, 7b, and 8...... 125 Table E-2: Selected crystallographic parameters for complexes 4 and 5...... 127 Table E-3: Selected bond lengths and angles for complexes 2, 3, and 4...... 128 Table E-4: Selected crystallographic parameters for complexes 1, 7, and 8...... 130

List of Figures

Figure 1-1: Lifecycle of PLA from corn starch...... 2 Figure 1-2: Homoleptic initiators commonly used in industrial ROP of PLA, respectively tin(II)octanoate and aluminumisopropoxide...... 3

Figure 1-3: Coordination-insertion mechanism for the ROP of lactide...... 5

Figure 1-4: Activated monomer mechanism for ROP of lactide...... 6

Figure 1-5: Living polymerization versus immortal polymerization...... 7

Figure 1-6: Different microstructure of PLA arising from lactide polymerization...... 9

Figure 1-7: PLA presenting an mrr or iss tetrad (A), 1H{1H} NMR representative of PLA arising from rac-LA (B), and meso-LA (C), tetrad associated to each tacticity value (D)...... 10

Figure 1-8: Different classes of bimetallic catalysts ...... 15

Figure 1-9: Dinuclear systems based on monodentate ligand reported for ROP of lactide ...... 17

Figure 1-10: Dinuclear β-diiminate, β-diketiminate based complexes...... 19

Figure 1-11: β-enaminoketonate pyrazonolate zinc alkoxides, β-enaminoketonate zinc alkyls and β- diketonate aluminum alkoxides...... 20

Figure 1-12: Scorpionate and alkoxide imino zinc alkyl complexes...... 21

Figure 1-13: Diiminopyrrolide copper alkoxides reported by Schaper...... 22

Figure 1-14: P,O phosphinophenolate zinc alkoxide reported by Dagorne and Avilés...... 22

Figure 1-15: Zinc aryloxides and gallium alkoxide initiators supported by bidentate ligands ...... 23

Figure 1-16: N,N,O zinc and indium di(amino)phenolate alkoxide complexes ...... 24

Figure 1-17: N,N,O mono-methylether Salen-type and Schiff-base zinc and magnesium complexes. 26

Figure 1-18: N,N,O Schiff-base group IV alkoxides and O,N,O Schiff base iron complexes...... 27

vii

Figure 1-19: N,N,O β-ketiminate magnesium benzoxide and O,N,O β-ketiminate aluminum and zinc complexes...... 28

Figure 1-20: N,N,O pyrazonolate zinc and magnesium benzoxide reported by Lin...... 30

Figure 1-21: N,O,N aluminum benzoxide and O,E,O chalcogen diaryloxides titanium chloride...... 30

Figure 1-22: N,N,O biaryl based aluminum chloride complexes reported by Zi ...... 31

Figure 1-23: Triphenolate titanium alkoxide reported by Hofmeister...... 32

Figure 1-24: Phenoxy aminidate lanthanide amides and N,O,N diamido ether actinide alkoxides...... 33

Figure 1-25: SalBinap yttrium alkoxide reported by Coates ...... 33

Figure 1-26: SalBinap indium alkoxides and their different coordination mode...... 34

Figure 1-27: Salen-like yttrium lactate and chiral salen indium alkoxides ...... 35

Figure 1-28: Bis(phosphonic)diamido yttrium amides and alkoxides reported by Williams...... 36

Figure 1-29: Bis(phenolate) scandium lactate and indium alkoxide reported by Okuda...... 37

Figure 1-30: Bis(phenolate) trivalent and divalent lanthanide complexes...... 38

Figure 1-31: Bis(phenolate) lanthanide generated from bifunctional aryloates...... 39

Figure 1-32: Asymmetric bis(phenolate) indium and group IV complexes reported by Sun ...... 40

Figure 1-33: Oxo-bridged aminobis(phenolate) group IV alkoxides...... 41

Figure 1-34: Titanatrate complexes reported by Verkade...... 41

Figure 1-35: Pentadentate yttrium complexes...... 42

Figure 1-36: Calixarene and scorpionate zinc alkyls...... 43

Figure 1-37: Trisamido aluminum alkyls reported by Fontaine...... 44

Figure 1-38: Amino and imino bis(phenolate) aluminum alkyls ...... 45

Figure 1-39: Bis(phenoxyimine) and diphenylethylene salen group 13 alkyls ...... 46

viii

Figure 1-40: Di(salen) aluminum alkyls reported by Chen...... 47

Figure 1-41: Piperazidine-bridged bis(phenolate) lanthanide amides reported by Shen...... 48

Figure 1-42: Phenylenediamidinate lanthanide amides...... 49

Figure 1-43: Phenylenebis(β-ketiminate) lanthanide amides reported by Yao ...... 49

Figure 1-44: Heteroscorpionate magnesium alkyl complexes reported by Rodríguez...... 50

Figure 1-45: Di(amino) and di(imino)phenolate zinc, cobalt and magnesium alkoxides...... 51

Figure 1-46: Aminobis(pyrazolil) and scorpionate zinc complexes...... 52

Figure 1-47: Bis(Salen) magnesium and zinc alkoxides reported by Wu ...... 53

Figure 1-48: Bis(iminopyrrolide) zinc, mangnesium and aluminum alkoxides reported by Wang...... 53

Figure 1-49: Phenylenedi(amino)bis(phenolate) titanium alkoxide reported by Coates and Kol ...... 54

Figure 1-50: Group IV Salen alkoxides and related complexes...... 55

Figure 1-51: Binap-like yttrium lithium complex reported by Carpentier ...... 56

Figure 1-52: Heterometallic complexes comprising residual lithium centers...... 57

Figure 1-53: Nickel(II) lanthanide(III) Salen-type nitrates reported by Lü and Bao ...... 58

Figure 1-54: Heterobimetallic titanium zinc, magnesium, lithium and sodium alkoxides reported by

Li...... 59

Figure 2-1: Structure of the achiral ligands L1-3 synthesized and characterized by Kim Osten...... 63

Figure 2-2: Synthesis of [(L3)In2Cl4](μ-Cl) (1), [(L3)In2Cl4](μ-OEt) (2), and [(L3)In2Cl4](μ-OH) (3) established by Kim Osten and Insun Yu...... 64

Figure 2-3: Solid state structures of 2 (left) and 3 (right) obtained by single crystal X-ray diffraction by Kim Osten and Insun Yu respectively...... 65

Figure 2-4: Structure of the two possible enantiomers obtained from complexes 1, 2 and 3...... 66

Figure 2-5: Synthesis of proligands L4-7...... 67

Figure 2-6: Diastereoisomeric distribution in the synthesis of rac-L7 when starting with the racemic

1 trans-N,N-dimethylcyclohexyldiamine (top). H NMR spectra (300 MHz, CDCl3, 25 °C) of enantiopure RR/RR-L7, SS/SS-L7, rac-L7 and crossover experiments associated of the SS/SS-L7 +

RR/RR-L7 and rac-L7 + RR/RR-L7 (bottom)...... 68

Figure 2-7: Pathways for the synthesis of zinc and indium alkoxide complexes supported by L1-7...... 69

Figure 2-8: One pot syntheses of RR/RR-[(L7)In2Cl4](μ-OEt) (4) and RR/RR-[(L7)Zn2Cl2](μ-OEt)

(5)………………………………………………………………………………………………………70

Figure 2-9: Solid state structure of complex SS/SS-4 obtained by single crystal X-ray diffraction...... 71

Figure 2-10: Structure of the different chiral conformations of the chelating ring arising from cyclohexyldiamine and ethylenediamine……………………………………………………...…...... 87

Figure 2-11: Structure of the different isomers of RR/RR-4 and SS/SS-4 and their relationship...... 73

Figure 2-12: Synthesis of RR/RR-[(L7)Zn2Cl2](μ-OEt) (2) in a stepwise reaction...... 89

Figure 2-13: Attempted synthesis of zinc alkoxide complex with imine ligand RR/RR-L5...... 75

Figure 2-14: Solid state structures of complexes RR/RR-7 (top) and RR/RR-8 (bottom) obtained by single crystal X-ray diffraction...... 76

Figure 2-15: Structure of the zinc diastereomers 7a and 7b as observed by single crystal X-ray diffraction (top), and hypothetic structure of the indium analogue (bottom)...... 77

Figure 2-16: One-pot synthesis of complex 2...... 78

1 Figure 2-17: H VT NMR (400 MHz, toluene-d8) of complex 2 from −35 to 75 °C...... 79

Figure 2-18: Possible mechanism of equilibrium between the two enantiomers of complexes 1, 2 and

3………………………………………………………………………………………………………...80

Figure 2-19: Structure of a dinucleating indium complexes bearing asymmetric ligand backbone (A) and a dinucleating heterobimetallic indium zinc complexes (B)...... ………………..……..…99

1 Figure B-1: H NMR (400MHz, RT, CDCl3) of complex RR/RR-5 compared to rac-5 obtained from proligands respectively RR/RR-L5 and rac-L5...... 109 1 Figure B-2: H NMR (400MHz, RT, CDCl3) of respectively complex 7, mixture of products from the attempted alkoxilation and complex 8...... 110 1 Figure C-1: H NMR spectrum (400 MHz, CDCl3, 25 °C) of rac-L4...... 112

1 Figure C-2: H NMR spectrum (400 MHz, CDCl3, 25 °C) of RR/RR-L5...... 113

13 1 Figure C-3: C{ H} NMR spectrum (100 MHz, CDCl3, 25 °C) of (RR/RR)-L2...... 113

1 Figure C-4: H NMR spectrum (400 MHz, CDCl3, 25 °C) of rac-L6...... 114

13 1 Figure C-5: C{ H} NMR spectrum (150 MHz, CDCl3, 25 °C) of rac-L3...... 114

1 Figure C-6: H NMR spectrum (400 MHz, CDCl3, 25 °C) of RR/RR-L7...... 115

13 1 Figure C-7: C{ H} NMR spectrum (100 MHz, CDCl3, 25 °C) of (RR/RR)-L4...... 115

1 Figure C-8: H NMR spectra (300 MHz, CDCl3, 25 °C) of rac-L6 and crossover experiments associated of the SS/SS-L6 + RR/RR-L6 and rac-L6 + RR/RR-L6...... 116

1 Figure C-9: H NMR spectrum (400 MHz, CDCl3, 25 °C) of RR/RR-[(L7)In2Cl4](μ-OEt) (4)...... 117

13 1 Figure C-10: C{ H} NMR spectrum (100 MHz, CDCl3, 25 °C) of RR/RR-[(L7)In2Cl4](μ-OEt)

(4)……………………………………………………………………………………………………..117

1 Figure C-11: H NMR spectrum (400 MHz, CDCl3, 25 °C) of RR/RR-[(L7)Zn2Cl2](μ-OEt) (5)...... 118

13 1 Figure C-12: C{ H} NMR spectrum (100 MHz, CDCl3, 25 °C) of RR/RR-[(L7)Zn2Cl2](μ-OEt)

(5).………………………………………………………………………………………………...... 118

1 Figure C-13: H NMR spectrum (400 MHz, CDCl3, 25 °C) of RR/RR-[(L7)Zn2Cl2](μ-Cl) (6)...... 119

13 1 Figure C-14: C{ H} NMR spectrum (150 MHz, CDCl3, 25 °C) of RR/RR-[(L7)Zn2Cl2](μ-Cl)

(6)….………………………………………………………………………………………………….119

1 Figure C-15: H NMR spectrum (600 MHz, CDCl3, 25 °C) of RR/RR-[(L5)Zn2Cl2](μ-Cl) (7)...... 120

13 1 Figure C-16: C{ H} NMR spectrum (150 MHz, CDCl3, 25 °C) of RR/RR-[(L5)Zn2Cl2](μ-Cl)

(7)….………………………………………………………………………………………………….120

1 Figure C-17: H NMR spectrum (600 MHz, CDCl3, 25 °C) of [(L3)In2Cl4](μ-OEt) (2)...... 121

13 1 Figure C-18: C{ H} NMR spectrum (150 MHz, CDCl3, 25 °C) of [(L3)In2Cl4](μ-OEt) (2)...... 121

1 Figure D-1: H NMR of heterotactic PLA (600 MHz, CDCl3, 25 °C) obtained from rac-LA using complex 1 in CH2Cl2 at room temperature...... 122 1 Figure D-2: HNMR of atactic PLA (600 MHz, CDCl3, 25 °C) obtained from meso-LA using complex

7 in CH2Cl2 at room temperature...... 122

Figure D-3: MALDI-TOF mass spectra of PLA oligomers obtained using complex 2 in CH2Cl2 at room temperature with 20 equivalents of lactide...... 123 Figure D-4: MALDI-TOF mass spectra of PLA oligomers obtained using complex 2 at 100 °C in toluene with 40 equivalents of lactide...... 123 Figure E-1: Solid state structures of the two diastereoisomers present in complex 7, respectively 7a (top) and 7b (bottom)...... 124

xii

List of Abbreviations and Symbols

(+) dextrorotatory (-) levorotatory (±) racemic Ac acetate Ad adamantyl Binap 2,2’-bis(diphenylphosphino)-1,1’-binaphthyl Bn benzyl tBu tert-butyl CEM chain end control mechanism CIP Cahn-Ingold-Prelog 13C{1H} proton decoupled carbon Cm cumyl COSY correlation spectroscopy Cy cyclohexyl Ð dispersity Deap 3-diethylamino-1-propanoxy DFT density functional theory DME dimethylether DMSO dimethylsulfoxide ee enantiomeric excess Et ethyl i iso linkage ( = meso linkage) 1H{1H} proton decoupled proton HMPA hexamethylphosphoramide ki initiation rate kp propagation rate kt chain transfer rate LA lactide D-LA D-lactide ( = (R,R)-lactide) L-LA L-lactide ( = (S,S)-lactide) rac-LA 50:50 racemic mixture of D-lactide and L-lactide LCA life cycle assessment m meso linkage m- meta substituent MALDI-TOF matrix-assisted laser desorption time of flight mass spectroscopy Me methyl Myr (R)-1-[(R)-6,6-dimethylbicyclo[3.1.1]-2-hepten-2-yl]-2,2-bis(3,5- dimethylpyrazol-1-yl)ethoxide NMR nuclear magnetic resonance spectroscopy o- ortho substituent p- para substituent PET poly(ethyleneterephtalate)

xiii

Ph phenyl Pip 4-(1-benzylpipiridine) PLA poly(lactic acid) PLLA – PLDA poly(L-lactide) and poly(D-lactide) Pm probability of finding a meso linkage in a PLA chain Pr probability of finding a racemic linkage in a PLA chain iPr iso-propyl nPr propyl PS poly(styrene) r racemic linkage ROP ring opening polymerization RT room temperature s syndio linkage ( = racemic linkage) SCM site control mechanism Tg glass transition temperature THF tetrahydrofurane UV ultraviolet VT variable temperature  Relate to the coordination mode μ Indicate a bridging moiety

xiv

Acknowledgements

I am truly grateful to the Department of Chemistry for recruiting me. I have to admit that when I first applied to UBC, I never imagined that I would have the chance to get in, and today still, I am wondering how this happened.

My first thanks go to my supervisor Dr. Parisa Mehrkhodavandi. She was a demanding but fair mentor, always pushing me to give the best of myself. She made the chemist I am today and I am truly honoured that I have been able to work with her. Her passion for life and science will forever stay with me.

A huge thanks to the MehrCats, with a very special dedication to my mentors Dr. Paul Kelley, Dr.

Kim Osten, Dr. Dinesh Aluthge, Love-Ese Chile and Emiliya Mamleeva, who helped me constantly in the course of these two years and from whom I have learnt a great deal. I could not have achieved any of it without their patience, suggestions and presence. Thanks to the rest of the group too, for their friendship and support: Steve Chang, Tannaz Ebrahimi, Xiaofang Zhai, Carlos Andres Diaz Lopez and

Alan Wong. I would also like to thank the amazing undergraduate students I had the chance to work with, although their work is not presented here: Sabrina Zhang and Matt Milner. I am grateful I have been able to work surrounded by truly unique, passionate and fascinating people.

Then, I need to thank my family for their constant support, especially my girlfriend, Laura. My time here would have been much harder without her. Thank for following me in this crazy adventure and moving on the other side of the world with me – so far away from our families and friends. A dedication to my brothers and sisters too: Christelle, Aurélie, Arnaud and Thomas.

Finally, I want to thank all the precious people I have met in this department: friends, colleagues and professors. They made every second unique.

To my parents,

xvi

Chapter 1: General Introduction

Assessments of plastic impacts on the environment over the past decades have raised concerns over the end of life of plastics as organo-polymers,1-3 leading to a growing ambition to substitute fossil-based plastics, such as polyolefins, with bio-based or biorenewable biodegradable analogues at an industrial scale.4, 5 Whereas the annual production of conventional plastic reached over 300 MT in 20146 – and is still expected to keep growing in the next years – the development of waste disposal technologies has not yet bridged the gap and plastics accumulate into landfills but also in nature, especially oceans.1, 2

Plastic toxicity mainly arises from accumulation of non-biodegradable plastic nanoparticles and additives into food and water streams.2, 7, 8

In addition, in recent years, interests have shifted from fossil-based to bio-based resources.9 Although legislation to encourage the development of bio-based plastic has not been prevalent,10 the recent development of bio-based and biodegradable polyesters to substitute commodity plastic has emerged as a powerful solution for the next generation of sustainable plastics.4, 11 Research in the recent years have focused on developing cost and performance wise competitive catalysts for the polymerization of this new class of monomers, which is one of the keys for economically sustainable emergence of this new market.10, 12

1.1. Introduction to poly(lactic acid)

Poly(lactic acid) (PLA) is by far the most commonly produced biodegradable polyester with 205 kT in 2014 and the production is expected to double by 2017.10 NatureWorks LLC (an entity of Cargill and

PTT Global Chemical) is the pioneer and world leader producer of PLA with more than half of the market production.10

Industrial production of PLA is commonly performed from readily available corn (Figure 1-1) but can also be produced from rice or sugar beets.12 A recent communication from Corbion Purac – a Dutch pharmaceutical company producing lactide – shows PLA production from second generation feedstocks

– agricultural residues or waste - such as bagasse or corn stover.13 This would be a major improvement as corn starch-based PLA, which is competing directly with the food industry, is considered as a major ethical drawback in developing countries. NatureWorks LLC, with production lines in US and Thailand, is also looking in that direction and has received support from the US Department of Energy to extend ongoing research on the production of lactide from biomethane – methane obtained biologically.14

Figure 1-1: Lifecycle of PLA from corn starch.

Industrially, PLA is predominantly produced from the ring opening polymerization (ROP) of lactide, but can also be obtained directly from lactic acid via polycondensation.15 Uncontrolled molecular weight, in addition to the necessity of water removal by energetically costly distillation and solvent to reach high molecular weight, make polycondensation less attractive.16 Alternatively, metal catalyzed

2 ring opening polymerization produces highly controlled PLA via living polymerization but requires isolation of the lactide intermediate.

Lactic acid is obtained from corn sugar by bacterial fermentation; the natural abundance of D-glucose leads to the exclusive formation of L-(+)-lactic acid (ee > 99.5%).15 Oligomerization through polycondensation at high temperature under vacuum followed by metal-catalyzed depolymerization at reduced pressure affords lactide.15 This last step leads to partial racemization and formation of the three isomers of lactide: D-LA, L-LA and meso-LA. Purification of lactic acid and separation of the different lactide isomers represent a significant component of the production cost of PLA.15 Common impurities are acids, water and metal ions which decrease polymerization rates, accelerate hydrolysis and degrade physical properties in general.15 While meso-LA can be easily isolated by recrystallization, the chiral resolution of D-LA and L-LA required the use of chiral auxiliary. In industry, mixture of D-LA and L-

LA are often used as a starting material for PLA production. Due to the natural abundance of L-LA, a ratio 98:2 (L-LA:D-LA) is typically used in industry.15

PLA is obtained by ring opening polymerization of lactide using simple homoleptic initiators such as tin(II) octanoate or aluminum isopropoxide (Figure 1-2) to afford highly isotactic poly(L-lactic acid)

(PLLA) although no stereocontrol is typically observed during polymerization.17 This step is commonly performed neat – without solvent – at 140 °C and at low catalyst loading.17 Toxicity of tin as a metal have led scientists to focus on biocompatible metals such as K, Mg, Ca, Ti, Fe, Zn and Al for instance.18,

Figure 1-2: Homoleptic initiators commonly used in industrial ROP of PLA, respectively tin(II)octanoate and aluminum isopropoxide.

PLA is comparable to polystyrene (PS) and poly(ethyleneterephtalate) (PET) and aims to substitute them in most applications such as food packaging and fibers.20, 21 PLA biocompatibility and biodegradability make it attractive for sectors like agriculture4 and medicine, leading, in the pharmaceutical sector for example, to extensive research in the fields of nanoscale drug delivery systems for drugs, medical implants, beads for bones or tissue engineering, sutures and surgical fixtures.22, 23

Poly(L-lactide) (PLLA) - mostly produced in industry - is a semicrystalline thermoplastic polymer

24 24 susceptible to brittleness and stiffness. It has a low Tg, 50-60 °C and a melting point of 175 °C. As a biodegradable polymer, PLA is sensitive to heat and therefore difficult to process. As a result, PLA cannot be used for a wide range of applications comprised by ‘petro-polymers’.15

Ultimately, PLA can degrade to produce carbon dioxide and water.25 Contrary to common misconception, PLA does not fully degrade without the presence of heat (> 50 °C ), radiation such as

UV (315 nm) or/and humidity (> 30%) which initiate non-enzymatic depolymerization by hydrolysis of the ester bond.26 Optimal degradation conditions highlight a two-step process where hydrolysis is followed, under a certain molecular weight (< 10000 Da), by bacterial or enzymatic degradation which considerably increases degradation rate.25, 26 Complete degradation of PLA in the environment can take up to 2 years in presence of light and water, and down to 40 days under optimal conditions such as industrial composting.25, 26 However, PLA will take much more time to degrade in landfills or deep submarine environment where no light and heat is available. In contrast with fossil-based plastics, PLA does not produce harmful chemicals upon degradation.7 Lactide can also be copolymerized with more aliphatic cyclic ester such as ε-caprolactone to decrease crystallinity and to increase degradation rates.15

1.2. Metal mediated ring opening polymerization mechanism of lactide

Metal-mediated ring opening polymerization is by far the most studied route for the synthesis of PLA with hundreds of catalysts reported to date for the sole polymerization of lactide.18, 19 Discrete metal initiators comprise an electropositive metal center, an ancillary ligand and a initiating group such as an amide or an alkoxide. Metal complexes are attractive because they are easily tunable, highly active and highly stereoselective.18, 19 Recent years have seen the emergence of competitive organocatalysts and enzymatic systems,27 although this new class of compounds are interesting in their own right, they will not be discussed here. Metal-assisted ROP of lactide can undergo two distinct mechanisms: coordination insertion or activated monomer.

Figure 1-3: Coordination-insertion mechanism for the ROP of lactide.

The coordination insertion mechanism (Figure 1-3) was first described by Dittrish and Schulz28 for cyclic ester polymerization and can be assimilated to the related Cossee and Arlman mechanism first described for olefins polymerization.29 Two major steps are involved, first the coordination of the lactide electron rich carbonyl moiety to the metal center. Secondly, through the formation of a 4-membered

5 ring intermediate, the nucleophilic attack of the initiating group or later the active polymeryl moiety provokes the lactide ring to open by scission of the activated ester bond. As a result, the new monomer unit is inserted in the polymer chain as a new chain-extended metal alkoxide. Stabilization of the propagating species is postulated through the close proximity of the carbonyl moiety. The free coordination site is recovered and the newly formed complex can undergo the next coordination insertion reaction until monomer is no longer available. This mechanism allows a living polymerization, meaning that no termination mechanism occurs and direct correlation can be found between monomer concentration, catalyst loading and molecular weight.30 Subsequently, the polymer can be quenched to allow removal of the residual organometallic moiety; excess alcohol is typically used and substitutes the chain end.

The activated monomer mechanism (Figure 1-4) relies on Lewis acidity of the metal catalyst.31

Coordination of lactide will induce strong polarization and activation of the ester bond to allow a nucleophilic attack from an external hydroxyl-ended cocatalyst such as alcohol. Opening of the ring will occur to regenerate a new polymeryl cocatalyst for the next monomer addition.31 This polymerization mechanism is also living.

Figure 1-4: Activated monomer mechanism for ROP of lactide.

Although living polymerization is desirable, immortal polymerization (Figure 1-5) - firstly reported in 198732 - has caught much attention allowing even lower catalyst loading while conserving high control during polymerization.33 Immortal polymerization differs from living polymerization by the presence of a chain transfer agent, typically an alcohol - ethanol or benzyl alcohol for immortal polymerization of

33 lactide for example. If the propagation rate (kp) and the initiation rate (ki) are negligible compared to the chain transfer reaction rate (kt), the catalyst will undergo immortal polymerization which results in the formation of one polymer chain per equivalent ratio of the chain transfer agent over the initiator.33

In other words, each initiator can polymerize as much chain as there is excess of transfer agent, drastically decreasing the catalyst concentration. In contrast, in a living polymerization each initiator can only grow one unique polymer chain. The major drawback is the common instability of organometallic initiator in presence of large amount of transfer agent especially alcohols. The development of more tolerant initiator is a current area of research.33

Figure 1-5: Living polymerization versus immortal polymerization.

Lactide polymerization is not free of side reactions such as transesterification reactions leading to deep change in terms of microstructure. Transesterification can be either intermolecular or intramolecular.34 Intermolecular transesterification is the reaction of a growing polymer chain with an ester moiety of an adjacent polymer chain instead of a new monomer unit.34 In contrast, intramolecular 7 transesterifications also known as “back-biting” results from the reaction of a growing polymer chain with an ester moiety of its own chain.34 In that particular case, it leads to the formation of cyclic polymer chains. Matrix assisted laser desorption or ionization coupled with a time of flight mass spectrometry analyzer (MALDI-TOF) is a commonly used technique to confirm the presence of transesterification during polymerization.34 No transesterification results in a PLA sample where polymer chains are distributed evenly and 144 m/z separation, corresponding to one unit of lactide, can be observed between each peak on the mass spectrometer. The presence of peaks separated by 72 m/z corresponds to a lactic acid unit and will be characteristic of transesterification.34 Intermolecular and intramolecular transesterification can be differentiated by chain end analysis via 1H, 13C{1H}, COSY NMR or MALDI-

TOF, giving that a cyclic polymer has no chain end.34

1.3. Tacticity in PLA

As with all chiral based polymers, PLA physical properties are intimately related to tacticity.15 Thus, making highly active and stereoselective catalysts for ROP of lactide has been widely studied18, 19 in order to target one microstructure and tune the properties of the final product.

Starting from rac-LA (a 50:50 mixture of D-LA and L-LA), three different microstructures can be obtained: atactic, isotactic and heterotactic (Figure 1-6). Atactic PLA results from non-stereocontrolled polymerization to yield a randomly distributed polymer. Isotactic PLA results from polymerization of only L-LA or D-LA. Note that the use of rac-LA often leads to the formation of isotactic stereoblocks where one monomer is preferentially polymerized over the other, resulting in one poly(L-lactic acid)

(PLLA) block and one poly(D-lactic acid) (PDLA) block within the same chain. Finally, heterotactic

PLA arises from the alternating insertion of L-LA and D-LA.

The use of meso-LA instead can generate the same atactic and heterotactic microstructure in addition to syndiotactic PLA (Figure 1-6). Syndiotactic PLA results from the insertion of meso-LA at the same stereocenter at each insertion, to form a polymer with alternating stereocenters.

Figure 1-6: Different microstructure of PLA arising from lactide polymerization.

Stereocontrolled polymerization can arise from two distinct mechanisms: enantiomorphic site control mechanism or chain-end control mechanism (respectively reported in the literature as SCM and CEM).

A site control polymerization takes advantage of a chiral active site to generate stereocontrol, thus most site-control catalysts display chiral ligands. In contrast, in a chain-end control mechanism, stereocontrol arises from the polymer chain and thus from the last inserted chiral monomer. While it is often admitted that transesterification has a negative impact on stereocontrol due to redistribution of the polymer architecture, it is interesting to notice that, in some systems, high stereocontrol was reported along with transesterification reactions.118,119

Measurement of the stereoselectivity of PLA is commonly performed by 1H{1H} or 13C{1H} NMR.

Proton decoupled NMR is a powerful technique to assign each peak based on the shift by decoupling the methylene proton from the methyl protons. Each sequence of 4 stereocenters – also called tetrad - has been previously assigned35, 36 and stereocontrol is measured by statistical ratio of these peaks (Figure

1-7). The Tacticity is therefore defined as the probability of finding a meso (RR) or racemic (SR) linkage in the polymer chain. The notation iso (RR) and syndio (SR) is also used to describe these linkages.

Different stereoselectivity values arise: Pr (or Ps) to describe the probability of finding a racemic (or syndio) linkage, and Pm (or Pi) to describe the probability of finding a meso (or iso) linkage. Thus starting from meso-LA, syndiotacticity will be described by Pr and heterotacticity will be described by Pm.

However when starting from rac-LA, heterotacticity is now described by Pr and isotacticity by Pm. Due to the rise of stereoerrors overlapping with the desired peaks, the simplistic equations shown (Figure

1-7-D) below cannot be used. A model based on Bernouillan statistics is commonly accepted to access tacticity (Appendix A).37, 38 The analysis of tetrad can also give insight into the mechanism of stereoselectivity (SCM vs CEM) by comparing the ratio of the peaks resulting from stereoerrors. 37-39

Figure 1-7: PLA presenting an mrr or iss tetrad (A), 1H{1H} NMR representative of PLA arising from rac-LA (B), and meso-LA (C), tetrad associated to each tacticity value (D). 10

1.4. ‘PLA toward sustainability’ or ‘Towards sustainable PLA’?

This thesis started with an easy word –sustainability – which cannot be used without being justified.

It is therefore important to explain why and how this word comes along. Although some advantages inherent to the nature of PLA are obviously attractive, PLA has also been facing criticism and skepticism.40, 41 In this section, different concerns, along with life cycle assessments (LCA) will be reviewed to understand what really makes PLA sustainable … or not.

Before discussing the matter, it must be acknowledged that PLA’s success is based on competitiveness and economic viability, sustainability becoming a marketing feature or a company’s commitment. By being the first manufacturer to produce PLA on an industrial large scale, NatureWorks

LLC has proven that PLA was not only a case study but a real market opportunity with beneficial environmental repercussions.12, 42, 43 PLA production is expected to quadruple in the next 4 years,10 attesting a real potential that justifies academic interest.

PLA is most commonly mentioned for its biodegradability but was, ironically, not initially defined as biodegradable but as compostable. Recent advances in bio-compostable plastics in general have led to a new definition of biodegradability which was solely restricted to materials decomposing at ambient temperature.15, 41 In other words, PLA’s degradation in the environment can be slow although plastics residues will not be visible, and, as a matter of fact, will lead to confusion among unaware communities.4

As a result, PLA cannot be considered as a solution for worldwide plastic accumulation in the environment, oceans or landfills, as it requires specific conditions for complete degradation. It could even make it worse if consumers’ misconception of ‘biodegradability’ leads to an increasing amount of plastics purposely being released in the environment.4 However, it cannot be denied that plastic accumulation has been a powerful weapon to increase awareness of the impact of plastic’s end life and

11 waste in general, leading to an increasing demand to access more sustainable material with concrete end of life solution.1, 4

However, end of life solutions are not always synonym of direct sustainable impact. Polyolefins like

PET for example can be efficiently recycled3 but the truth is, most of it still end up in landfills. It is therefore legitimate to wonder, if PLA can be composted, will it be?26 Compostable PLA requires different facilities from the one recently implemented for recycling fossil-based plastics. Aside from the economic input required, biodegradable plastics will require crucial awareness among consumers to separately sort biodegradable and non-biodegradable plastics.44 Sorting plastics is a non-trivial challenge that industries are facing as structurally different plastics can be similar in aspect, color and density.45

The contamination of biodegradable plastics in recycling process and vice versa of non-biodegradable plastics in compost can lead to dysfunctions in both facilities.44, 45 Note that plastic recycling also leads to a decrease in plastic quality also known as “downcycling”. In addition, much of the current plastics are not recycled as consumers usually understand it, meaning that the plastic will be processed for another use. In most cases (in Europe especially), plastics are recycled through energy recovery facilities.3, 45 These facilities burns clean plastic waste to produce steam or electricity and therefore do not recycle the carbon that still accumulate in the atmosphere.3, 45

And this is where the main advantage of PLA comes: the low carbon footprint.4 Made from biorenewable resource, PLA has a considerable advantage in comparison to fossil-based polymer. In

PLA, all carbon atoms present arise from CO2 photosynthesis. Release of that same CO2 in the environment upon end of life will not lead to any increases in the atmosphere. PLA as all bio-based plastics is therefore called carbon neutral although it is not entirely true. Carbon emitted from energy consumption required to process PLA and its intermediates still impact its carbon footprint. Interestingly, recent studies have shown that in contrast with common plastics, PLA is usually locally sourced. Local

12 corn productions make transportation carbon footprint - which is a major factor when calculating carbon footprint - much lower than fossil-based polymer.46 However, one might wonder if mass production of

PLA will not lead to the same transportation issue with term.

Recent LCA analyses are all consistent to show that, although PLA production is energetically significant due to corn growth, lactic acid and lactide production and purification, the footprint of PLA materials are lower and they require less energy to produce than its fossil-based analogues.12, 42, 43, 47-51

Also the production of 1 kg of PLA requires from 1.3 to 1.5 kg of glucose and therefore about 2.5 kg of corn, which might seem a poor yield in contrast with fossil-based olefins but it is in fact a rather good score compared to bio-based polyolefins and other bio-based polymers.52 It must be noted that an LCA performed on PET bottles compared to PLA ones have shown that upon full recycling of PET bottles, these would be more sustainable than the PLA analogues.53 Note also that this same LCA is used by the

USA government to support their campaign on PET recycling.

Although biodegradable polymers are not suitable for recycling – as it is usually defined for polyolefins - they can undergo straightforward depolymerization to produce their monomer in good yield under control conditions as it was shown for poly(γ-butyrolactone).54 The use of LCA analyses have to be done with care, for example, they rarely take into account the use of additives which is a non- negligible part of plastic pollution and carbon footprints for both categories of plastics.51 LCA studies also widely rely on assumptions and criteria and therefore are difficult to compare, for this reason no data will be pulled out which might be irrelevant out of context. Finally, most LCA only focus on climate change and carbon footprint but other factors deeply related to bio-based plastic production can impact environment but are harder to evaluate such as biodiversity, land management, use of pesticide and geopolitics in general.51 Finally, it worth noting that readers have to pay attention to the authors who conduct the LCA, as their political and economic interests can bias the analysis.

The price of PLA remains an issue though, and to date, PLA remains a plastic for consumers and companies willing to pay the price of sustainability. It is also important to notice that fossil-based industry benefits from almost 100 years of industrial optimization, research and development leading to incredibly competitive price among plastic retailers. A lot more research in the different direction to access PLA and its intermediate lactide and lactic acid are therefore necessary to reach an equal amount of knowledge in the field. The introduction of competitors into the market to challenge NatureWork

LLC prices will also be a decisive factor that will drive research and development.

Another issue raised is the fact that PLA does not actually completely degrade to form CO2 and H2O, but leads to acidic intermediate such as lactic acid and other derivatives. Degradation studies show consistent acidification of soils upon PLA degradation. Although it is not clear what would be the effect of this acidification, it is usually considered far less harmful than common fossil-based by-products.7, 26

Finally, PLA and biodegradable polymers have been challenged by the rise of controversial oxo- biodegradable plastics.55 This class of polymer has been developed to accelerate the degradation of commonly non-biodegradable materials by addition biocompatible pro-oxidant or prodegradant.55 What seems to be a good compromise for industry has led to concerns over the nature of the actual degradation species and their toxicity which is little known to date.55 The ability of this class of polymer to be recycled is also a topic of discussion. Further research in this field will be necessary to say if this class of material already widely used for the production of plastic bags will find consensus.

PLA stays attractive as a symbol, a symbol of a new generation of plastic carbon neutral and biodegradable. In that sense, PLA is one of the numerous indicators of the society’s willingness to move away from fossil-based energy. Although, PLA sustainability is widely challenged, it is PLA’s potential which drives society interest to invest in this new generation of plastic.

1.5. Defined bimetallic catalysts for the ROP of lactide

Synthetic chemists have long taken inspiration from tandem catalysis, prevalent in nature, to develop multi-metallic systems in catalysis.56, 57 In particular, studies have explored cooperative bimetallic

58, 59 60, 61 mechanisms in polymerization catalysis, for example for polyolefins and CO2-based polymers.

The involvement of two metals in catalysis can take different forms (Figure 1-8): dinuclear or dimeric species arise from aggregation of two discrete metal centers through bridging ligands, tethered species involve two non-bridged metal centers on the same ligand architecture that can react independently, and dinucleating catalysts involve a multidentate ligand platform bound to two different metals, which are also bridged by a secondary ligand and can react in tandem. Bimetallic catalysts, either dimeric, tethered, or dinucleating, have also received attention as catalysts for the ring opening polymerization of lactide

(LA) to form PLA. In this part, we will review the different bimetallic species reported for PLA polymerization.

Figure 1-8: Different classes of bimetallic catalysts

1.5.1. Dimeric initiators

Dimeric species are the most widely reported bimetallic initiators for the ROP of lactide and are almost exclusively bridged by an alkoxy moiety. It does not come as a surprise that dimeric complexes have been widely encountered in the field as alkoxides species are by far the most efficient initiating group for ROP of lactide and easily bridge in the presence of Lewis acidic metal. Originally, it was

15 postulated that aggregation would slow down polymerization by generating an induction period where the complex would need to break up to form the mononuclear active species. However, after more than two decade of intense research in the field of ROP of lactide, it is clear that dinuclear propagating species can also carry out polymerization and are not necessarily slower, although the mechanism of ROP is greatly affected by the nature of the ancillary ligand and the metal center.

1.5.1.1. Monodentate ligands

Monodentate species are often the simplest system in the literature. Although in the early stage of research a wide range of homoleptic monodentate metals has been screened for ROP of lactide,17 there is no structural information regarding the nuclearity of these species. We will only report well characterized complexes exhibiting clear dinuclearity in solution or in the solid state.

The first studies of dinuclear catalysts with monodentate ligands were pioneered by Hillmyer, Tolman and co-workers with the study of the homoleptic iron alkoxide (I-1) in 2001.62, 63 Although no stereocontrol was reported, kinetic comparison with a monomeric analogue shows same order relative to the monomer but different order relative to the catalyst was obtained suggesting different

63 mechanisms. Calculation using a chain aggregation model shows that kp is 50 times higher for the dimeric catalyst.63 It was therefore postulated that interaction between poly(cyclic ester) and ancillary ligand can occur for the monomeric analogue and drastically inhibit propagation steps.63 Later on, the same group developed a simple three-component system based on indium(III) chloride, benzyl alcohol and trimethylamine.64, 65 The system is highly active and surprisingly stereoselective yielding heterotactic PLA (Pr > 0.9) from rac-LA via chain-end control mechanism although little stereoselectivity was observed as higher temperature (> 60 °C) or from meso-LA.64, 65 Mechanistic studies confirmed the living character of the system and revealed a dinuclear key intermediate (I-2-5).65

The system is believed to retain its dinuclearity during polymerization and allowing two-polymer chains

16 to grow from each benzyl alcohol initiator.65 The rate of polymerization is also shown to decrease with increasing size of the halide.65

Figure 1-9: Dinuclear systems based on monodentate ligand reported for ROP of lactide.62-68

More recently, Dagorne, Avilés and co-workers have reported a dinuclear zinc N-heterocyclic carbene alkoxide complex (I-6-7).66 Structurally characterized in the solid state to be dinuclear, computational study suggests that the structure is retained in solution and during polymerization with assistance of the second zinc to lower the energy barriers during ROP.66 Polymerization of lactide was controlled but not stereoselective. Immortal polymerization was also reported up to 10 equivalents of benzyl alcohol.66

Carpentier, Sarazin and co-workers reported a series of multi-nuclear lead species.67 Among them, the heteroleptic dinuclear lead alkoxide complex I-8 shows active and controlled ROP of lactide but once again not stereocontrol.67 Dissociation is believed to occur and polymerization is initiated by the

17 bridged alkoxy group, as no evidence was found for amide chain-ends by NMR spectroscopy. Rates are on the same order of magnitude than previously reported monomeric lead species.67

Finally, Mountford and co-workers reported the dinuclear indium alkoxide I-9, initially it was synthesized as a starting material to access a related alkoxide mononuclear species with a more elaborated ligand, but unsuccessfully as only the alkyl indium analogues were form upon elimination of the alcohol.68 This structure did not come as a surprise as indium alkoxides are known to aggregate.69

Non-characterized related amino indium species were also studied for immortal ROP of lactide in an analogous fashion as the Hillmyer and Tolman system described above, but in contrast an activated monomer mechanism is proposed.68 No stereocontrol was observed for these systems.68

1.5.1.2. Bidentate ligands

In 2001, alongside with the work of Hillmyer and Tolman described above, Coates and co-workers reported a series of highly active dinuclear zinc and magnesium β-diiminate alkoxide and acetate complexes (I-10-14) for the ROP of lactide.70 The zinc alkoxide species I-11 shows higher activity than the acetate analogue I-13 and highest stereocontrol to yield heterotactic PLA (Pr > 0.9) from rac-LA and syndiotactic PLA (Pr = 0.76) from meso-LA. In contrast, the magnesium complex I-14 exhibits no stereocontrol.70 The mechanism was unclear as the reaction was found to be first order with respect to the monomer but 1.5 to the catalyst, although controlled molecular weights and a narrow dispersity confirm that polymerization is living. The substituents on the β-diiminate ligand (I-10-12) were also found to have a great impact on selectivity and polymerization rate (iPr > Et > nPr).70 Interestingly mononuclear analogues of complex I-14 were further studied by Phomphrai and co-workers and exhibit high stereoselectivity in THF clearly highlighting a different mechanism.71

Figure 1-10: Dinuclear β-diiminate, β-diketiminate based complexes.70, 72-76

This class of ligand was further studied by Schaper and co-workers which reported several dinuclear

β-diketiminate copper, magnesium and zinc complexes.72-76 Zinc complex I-15 was active for rac-LA

72 ROP and yields heterotactic PLA (Pr = 0.84-0.87), whereas the magnesium (I-16) and copper (I-17-

18) analogues yield essentially atactic PLA with a slight isotactic and heterotactic bias respectively.73, 75

The square planar copper complexes I-17-18 undergo dimerization with the presence of less sterically demanding ligands.75 Interestingly, the dimeric species feature lower rate that their monomeric analogues for the immortal ROP.75 A dissociation mechanism is proposed for these species.74, 75

Similar β-enaminoketonate complexes were reported by different groups.77-79 The β-enaminoketonate pyrazonolate I-19 reported by Lin and co-workers was found to be dinuclear in presence of bulky substituted aryl while less bulky substituent leads to the mononuclear analogues.77 VT NMR reveals an equilibrium between I-19 and its monomeric form in solution and thus the authors postulated a dissociative ROP mechanism.77 The complex was highly active as expected, but poorly selective upon addition of rac-LA.77 Jones, Johnson and co-workers generated the less crowded β-enaminoketonate I-

20-21.78 These complexes were active in presence of benzyl alcohol and feature modest to good activity and poor stereoselectivity.78 Finally, similar zinc alkyl complexes I-22-24 investigated by Muños-

Hernández and co-workers feature better activity than their mononuclear analogues but still poor selectivity for the catalyst themselves.79 Addition of one equivalent of isopropanol to form the alkoxide in situ was found to significantly improve the activity, as expected.79

Figure 1-11: β-enaminoketonate pyrazonolate zinc alkoxides, β-enaminoketonate zinc alkyls and β- diketonate aluminum alkoxides.79, 80

The related β-diketonate aluminum alkoxide complexes (I-25-26) were reported by Carpentier and co-workers.80 Activities of these complexes are modest and ROP of rac-LA yields atactic PLA.80

Complexes were tolerant to 2-propanol and therefore immortal polymerization were conducted.80

Bidentate scorpionate (I-27-31) and alkoxide imino (I-32-35) zinc complexes were studied by

Rodríguez81 and Carpentier82 respectively. Interestingly, the scorpionate complexes were found to be mononuclear upon addition of ZnMe2 but dinuclear when larger zinc alkyls were used with bridging of the alkoxide moiety.81 Chain end analysis evidenced that the alkyl groups are initiating the polymerization and activity was found to increase with decreasing Zn-C bond strength (CH2SiMe3 >>

t 81 Et > Bu). Heterotactic-enriched PLA was obtained upon ROP of rac-LA with complexes I-29-31 (Pr

= 0.67-0.77).81

Figure 1-12: Scorpionate and alkoxide imino zinc alkyl complexes.81, 82

Dinuclearity of alkoxide imino analogues were found to be closely controlled by steric hindrance, for instance, complex I-32-33 would slowly decompose to yield the mononuclear homoleptic analogue and

82 release ZnEt2. These complexes yield atactic PLA and, in contrast with the scorpionate complexes described above, required the use of an external alcohol as initiating agent. Immortal polymerization was also studied upon higher equivalency of alcohol.82

Diiminopyrrolide copper alkoxides I-34-35 were reported by Schaper and co-workers.83 Interestingly, the use of chelating alkoxides yield the related dinuclear complexes in contrast with the tetradentate salen indium initiators that we will introduce later.83 More surprising is that complex I-34 was not

83 stereoselective for the ROP of rac-LA while I-35 yields isotactic-enriched PLA (Pm = 0.68).

Polymerizations with complex I-34 were carried out in presence of excess pyridine without increasing the stereoselectivity, whereas the use of half an equivalent of pyridine alcohol leads to similar stereoselectivity than I-35.83 In addition, complex I-35 features similar stereocontrol under immortal polymerization with benzyl alcohol used as a transfer agent.83 If a dissociative mechanism was undertaken, complex I-34-35 should display similar reactivity, thus this is the proof that a dinuclear propagating species carries the ROP where the polymer chain is growing on only one side of the dimer.83

The second bridge features the pyridine alcohol and allows stereocontrol.83 Analyses of stereoerrors suggests a chain end mechanism.83

Figure 1-13: Diiminopyrrolide copper alkoxides reported by Schaper and co-workers.83

Dagorne, Avilés and co-workers reported the P,O phosphinophenolate zinc alkoxide I-36.84 The complex is highly active but not stereoselective with a first order kinetic dependency with respect to the monomer.84 The dinuclear complex I-36 undergoes a coordination insertion mechanism whereas its mononuclear homoleptic analogue was found to follow an activated monomer mechanism.84 Complex

I-36 is stable and still active under immortal conditions.84

Figure 1-14: P,O phosphinophenolate zinc alkoxide reported by Dagorne, Avilés and co-workers 84

Bidentate zinc and magnesium aryloxides (I-37-40) were reported by Lin and co-workers.85, 86 The magnesium complex I-37 was found to be multinuclear and was broken down as dinuclear species in

86 coordinating solvent such as THF or Et2O. An external alcohol, in that case benzyl alcohol, was required to initiate polymerization and was observed to be activated by the Lewis acid metal center.85, 86

A dinuclear propagating species is postulated for the coordination insertion of LA.85, 86 Chain end analysis is consistent with benzyl alcohol initiating the polymerization and no evidence for the presence of the aryloxide ligand was found in the resulting polymer.85, 86 Increasing steric bulk for the zinc complexes leads to higher activity (I-40 > I-39 > I-38).85 Stereocontrol was not studied and only L-LA was used.

Figure 1-15: Zinc aryloxides and gallium alkoxide initiators supported by bidentate ligands.85-87

Finally, Kruk, Pécaut and co-workers studied ROP mechanism using the chiral and achiral dialkylgallium alkoxide I-41-42.87 The chiral complex I-41 yielded atactic PLA from rac-LA and increasingly heterotactic-enriched PLA (Pr = 0.60-0.78) upon addition of a Lewis base such as THF or

γ-picoline.87 It was found that the stronger the base, the more stereoselective the ROP.87 The use of the achiral analogue I-42 leads to same results highlighting a chain end control mechanism.87 Note that the use of the Lewis base did not decrease the overall activity of the catalyst suggesting that it does not compete with LA coordination. The authors found spectroscopic evidences for the presence of dinuclear gallium species during polymerization suggesting a dinuclear propagating species.87

1.5.1.3. Tridentate ligands

Tridentate ligands are most probably the most studied class of ligands in the literature for the ROP polymerization of lactide. The most famous example is from Hillmyer, Tolman and co-workers with a

N,N,O diaminophenolate zinc alkoxide (I-43-44) which is one of the most active catalysts reported for the ROP of lactide.88 The complex was dinuclear in the solid state, as determined by X-ray crystallography, but in solution the monomeric form seems to be prevalent.88 Kinetic studies showed first order dependencies with respect to the monomer and the catalyst confirming dissociation during polymerization and is consistent are with a coordination insertion mechanism.88 No stereocontrol was observed for these highly active zinc initiators.88 Immortal polymerization is reported using benzyl alcohol and does not affect rate of polymerization consistent with a fast exchange rate of the chain transfer agent.88

Figure 1-16: N,N,O zinc and indium di(amino)phenolate alkoxide complexes.88-97

Mehrkhodavandi and co-workers further studied this class of ligand by addition of a chiral moiety and generated the indium analogues and derivatives (I-45-59).89-97 Interestingly, this is the only one dimeric initiator exhibiting asymmetric bridges with only one initiating alkoxide, although the bis bridged alkoxide could also be obtained (I-53-54 and I-56).92 The initial complex I-45 exhibits the highest activity and yields modest isotactic-enriched PLA (Pm = 0.53-0.63) upon polymerization of rac-

LA.89 Further studies along with DFT calculations highlight a clear dinuclear propagating species during polymerization and therefore the growth of one bridged polymer chain per dinuclear complex.92, 97

Ligand design indicates that increasing halide size (I-45-47) decreases the overall activity (Cl > Br > I) while stereo bias is retained.90, 92 Increasing the steric bulk on the terminal or internal amine (I-48-49) provokes dissociation of the complex during polymerization as evidenced by the presence of the related dichloroindium analogue. Slightly higher rate was obtained for I-48 and loss of stereocontrol for both.93

Complex I-49 and I-58-59 evidenced the importance of the central amine where switching from a proton to a methyl leads to significant decrease in the activity regardless of the backbone.94 Substituents on the phenolate moiety were also examined (I-50-52), while I-51-52 behaved comparatively to the parent system (I-45), I-50 suggests dissociation mechanisms.96 Finally, the related benzyl alkoxides were generated (I-55-56) and highlight dissociation during polymerization in coordinating solvent like THF,

95 but retain dinuclear in less polar solvent like CH2Cl2. Finally, a unique 1,3,5-tris(hydroxymethyl)- benzene bridged complex (I-57) was generated and is the only one example of a one-component system for the synthesis of star-shaped polymer.95

Schiff-base N,N,O motif was widely studied by Lin and co-workers to generate zinc and magnesium benzoxide (I-60-80).98-101 The asymmetric mono-methylether Salen-type complexes I-60-61 highlight interesting mechanistic insights on the nature of the metal in ROP.98 Whereas the zinc complex I-60 shows first order dependency on monomer and catalyst and suggests a mononuclear propagating species,

25 the magnesium analogue I-61 exhibits first and second order for catalyst and monomer respectively suggesting a dinuclear propagating species.98 Upon polymerization of rac-LA, I-60 yields heterotactic- enriched PLA (Pr = 0.75) in CH2Cl2 while I-61 had a poor stereocontrol (Pr = 0.57) in THF, and switches

98 to yield isotactic-enriched PLA at low temperature in CH2Cl2 (Pm = 0.67).

Figure 1-17: N,N,O mono-methylether Salen-type and Schiff-base zinc and magnesium complexes.98- 102

A series of similar Schiff base N,N,O ligands were examined to yield the magnesium complexes I-

62-69.101 Substituents were shown to have a great effect on polymerization with I-63 exhibiting the highest activity.101 The decrease in activity was rationalized for withdrawing groups (such as Cl or Br) by increasing acidity of the magnesium center and thus increasing bond strength of the metal-alkoxide bond.101 The related zinc analogues (I-70-80) were also generated.99, 100 For complexes I-70-74, the same effects were observed than the ones described above for the magnesium analogues.99 However,

26 complexes I-75-80 were highly active polymerizing 100 equivalents of lactide within 5 min regardless of the substituents.100 These later zinc and magnesium complexes undergo the same ROP mechanisms that previously described analogues I-60-61.99-101 No stereocontrol was observed for the magnesium

99-101 whereas heterotactic-enriched PLA was obtained for the zinc analogues (Pr = 0.59-0.74).

The related Schiff-base zinc benzoxide I-81 was reported by Darensbourg and co-workers and exhibits a dimeric structure with an asymmetric coordination mode while its more bulkier chiral analogues are mononuclear.102 The complex was active but shows not stereocontrol, mechanistic insights confirm dissociation as probed by Lin and co-workers.102

N,N,O Schiff-bases were further investigated and the analogous group IV alkoxides I-82-90 were generated by Chand and co-workers.103 All complexes were highly active and yielded heterotactic-

103 enriched PLA (Pr = 0.62-0.80) from rac-LA. Stereocontrol and activity increased with increasing ionic radii of the metal center (Ti > Zr > Hf), but also with increasing steric hindrance on the aryl substituent (tBu > Me > H).103 Kinetic studies highlighted a first order dependency with respect to the monomer.103

Figure 1-18: N,N,O Schiff-base group IV alkoxides and O,N,O Schiff base iron complexes.103, 104

The dinuclear N,N,O iron complexes I-91-92 with iron centers featuring different ligand environment were reported by Li and co-workers.104 Whereas complex I-91 features higher activity that the related homoleptic iron complex, bulkier complex I-92 was similar.104 No stereocontrol was obtained for these complexes. The authors suggest a living mechanism involving a cationic intermediate.104

A series of N,N,O ketiminate magnesium benzoxide (I-93-96) developed also by Lin and co- workers.105 Ligand design highlights great effects of the substituent on the activity toward ROP (I-95 >

105 I-93 = I-94 >> I-96). The electron withdrawing group -CF3 probably leads to increase the Lewis acidity of the magnesium center increasing the strength of the Mg-OBn bond. Kinetic studies suggest a mechanism involving dissociation which was correlated by the higher activity with the bulkier tBu substituents that facilitate the formation of the mononuclear species.105

Figure 1-19: N,N,O β-ketiminate magnesium benzoxide and O,N,O β-ketiminate aluminum and zinc complexes.78, 105, 106

Analogous O,N,O β-ketiminate aluminum complexes (I-97-99) were generated by Clegg and co- workers.106 As expected these complexes were poorly active for ROP, however, upon addition of epoxide used as an external initiator they exhibit rather high activity.106 The author found evidences for the nucleophilic substitution of the epoxide at the metal center while the dinuclearity of the complex was

28 retained during the process.106 Note that complexes I-97-99 are not active for the polymerization of epoxides. No stereocontrol was observed for these complexes.

The O,N,O-type β-ketiminates were also investigated by Jones, Johnson and co-workers which generated the zinc analogues I-100-104.78 The complexes are exposed to “unpurified rac-LA” to mimic

78 industrial conditions and exhibit fair activity with slight heterotactic bias (Pr < 0.6).

Zinc and magnesium pyrazonolate benzoxides (I-105-126) were also examined by Lin and co- workers.107, 108 The zinc complexes I-105-110 were highly active, and can polymerize up to 6000 equivalents of L-LA.107 Kinetic studies are consistent with previously reported zinc species - first order dependencies in catalyst and monomer - suggesting dissociation.107 For the magnesium complexes (I-

112-126), however, substituents were found to have a greater effect on activity with electron donating

108 group on the position R3 and R4 enhancing ROP according to polymerization rate. Although these magnesium complexes were characterized dinuclear in solution and in the solid state, it is not clear if the dinuclearity is retained during polymerization. The authors suggest that it is in contrast with the ketiminate analogues, and proposed the decoordination of the tertiary amine to accommodate lactide coordination.108 Unfortunately, none of these complexes exhibit stereocontrol.107, 108 Interestingly, the substitution of the tertiary amine by a pyridine moiety (I-106 and I-111) leads to higher activity and

109 allows stereocontrol to yield heterotactic-enriched PLA (Pr = 0.60-0.87) from rac-LA .

Figure 1-20: N,N,O pyrazonolate zinc and magnesium benzoxide reported by Lin and co-workers.107- 109

Dagorne and co-workers investigated a series N,O,N aluminum and gallium alkyl complexes and could isolate the dinuclear aluminum complex I-127 but not the gallium analogue.110 The complex retains a dinuclear structure in solution and affords isotatic-enriched PLA (Pm = 0.62) from rac-LA with same activity as its mononuclear alkyl counterpart.110

Figure 1-21: N,O,N aluminum benzoxide and O,E,O chalcogen diaryloxides titanium chloride.110, 111

O,E,O chalcogen diaryloxide titanium chloride (I-128-129) were reported by Harada and co- workers.111 These complexes were modestly active and only L-LA polymerization was studied.111 A mononuclear analogue with one equivalent of monomer coordinated was isolated and exhibits same performance, clearly probing dissociation during ROP.111

N,N,O biaryl based aluminums (I-130-133) were investigated by Zi and co-workers.112 The complexes can only initiate the ROP of lactide in presence of propylene oxide to generate an alkoxide initiating group in situ as evidenced by NMR.112 Complexes I-130-133 exhibit similar modest activity

112 yielding heterotactic PLA (Pr = 0.64-0.72) from rac-LA. The use of THF was found to decrease significantly the activity, probably due to competition with lactide.112

Figure 1-22: N,N,O biaryl based aluminum chloride complexes reported by Zi and co-workers.112

Dinuclear tridentate titanium complexes were reported by Hofmeister and co-workers with a series of triphenolate titanium alkoxide (I-134-136).113 Complexes were also modestly active but yielded

113 heterotactic-enriched PLA (Pr = 0.80) via a chain end control mechanism. Sterics seem to have little effect on polymerization for this ligand structure. Kinetic studies suggest that a dissociative mechanism occurs.

Figure 1-23: Triphenolate titanium alkoxide reported by Hofmeister and co-workers.113

Finally, some phenoxy amidinate lanthanide amides (I-137-139) and diamido ether actinide alkoxides

(I-140-141) have been reported by Carpentier and co-workers.114, 115 The lanthanides complexes were highly active to yield broadly dispersed heterotactic-enriched PLA (Pr = 0.64-0.76), and can undergo immortal polymerization in presence of an external alcohol exhibiting narrower dispersity but loss of stereocontrol.115 The N,O,N actinide complexes I-140-141 also feature good activity with complex I-

141 being the most active and yielding heterotactic-enriched PLA (Pr = 0.61-0.73) upon addition of rac-

LA.114 Stereocontrol was found to be better in presence of a coordinating solvent such as THF but activity decreased due to a probable competition with lactide.114

Figure 1-24: Phenoxy aminidate lanthanide amides and N,O,N diamido ether actinide alkoxides.114, 115

1.5.1.4. Tetradentate ligands

Early report of dinuclear initiator bearing tetradentate ligand came from Coates and co-workers, in

2001, with a SalBinap yttrium alkoxide (I-142).38 In contrast with the mononuclear aluminum analogue,

I-142 features higher activity but not stereocontrol.38 Although the kinetics and mechanism of the mononuclear aluminum related complex were well studied, the yttrium was not further investigated due to its lack of stereocontrol.38

Figure 1-25: SalBinap yttrium alkoxide reported by Coates and co-workers.38

Mehrkhodavandi and co-workers further studied the role of aggregation in the related SalBinap indium alkoxides (I-143-144).116 Aggregation follows a similar pattern that the aluminum related

4 2 complexes with the presence of two complexes [( -O,N,N,O)In(OEt)]2 (I-143) and [(μ- -

116 O,N,N,O)In(OEt)]2 (I-144) which were separated and studied for ROP. Both complexes feature different reactivity with I-144 much slower and less controlled than I-143.116 Compared to the

33 mononuclear aluminum SalBinap analogues, both complexes exhibit poor activity and no stereocontrol.116

Figure 1-26: SalBinap indium alkoxides and their different coordination modes.116

Carpentier and co-workers reported the salen-like fluorous dialkoxy-diimino yttrium lactate I-145.117

The complex was highly active, yielded atactic PLA and was stable under immortal polymerization conditions when an external alcohol was used.117

Salen dinuclear indium alkoxides (I-146-150) were investigated by Mehrkhodavandi and co- workers.118, 119 These complexes were highly active compared to their mononuclear aluminum analogues and yielded stereogradient isotactic to isotactic-enriched PLA (Pm = 0.70-0.85) although extensive transesterification reactions occured.118, 119 tBu substituents were found to induce higher stereoselectivity than any others suggesting that a delicate steric balance is required for stereocontrol.119 Mononuclear analogues of complex I-146-150 were generating by using a bulky chelating alkoxide (O-CH2-Pyr) that prevent aggregations.119 Note that the less bulky chelating alkoxide used by Coates for complex I-141 did not prevent aggregation, same with the copper complexes reported by Schaper (I-34-35).38 Upon polymerization of lactide dinuclear complexes I-146-150 feature an initiation period not present for the mononuclear counterparts, but behave similarly in terms of propagation and stereocontrol, attesting that a dissociative mechanism takes place.119

Figure 1-27: Salen-like yttrium lactate and chiral salen indium alkoxides.117-119

A series of bis(phosphonic)diamido yttrium amides and alkoxides (I-151-160) were investigated by

Williams and co-workers.120-122Although the related bis(thiophosphonic)diamido yttrium amide was mononuclear, the dinuclear structure of complexes I-151-153 was relatively stable and no dissociation was observed in solution upon addition of THF or pyridine.120 Increasing the steric bulk on the phosphine, however, leads to the mononuclear amide complexes.121 In contrast, the alkoxide related complexes are all dinuclear, except when an aryloxide was used instead.121, 122 All complexes were highly active but poorly stereoselective from rac-LA affording heterotactic-enriched PLA (Pr = 0.61-

0.64) whereas the mononuclear amide and aryloxide counterpart were more stereoselective (Pr = 0.71-

0.85).120-122 ROP of lactide evidences that only one amide group is initiating polymerization and that the initiator is retaining its dinuclear structure.120 Possible explanation for the poor stereocontrol of the dinuclear species arises from the coordination of lactide which in this case can coordinate on either of the two stereo-different metal centers leading to two different chain end control mechanisms.121, 122

Increasing the steric bulk on the amide initiating group leads to slower initiation rates as evidenced with molecular weight twice higher than expected for complexes I-151-152, in contrast with I-153 which fully initiates polymerization.120 This was confirmed by the kinetic analysis of the related alkoxides, which are better nucleophile and thus does not exhibit this long induction period fully initiate.122 All reactions were found to have first order dependency in respect to the monomer.120, 122 Finally, complex

I-160 bearing ethylene backbone exhibits the highest activity compared to the 2,2-dimethylpropylene and phenylene backbones.122

Figure 1-28: Bis(phosphonic)diamido yttrium amides and alkoxides reported by Williams.120-122

In 2006, Okuda and co-workers reported a bis(phenolate) scandium lactate intermediate (I-161) isolated from the reaction of mononuclear scandium amides with lactate in THF.123 The complex exhibits high stereocontrol toward rac-LA (Pr = 0.93) as its mononuclear counterparts but rather low activity and high molecular weights. Later on, the same group also reported the indium related bis(phenolate) complexes.124, 125 Although, the amides and alkyls were mononuclear, the alkoxide complex I-162 was

36 dinuclear.124 Increasing the steric bulk on the aryl substituent leads to the mononuclear indium analogue which exhibits higher activity.124, 125 123 Thus, the authors believe that initiation mechanism for these scandium and indium complexes required a dissociation step leading to a mononuclear propagation species.123, 124 This mechanism was further confirmed by kinetic studies.125 No stereocontrol was observed with the bis(phenolate) indium alkoxide I-162 starting from rac-LA, however, highly

124, 125 syndiotactic PLA was obtained from meso-LA (Pr = 0.82-0.93).

Figure 1-29: Bis(phenolate) scandium lactate and indium alkoxide reported by Okuda and co- workers.123-125

Bis(phenolate) ligands were also investigated by Mountford and co-workers that reported a series of dinuclear lanthanide borohydride complexes (I-163-165).126, 127 Although, the borohydride groups were initially postulated to be bridging, later studies show that the dimer is actually formed by bridging oxygen from the bis(phenolate) ligand.126, 127 Upon addition of THF the dimer breaks to generate the mononuclear analogue.127 The complexes are modestly active featuring Sm > Nd > Y. I-163 exhibit the highest stereocontrol yielding heterotactic-enriched PLA (Pr = 0.87) in THF at RT, however, in toluene at 70 °C no stereocontrol was observed.126 Kinetic studies highlight a different mechanism depending on the metal center, with I-163-164 exhibiting first order dependencies in respect to the monomer whereas I-165 exhibits a second order dependency.126 Given that polymerization was undertaken in

THF, dissociation occurred; polymerization in toluene was not further investigated. A mechanism is proposed for the ROP of cyclic esters by borohydrides moieties. 127

Figure 1-30: Bis(phenolate) trivalent and divalent lanthanide complexes.126-129

Delbridge and Carpentier separately investigated the dinuclear divalent and trivalent respectively bis(phenolate) lanthanide complexes I-166-168.128, 129 Divalent dinuclear lanthanide complexes I-166-

167 were found to be in reversible equilibrium with their mononuclear counterpart in coordinating solvent such as THF.128 Increasing the bulk on the side amine donor moiety also generated mononuclear complexes.128 These complexes were modestly active for the polymerization of lactide and exhibit first order dependencies in respect to the monomer.128 Stereoselectivity was not studied. The related trivalent complex I-168 reported by Carpentier and co-workers turned out to be less active and stereoselective that its mononuclear counterparts, however it is not clear if this effect arises from the chloro-substituents, the dimeric structure or both.129

Interesting bis(phenolate) lanthanide analogues (I-169-174) where obtained from the use of bifunctional aryloate by Shen and co-workers.130 These complexes were highly active and highly stereoselective yielding heterotactic PLA (Pr = 0.81-0.99) with similar trends that their mononuclear

38 analogues.130 Therefore the author postulates that the metal center does not interact with each other during polymerization and the mechanism most likely involves a mononuclear propagating species.130

Figure 1-31: Bis(phenolate) lanthanide generated from bifunctional aryloates.130

Asymmetric bis(phenolate) ligands were also investigated and Sun and co-workers reported several complexes included a dinuclear indium water-bridged (I-175) and two group IV hydroxy-bridged complexes (I-176-177).131, 132 The indium complex is highly active and surprisingly tolerant with comparable reactivity under air and nitrogen.131 Stereocontrol was modest yielding heterotactic-enriched

131 PLA (Pr = 0.63-0.67) from rac-LA. Addition of one equivalent of alcohol improved significantly the control of the catalyst during polymerization.131 The mechanism is unclear but the authors suggest a possible dissociation pathway.131 The related group IV hydroxyl-bridged complexes I-176-177 were also highly active and yielded isotactic-enriched PLA (Pm = 0.65-0.72) from rac-LA with higher stereocontrol that their mononuclear symmetric bis(phenolate) analogues.132 Although the mechanism has not been investigated, these results highlight the importance of the dinuclear nature of the active site.132 The nature of the initiation step is also unclear.132 39

Figure 1-32: Asymmetric bis(phenolate) indium and group IV complexes reported by Sun and co- workers.131, 132

Chen and co-workers reported the oxo-bridged aminobis(phenolate) group IV alkoxides (I-178-

181).133 Whereas, the related non-oxo-bridged complexes are mononuclear, they can react with water to form these dinuclear oxo-bridged structures. Interestingly, the titanium complexes I-178-179 are inactive in contrast with their mononuclear alkoxide counterparts. However, the zirconium and hafnium complexes I-180-181 were active probably due to the better nucleophilicity of the alkoxide group, and exhibit similar activity as the related mononuclear complexes. Molecular weight analyses evidenced that both alkoxide groups initiate polymerization for complexes I-180-181. These complexes were also successfully used for immortal polymerization. Although very poor selectivity was reported in the melt, in solution complex I-180 can afford isotactic-enriched (Pm = 0.68-0.70) PLA from rac-LA.

Figure 1-33: Oxo-bridged aminobis(phenolate) group IV alkoxides.133

Finally a family of titanatrate complexes bearing alkoxide initiating group (I-182-183) was generated by Verkade and co-workers.134, 135 Although some mono-, di-, tri- and tetra-nuclear complexes were obtained the authors suggest that the nuclearity has no impact on the reaction control and mechanism.134,

135 Dinuclear complexes I-182-183 were modestly selective compared to their pairs.134, 135 The stereoselectivity was not studied for these complexes.

Figure 1-34: Titanatrate complexes reported by Verkade and co-workers.134, 135

1.5.1.5. Pentadentate ligands

The only one example in the literature of pentadentate dinuclear catalyst for ROP of lactide was reported by Hillmyer, Tolman and co-workers.136 The yttrium complexes generated feature chiral (I-

184-186) and achiral moieties (I-186). Complex I-186 differs from one coordinated water molecule on each yttrium center. All three complexes exhibit very different reactivity toward lactide with I-186 being significantly more active (I-186 > I-185 > I-184).136 I-186 also exhibits the highest fluxional behavior in solution according to VT NMR studies, although that might not be the rationale for the high activity which also can arise from the presence of water.136 No stereocontrol was observed for these complexes.136

Figure 1-35: Pentadentate yttrium complexes.

1.5.2. Tethered initiators

Tethered initiators for the ROP of lactide were far less studied with only few reports in the literature.

Early reports came from Vigalok and co-workers with a series of calixarene zinc alkyls (I-187-190).137

The complexes were modestly active and the alkyl group was found to initiate polymerization, although very slowly as a large amount of unreacted catalyst was found in solution.137 Also, methyl was found to be a superior initiating group than ethyl.137 Interestingly, spectroscopic evidence in addition to the synthesis of the cross-over complexes I-189-190 probed the unreactivity of the internal zinc center.137

Although it is not clear if the second metal plays a structural role during polymerization, it was clear that

42 it did not participate to the polymerization, and thus the catalyst is considered as a single site bimetallic catalyst.137 Stereocontrol was not assessed for those complexes.

Figure 1-36: Calixarene and scorpionate zinc alkyls.137, 138

Rodríguez and co-workers also reported a tethered asymmetric ( 3-N,N,O ;  1-O) zinc alkyl initiator

(I-191) for ROP but based on an N,N,O scorpionate structure.138 The complex was active for polymerization but not stereoselective and thus the author did not investigate this particular complex further.138

In 2010, Fontaine and co-workers reported the asymmetric ( 3-N,N,N ;  2-N,N) trisamido aluminum complexes I-192-193.139These complexes were modestly active without the presence of cocatalyst such as alcohols whereas the mononuclear counterpart generated by addition of only one equivalent of trimethylaluminum were not active.139However, reaction of complex I-193 with one equivalent of caprolactone suggests that no cooperativity takes place during polymerization, as the caprolactone chain is interacting with one of the amido groups.139 Thus, the second aluminum might only have a structural role.

Figure 1-37: Trisamido aluminum alkyls reported by Fontaine and co-workers.139

Tethered aluminum complexes were further studied by Wang and co-workers who generated a series of amino and imino bis(phenolate) aluminum alkyls (I-194-202).140 Although poor activity, control and initiation were observed for the complexes themselves. They behaved better upon addition of benzyl alcohol most probably generating an alkoxide in situ.140 The use of several equivalents of benzyl alcohol was also reported and characteristic of immortal polymerization.140 In absence of benzyl alcohol, it seems that imino complexes displayed only one active aluminum center while the amino displayed two probably due to different initiation rates.140 However in the presence of benzyl alcohol, it seems that in all cases, both aluminum centers were active and separately initiate polymerization.140 Overall amino vs imino complexes displayed similar reactivity but the substituent were greatly affecting the activity although no clear trend was observed.140 In the case of the imino complexes, bulky I-194 was the most active catalyst and I-195 the least active one, I-196-198 were found to have similar activity.140

Concerning the amino complexes, I-199-201 feature similar activity and I-202 is less active.140

Bimetallic complexes were compared to their mononuclear analogues and systematically feature higher activity suggesting that a cooperative mechanism takes place.140 Kinetic studies highlight a first order dependency in respect to the monomer.140

Figure 1-38: Amino and imino bis(phenolate) aluminum alkyls.140, 141

Ma and co-workers reported similar bis(phenolate) aluminum alkyls (I-203-205).141 The complexes were used in the presence of two equivalents of 2-propanol, to form in situ an alkoxide initiating moiety.141 In presence of four equivalents of 2-propanol, the polymerization rate was found to

141 i 141 significantly increase. Spectroscopic evidences support the formation of Al(O Pr)2 center. Thus, the authors suggest that each alkoxide is initiating.141 Activity was found to decrease with increasing steric bulk (I-203 > I-204 = I-205).141 Selectivity was poor in presence of two equivalents of 2-propanol

141 yielding isotactic-enriched PLA (Pm = 0.57-0.62) and decreased upon addition of four equivalents.

Similar imino bis(phenolate) aluminum and indium alkyl complexes (I-206-207) were reported by

Carpentier, Kirillov and co-workers.142 First, it must be noted that the scaffold, in this case, is much more rigid that the one reported by Wang and thus the rotational barrier around the aryl-aryl bond is higher.142 Catalytic activity in presence of 2-propanol was studied in comparison with the mononuclear analogues.142 The aluminum complex I-206 displayed same first order kinetic dependency in respect to the catalyst that its mononuclear counterpart but its activity is one order of magnitude higher.142 Thus it was postulated that a cooperative mechanism takes place. In contrast, the indium complex I-207 exhibits same activity and control that its mononuclear analogue and the two systems feature first order dependencies in respect to the monomer and the catalyst.142 Thus, in that case, it was concluded that no cooperative mechanism takes place and the two indium centers act as nearly identically to their

45 mononuclear analogues.142 The difference kinetic systems displayed by aluminum and indium initiators were believed to arise from different ROP mechanism.142 The authors suggest that unlike their aluminum analogues, the indium complexes cannot generate in situ the initiating alkoxide (due to poor reactivity) required for coordination-insertion and thus undergo an activated monomer mechanism.142

Figure 1-39: Bis(phenoxyimine) and diphenylethylene salen group 13 alkyls.142, 143

The related aluminum, gallium and indium diphenylethylene salen complexes I-208-213 were further investigated by Carpentier, Sarazin and co-workers.143 Steric bulk of the aryl substituent was found to have a great impact on the nuclearity of the complexes.143 While the gallium alkyl complexes are all dinuclear, the aluminum and indium alkyl analogues required bulky substituent typically tBu otherwise the tetradentate mononuclear analogue was obtained.143 On the other hand, aluminum alkoxide complexes and indium chloride reported with this same ligand backbone are mononuclear as their less bulky salen analogues.143 Note the indium complex I-213 is unreactive toward alcohol (like its mononuclear counterparts), unlike the dinuclear aluminum and gallium complexes.143 All dinuclear complexes exhibit similar polymerization rates compared to the monomeric analogues but they were not

143 143 stereoselective (Pm < 0.60 vs Pm = 0.90). Thus the authors postulated a site-control mechanism.

Complexes I-208 and I-213 required the use of an external alcohol to initiate polymerization.143 Gallium complexes I-210-212 were poorly controlled with or without alcohol.143

Chen and co-workers also investigated tethered bimetallic aluminum alkyls using di(salen) ligand platforms (I-214-221).144-146 Both series of complexes exhibit similar reactivity in presence of one equivalent of 2-propanol with first order dependencies with respect to the monomer and the catalyst.144-

146 For the first series (I-214-216), complexes exhibit similar activity, however steric and electronic of the substituents were found to have a great impact on selectivity yielding isotactic-enriched PLA from

t 144 rac-LA (Pm = 0.68-0.92, Bu > Me > Cl). Interestingly, for the second series (I-217-221), a more

t delicate steric balance was required as evidenced by the trend in activity (Me > H > Bu > SiMe3 >

t 145, 146 Si( Bu)(Me)2). Stereoselectivity on the other hand is almost identical to the first set (Pm = 0.65-

t t 145, 146 0.91, Si( Bu)(Me)2 > SiMe3 > Bu > Me > H). As expected, upon increasing temperature activity increases while stereoselectivity decreases.144-146 The authors suggest a chain end control mechanism due to achiral nature of the ligand.144-146

Figure 1-40: Di(salen) aluminum alkyls reported by Chen.144-146

Tethered piperazidine-bridged bis(phenolate) lanthanide amides (I-222-226) were reported by Shen and co-workers.147 As reported for the mononuclear analogues and other bis(phenolate) lanthanides, the activity increased with increasing ionic radii (I-225 > I-224 > I-223 = I-222), and overall dinuclear lanthanide complexes feature higher activity that their mononuclear analogues.147 However, this enhancement is most probably due to a less steric environment rather than a cooperative mechanism.147

Both mononuclear and dinuclear complexes follow a first order dependency in respect to the monomer also suggesting related mechanism.147 Both dinuclear and mononuclear complexes exhibit same

147 stereoselectivity yielding heterotactic-enriched PLA (Pr = 0.56-0.66) from rac-LA.

Figure 1-41: Piperazidine-bridged bis(phenolate) lanthanide amides reported by Shen and co- workers.147

Chen and Zhang simultaneously reported the use of phenyldiamidinate for the synthesis of bimetallic tethered lanthanide complexes (I-227-234).148, 149 Complexes I-227-231 were highly active toward L-

LA and activity increases with ionic radii as expected for these lanthanides complexes (La > Y > Lu >

Sc).149 However, complexes I-227 and I-231 have similar reactivity.149 Comparison with the mononuclear analogues reveals that no cooperative mechanism takes place between the two metal centers acting as two distinct centers for polymerization.149 In contrast, Zhang reported that complexes

I-232-234 have similar polymerization rate but twice higher than their mononuclear counterparts suggesting a synergetic effect of the two metal centers.148

Figure 1-42: Phenylenediamidinate lanthanide amides.148, 149

The related phenylenebis(β-ketiminate) lanthanides (I-235-238) were generated by Yao and co- workers and also feature very high activitypolymerizing up to 40 000 equivalents in 30 min and still obtain a fair control (PDI = 1.38-1.56).150 Once again, ionic radii is found to have a great impact on activity (Nd > Sm > Y).150 Complex I-238 has slight better activity than its analogue I-235 in THF but not in toluene.150 Stereoselectivity was also found to be solvent dependent with heterotactic-enriched

150 PLA (Pr = 0.65-0.72) obtained from rac-LA in THF, and atactic PLA in toluene.

Figure 1-43: Phenylenebis(β-ketiminate) lanthanide amides reported by Yao and co-workers.150

Finally, a series of ( 3-N,N,N ;  2-N,C) heteroscorpionate magnesium alkyl complexes featuring an apical Mg-C bond was reported by Rodríguez and co-workers.151 While the use of THF leads to the

49 complexes I-239-240 as drawn, the use of 1,4-dioxane yields the dimeric tetranuclear magnesium analogues.151 Complexes were highly active and more efficient that the mononuclear analogues generated with only one equivalent of Grignard.151 Dinuclear complexes were also found to be faster than the tetranuclear ones, probably due to higher donating character of 1,4-dioxane compared to THF.151

Chain end analysis showed that the initiation occurs via the alkyl moiety.151 Tetranuclear complexes feature higher stereoselectivity yielding heterotactic-enriched PLA (Pr = 0.75-0.78) compared to

151 dinuclear analogues (Pr = 0.68-0.71).

Figure 1-44: Heteroscorpionate magnesium alkyl complexes reported by Rodríguez.151

1.5.3. Dinucleating initiators

In 2002, Hillmyer, Tolman and co-workers reported the first dinucleating initiator (I-241) for the

ROP of lactide.152 The di(amino)phenolate zinc monoalkoxide is highly active, but unfortunately not stereoselective.152 Kinetic studies indicate first order dependencies with respect to the monomer and the catalyst.152 Further investigations showed that the dimeric analogue (I-43) exhibits propagation rates six times higher.88 The authors proposed that the presence of the chloride moiety at the zinc centers decreases its Lewis acidity and thus lowers its activity.88 The magnesium and cobalt analogues were also generated (I-244-245) but exhibit lower activity - with propagation rate 90 times lower - and no improvement on stereoselectivity was observed.153 This comes as a surprise as magnesium complexes

50 usually exhibit higher activity that their related zinc analogues. The authors suggest that the crowded environment induces a switch in the rate determining step from insertion to coordination of LA.153 Later on, Williams confirms that the halide ligands (I-241-243) have a great impact on propagation rates attesting that a delicate electronic and steric balance is required for polymerization (Br > Cl > I).154

Encouraged by these results, Williams and co-workers also generated the imine and secondary amine bromide analogues (I-246-247).154 Whereas the imine counterpart is significantly slower, the secondary amine complex is incredibly active.154

Figure 1-45: Di(amino) and di(imino)phenolate zinc, cobalt and magnesium alkoxides.152-154

In 2007, Carpentier and co-workers reported the aminobis(pyrazolil) zinc alkoxide I-248.155

Interestingly, all aluminum and zinc alkyl analogues are mononuclear while the alkoxide I-248 is dinucleating and allows much more controlled ROP.155 This highlights the importance of the second metal in the coordination sphere during polymerization.

Figure 1-46: Aminobis(pyrazolil) and scorpionate zinc complexes.138, 155

Alongside, Rodríguez and co-workers generated the related scorpionate zinc complexes I-249-252.138

The complexes are modestly active compared to other zinc alkoxides, and increasing the steric of the

138 alkyl was found to decrease the overall activity (CH2SiMe3 < Et < Me). The author suggests that the steric of the alkyl groups prevent coordination of LA. Going from an alkoxide to a sulfoxide was also found to decrease the activity.138 Whereas the tethered analogue I-191 yielded essentially atactic PLA,

138 the dinucleating complexes I-249-252 afforded isotactic-enriched PLA (Pm = 0.59-0.74). Tacticity increases with the size of the alkyl groups.138 Analyses of stereoerrors suggests a SCM but surprisingly

I-252 is more stereoselective than I-249 highlighting that the mechanism might be more complex.138

More recently, Wu and co-workers generated the dinucleating complexes I-253-255 using bis(Salen) ligands backbone.156 Interestingly, the complexes were not active at room temperature and required 130

°C or the use of a co-catalyst such as benzyl alcohol.156 The authors suggest that deactivation of the alkoxide arises from the strong electronic donation of the ligand backbone.156 Thus the complexes are very stable and allow the use to up to 50 equivalents of benzyl alkoxide under immortal conditions.156

The authors found spectroscopic evidences that whereas an activated monomer mechanism occurs at room temperature in solution, a coordination insertion mechanism is observed in the melt.156 As observed by Hillmyer and Tolman for dinucleating complexes I-241 and I-245, the zinc complex I-254

52 is more active than the magnesium analogue I-253.156 Complexes yielded essentially atactic PLA from rac-LA.

Figure 1-47: Bis(Salen) magnesium and zinc alkoxides reported by Wu.156

The zinc, magnesium and aluminum bis(iminopyrrolide) alkoxides (I-256-258) reported by Wang and co-workers feature different coordination modes.157 The bimetallic aluminum is (4-N,N,N,N, 0), whereas the magnesium and zinc analogues are ( 2-N,N,  2-N,N).157 Once again the zinc complex I-

257 was found to be more active than the magnesium analogue I-256.157 In contrast, the aluminum

157 counterpart is modestly active. The complexes I-257-258 yield heterotactic-enriched PLA (Pr = 0.60-

0.70) from rac-LA.157 The complexes were found to have higher activity and better stereocontrol in THF than in toluene.157 Kinetic studies highlight that the complexes I-256-257 are first order with respect to the monomer.157

Figure 1-48: Bis(iminopyrrolide) zinc, magnesium and aluminum alkoxides reported by Wang.157

Dinucleating group IV initiators were pioneered by Coates, Kol and co-workers with the phenylenedi(amine)-bis(phenolate) titanium alkoxide I-259.158 The complex was obtained from the

i i addition of 2 equivalents of Ti( OPr)4, however the use of only one equivalent of Ti( OPr)4, a bulkier alkoxide (OtBu), or a larger metal (Zr) leads to the monomeric counterpart.158 I-259 was poorly active and not stereoselective in contrast with its mononuclear analogues.158

Figure 1-49: Phenylenedi(amino)bis(phenolate) titanium alkoxide reported by Coates and Kol.158

This class of ligand was further investigated by other groups.159-162 Chakraborty and Lin separately reported the use of Salen ligand to generate complexes I-260-264.159, 160Complexes I-260-261 are active for the ROP of lactide but not stereoselective and interestingly, the dimeric hydroxy analogue of I-261 was also active.159 These complexes exhibit no induction period and are first order with respect to the monomer.159 Molecular weights suggest that only one alkoxide is initiating polymerization.159 The related complexes with various backbones were studied by Lin.160 I-264 was found to have lower activity and stereocontrol than I-262-263 which yield heterotactic-enriched PLA (Pr = 0.62-0.64) from rac-

LA.160 Unlike Chakraborty, they found that two alkoxide groups were initiating polymerization and that the polymerization was first order with respect to the monomer and 1.5 with respect to the catalyst.160

Thus, they propose a mechanism where the alkoxide-bridged dissociate upon coordination of lactide and where the polymer chain grows on a non-bridging alkoxide.160

Figure 1-50: Group IV Salen alkoxides and related complexes.159-162

Chen and co-workers also reported a series of group IV Salen-like complexes (I-265-272).161 The steric bulk of the imine substituent was found to have a great impact on activity (H > Me > Ph).161

Electron donating group on the aryl substituents (I-267-268 and I-272) was also found to modestly increase the activity.161 Dinuclear complexes were found to be 10 to 60 times more active than their mononuclear counterparts.161 Complexes were not stereoselective.161 In this study, the authors found that

4 of the 6 alkoxides were initiating the polymerization and that the bridging alkoxides remained unreactive during polymerization.161 Note that, in contrast with the mechanism proposed by Lin and described above, no evidences for dissociation of the bridging alkoxides were found.

Finally, Huang and co-workers reported the hydroxyphenyl Salen-like group IV complexes I-273-

275.162 Increasing the size of the metal center from Ti to Zr and Hf significantly increased the activity

(Hf = Zr >> Ti). Molecular weight analysis reveals that, whereas complex I-273 seems to have 2-3 alkoxides group initiating polymerization, complexes I-274-275 have 4 initiating alkoxides. Again no evidence was found for the dissociation or initiation of the bridging alkoxide. Essentially atactic PLA was obtained from rac-LA.

1.5.4. Heterobimetallic complexes

By the nature of their structures, heterobimetallic complexes form a class of compound by themselves. As we will see in this part, a majority of them arise from the unexpected residual presence of alkali metals.

Figure 1-51: Binap-like yttrium lithium complex reported by Carpentier and co-workers.163

The heterobimetallic Binap-like yttrium lithium complex I-276 was reported by Carpentier and co- workers and inspired by complex I-142 reported by Coates.163 Surprisingly, complex I-276 is active for the ROP of lactide without the addition of a nucleophilic cocatalyst.163 Chain end analysis suggests that the ligand not only acts as an ancillary group but as a reactive nucleophile and then undergo transesterification with the quenching agent methanol or ethanol.163 Complex I-276 exhibits high stereocontrol yielding heterotactic and syndiotactic PLA (Pr = 0.99 and Pr = 0.80) upon addition of rac-

La and meso-LA respectively.163

Other heterometallic complexes comprising alkali metals (I-277-282) were reported by Thomas,

Carpentier and Leznoff supported by various ligand backbones.114, 164, 165 The tetradentate Salen-like complexes I-277-278 were obtained along with their mononuclear indium analogues bearing a more

164 bulky alkyl (CH2SiMe3) and catalyze the ROP of lactide without the need of a co-catalyst.

Mononuclear and dinucleating complexes feature similar activity although the dinucleating yielded

164 slight heterotactic-enriched PLA (Pr = 0.57-0.62) from rac-LA but with less control. Thomas reported bi(phenolate) complexes I-279-280.165 Although these are active for ROP of lactide, they were found to

56 be better behaved in presence of an external alcohol to for the alkoxide in situ.165 Polymerization carried out in THF was found to be much faster than in toluene.165 Low temperature (up to -30 °C) allows the formation of highly heterotactic PLA (Pr = 0.90) from rac-LA from complex I-279, while complex I-

165 280, less active, was stereoselective at room temperature (Pr = 0.84-0.88). The authors also screened several other structurally different diols, although there were not characterized.165 Finally the heterobimetallic analogue of complexes I-140-141 was obtained from thorium by Carpentier and

Leznoff (I-281-282).114 The complexes were highly active and comparable to their analogues but poorly controlled.114

Figure 1-52: Heterometallic complexes comprising residual lithium centers.114, 164, 165

The heterobimetallic nickel(II) lanthanide(III) nitrates supporting by Salen-type ligands (I-283-289) were reported by Lü, Bao and co-workers.166-168 The complexes were found to be modestly active under melt conditions with lower activity and control than the related mononuclear nickel analogue.168 The authors suggest that the decrease in activity arises from the steric hindrance of the lanthanide nitrates, inhibiting coordination of lactide.166-168

Figure 1-53: Nickel(II) lanthanide(III) Salen-type nitrates reported by Lü, Bao and co-workers.166-168

Finally, Lin and co-workers reported a series of heterobimetallic titanium zinc (I-290-293), magnesium (I-294-296), lithium (I-297) and sodium (I-298) alkoxides supported by bis(phenolate) ligand platforms.169 In contrast with the mononuclear titanium analogues, all magnesium and zinc complexes features enhanced catalytic activity (Zn > Mg) whereas lithium and magnesium complexes exhibit similar reactivity.169 Based on molecular analysis the authors found that probably two of the four alkoxides are initiating polymerization for complexes I-290-296. Ligand design evidenced that electron donor group were beneficial to the polymerization with I-294 > I-292 > I-291 > I-293 and I-296 > I-

295 > I-294.169 Kinetic studies show a first order dependence on the monomer and catalyst for the zinc complex I-292.169

Figure 1-54: Heterobimetallic titanium zinc, magnesium, lithium and sodium alkoxides reported by Lin and co-workers.169

1.5.5. Conclusion

Although it seems clear that nuclearity can play a major role in the stereoselective polymerization of lactide, it is hard to find trend and direct correlation between nuclearity and activity or stereoselctivity.

It must be noted that cooperativity of two metal centers is not always beneficial, in that sense cooperativity can be constructive (enhancing catalysis) or destructive (inhibiting catalysis). Here, we will attempt to draw conclusions that might be of interest for the reader.

For dimeric species, a wide majority of complexes (I-8, I-15-21, I-34, I-43-44, I-60, I-70-81, I-93-

96, I-105-110, I-146-150, I-161-165 and I-169-174) simply dissociate during polymerization. When comparison with monomeric analogues was possible, an initiation period was observed but overall behavior is usually comparable in terms of activity and stereoselectivity. In these cases, this initiation period is probably related to the dissociation of the complex and the formation of the active species.

More interestingly, a non-negligible number of dimeric complexes (I-2-7, I-35, I-41-42, I-45-59 and I-

61-69) retain their dinuclear structures during polymerization. Sadly, with the exception of complex I-

35, monomeric analogues of these complexes were not generated thus no direct comparison can be

59 undertaken. However, the surprising stereoselectivity of complexes I-2-7 and I-35 highlights the potential of dimeric propagating species to achieve stereocontrolled polymerization. In addition, some complexes (I-1, I-10-14, I-82-90, I-112-126, I-127, I-166-168, I-182-183) exhibit unresolved/unclear mechanisms which possibly indicates either a mixture of dissociated and un-dissociated catalyst, or a greater fluxionality with intermediates being successively mononuclear and dinuclear, or both.

In the case of tethered catalysts which mostly feature alkyl complexes, the question which arises does not concern the nuclearity but rather the role of the second metal. Therefore, the first question to answer concerns the number of active centers. Both metals can be involved in polymerization, regardless of the number of growing chain per catalyst. One of the metals can also have a structural role meaning that it will not be involved in polymerization (I-189-190 and I-192-193). It must be acknowledged that the use of alkyls - in comparison to more reactive alkoxides - as metal initiators can lead to two different mechanisms in presence of alcohol, coordination-insertion or activated monomer. As observed for complexes I-206-207, different mechanisms can lead to different synergic effects. Interestingly some tethered catalysts when compared to their monomeric shows enhanced activity for polymerization (I-

194-202, I-206 and I-232-234) while others just act as two distinct monomeric analogues (I-207-213 and I-227-231).

Dinucleating complexes are attractive because unlike dimeric or tethered complexes, the two metals must cooperate, as they must both be involved in the catalysis. The question remains if polymerization will be positively or negatively affected. Note that in the case of group IV metals, the question of the number of active initiating groups needs to be answered given the number of alkoxides on each metal.

Whereas magnesium complexes were more active than zinc for dimeric species (also note that mechanism were different, see I-60-61 for example), for dinucleating complexes the opposite was found, suggesting that forcing the two magnesium centers to cooperate in a more rigid environment impacts the

60 polymerization negatively. Compared to their monomeric counterparts some dinucleating complexes feature enhanced activity and control (I-248-252 and I-265-272), but other are considerably slower (I-

248 and I-259). Dinucleating complexes also seems to be more greatly affected by donating/coordinating solvent, thus it seems that the steric bulk generated by these ligand leads to the rate determining step being the coordination of lactide.

This study highlights the importance of understanding the mechanism of polymerization which often remains unclear in bimetallic catalysis. In the case of lactide polymerization, mechanistic studies are crucial to solve the two important challenges of the field: 1) resolve the origin of stereoselectivity, 2) assess the competition and contribution of site control mechanism versus chain-end control.

1.6. Scope of the thesis

In our group, we reported the first indium-based initiator (I-45), for the efficient and highly controlled ring opening polymerization of lactide.89 Since, other indium-based catalysts for ring opening polymerization have been reported.64, 65, 68, 89-97, 116, 118, 119, 124, 125, 131, 163, 164, 170-177 Although these include both mononuclear68, 163, 171, 173, 177 and dinuclear64, 65, 89, 91-97, 116, 118, 119, 124, 125, 131, 172, 173, 175, 176 indium species, there are only two reports of a tethered bimetallic indium catalyst for the ring opening polymerization of cyclic esters (I-206 and I-213),142, 175 the other being dimeric. Thus, to our knowledge, no examples of a dinucleating ligand platform for indium have been reported.

We are interested in the role of the nuclearity of indium complexes on polymerization mechanism.

Complex I-45 is a highly controlled catalyst for the living and immortal polymerization of cyclic esters, and we have demonstrated experimentally and computationally that the propagating species for this catalyst is dinuclear.89, 91, 92, 97 In contrast, indium complexes with tridentate or tetradentate supports developed in our group that do not retain a dinuclear structure during polymerization do not show the

61 same control of polymer molecular weight, dispersity, or selectivity. The importance of cooperativity is greatly reduced for a tethered system; in complexes I-206 and I-213 the two indium centers act as two distinct initiating points for immortal polymerization of lactide, with reactivity comparable to the monomeric analogue.142, 175 We were interested to see if replacing the dimeric or tethered motif in indium complexes with a dinucleating motif, such as the bis(diamino)phenolate supports reported for complexes

I-241-247, would enable greater cooperation between the metal centers.152-154 Initial studies for the polymerization of rac-LA with I-241 or its zinc analogues showed high activity but no stereocontrol.152,

154 Magnesium and cobalt analogues were also not selective and showed lower activity than the zinc analogues.153

The scope of this thesis is to synthesize a chiral bis(diamino)phenolate ligand platform inspired by I-

241 and explore the synthesis and reactivity of corresponding indium and zinc complexes through various synthetic strategies including an easy and efficient one pot reaction. Using this process, we report the first examples of indium metals supported by a dinucleating platform. The reactivity of these complexes as catalysts for the ring opening polymerization of lactide will also be discussed.

The first chapter of this thesis was written entirely by myself with minor corrections from my supervisor Dr. Parisa Mehrkhodavandi. The second chapter was also written by myself but more heavily edited by my supervisor. The work presented below was done in collaboration with Dr. Kim Osten and

Dr. Insun Yu who initiated this project and first isolated the indium complexes bearing achiral proligands. Dr. Insun Yu, Dr. Dinesh Aluthge and Tannaz Ebrahimi contributed to the solving of X-ray crystal structures. The second chapter of this thesis has been published in the American Chemical Society journal Inorganic Chemistry (in press, DOI: 10.1021/acs.inorgchem.6b00358).

Chapter 2: Dinucleating ligand platforms supporting zinc and indium catalysts for stereoselective lactide polymerization

2.1. Introduction

As discussed in Chapter 1, our group was interested in exploring the use of dinucleating platforms to support indium catalysts for the stereoselective ROP of lactide. Kim Osten and Insun Yu, past members of the Mehrkhodavandi group, started this project by synthesizing a series of achiral di(diamino)phenolate ligands similar to those reported by Hillmyer and Tolman and Williams (Figure 2-

152, 154 1) comprising imine (L1), secondary (L2) and tertiary (L3) amine moities.

Figure 2-1: Structure of the achiral ligands L1-3 synthesized and characterized by Kim Osten.

Ligand L3 was successfully used to generate dinucleating indium chloride (1), alkoxide (2) and hydroxy (3) complexes (Figure 2-2) with poor yield, whereas, intractable mixtures of compounds were obtained from L1 and L2. The crystal structures of complexes 2 and 3 were also obtained (Figure 2-3 and Table E-3). We were particularly interested in exploring indium complexes arising from L2 as we and Williams observed influence of the central amine moiety on the ROP rates of lactide as discussed in the previous chapter.94, 154

Figure 2-2: Synthesis of [(L3)In2Cl4](μ-Cl) (1), [(L3)In2Cl4](μ-OEt) (2), and [(L3)In2Cl4](μ-OH) (3) established by Kim Osten and Insun Yu.

We hypothesized that the great flexibility of ligand L2 was probably inhibiting the formation of a stable dinucleating complex and therefore targeted a more rigid complex. A di(diamino)phenolate ligand bearing cyclohexyl arms and secondary central amine was proposed. This work describes the synthesis of these rigid chiral ligands and their use to generate enantiopure indium and zinc complexes for the

ROP of lactide. Optimization of the synthesis of 2, as well as a more comprehensive analysis is also described.

Figure 2-3: Solid state structures of 2 (left) and 3 (right) obtained by single crystal X-ray diffraction by Kim Osten and Insun Yu respectively. Thermal ellipsoids are set at 50% probability and H atoms and solvent are removed for clarity. The unit cell for complex 2 contains two distinct conformations differing in the position of the ethoxide groups; only one is depicted here. Bond lengths available in appendix (Table E-3).

Note that all complexes described in this chapter are chiral, wherever they have been obtained from achiral or chiral ligands. For example, complexes 1, 2 and 3 are most probably obtained as a mixture of enantiomers (Figure 2-4) – chirality arises from the metal center. Note that only enantiomer A (Λ,Λ) was observed by single crystal X-ray diffraction for complex 2 and 3. From a theoretical point of view, the diastereomeric meso complex could also be present: (Δ,Λ). The “meso” complex is likely to be less stable due to possible interactions arising from the methyl of the terminal amines.

Figure 2-4: Structure of the two enantiomers arising from the dinucleating structure of complexes 1, 2 and 3 (Δ = right handed, Λ = left handed – viewed from a plane above the metal center containing the three atoms with highest CIP numbers - as defined for octacoordinated complexes202).

2.2. Synthesis and characterization of chiral proligands

Chiral pentadentate aminophenolate proligands L4-7 can be synthesized through modification of reported procedures from condensation of an excess amount of freshly distilled (±)-, (RR)- or (SS)-N,N- dimethylcyclohexyldiamine and 2,6-diformyl-4-tert-butyl-phenol or 2,6-diformyl-4-methyl-phenol

154, 178 1 (Figure 2-5). The H NMR spectra of L4 and L5 show broad singlets at 7.63/7.43 and 8.53/8.49 ppm corresponding to the Ar-H and imine N=CH protons, respectively, regardless of the stereochemistry

(Figures C-1-3). Excess of N,N-dimethylcyclohexyldiamine was essential to afford the ligand without the formation of mono-substituted imine. Reduction of L4-5 with NaBH4 forms proligands L6-7 (Figure

C-4-8) as viscous oils resistant to crystallization. Purification of L4-7 using various methods was attempted without success.

Figure 2-5: Synthesis of proligands L4-7.

1 The H NMR spectra of the enantiopure analogues of L6 and L7 show characteristic singlets for Ar-

H protons in addition to a set of diastereotopic protons corresponding to NH-CH2-Ar. In contrast, the

1 H NMR spectra of L6 and L7 generated from racemic amine show several sets of peaks, suggesting the

1 presence of several diastereomers. The H NMR spectra of RR/RR- and SS/SS-L7 are similar and different from that of the racemic species (±)-L7, which includes the two enantiomers as well as (SS/RR)-

L7 (Figure 2-6). The same observations can be made for L6 (Figure C-8). In order to facilitate the analysis and characterization of the metal complexes, only the enantiopure RR/RR-L7 will be used for metallation reactions.

Figure 2-6: Diastereoisomeric distribution in the synthesis of rac-L7 when starting with the racemic 1 trans-N,N-dimethylcyclohexyldiamine (top). H NMR spectra (300 MHz, CDCl3, 25 °C) of enantiopure RR/RR-L7, SS/SS-L7, rac-L7 and crossover experiments associated of the SS/SS-L7 + RR/RR-L7 and rac- L7 + RR/RR-L7 (bottom).

2.3. Synthesis of indium and zinc alkoxide complexes bearing chiral ligand backbones.

Dinucleating proligands L1-7 were screened as supports for both zinc and indium, but different strategies had to be used for different complexes. These strategies were not successful for L1-2.

Syntheses of indium and zinc alkoxide complexes of L1-7 were attempted through two pathways (Figure

2-7): (1) deprotonation of the proligand, salt metathesis with a metal chloride salt to yield the dinucleating chloride bridged intermediate, and further reaction with alkoxide salts to form the targeted

68 dinucleating alkoxide species, and (2) a one pot synthesis with the proligand, metal halide, and alkoxide salt.

Figure 2-7: Pathways for the synthesis of zinc and indium alkoxide complexes supported by L1-7.

Proligand, L7, is the only chiral species that forms isolable alkoxide complexes. Reaction of RR/RR-

L7 with four equivalents of NaOEt in toluene at room temperature for 16 h, followed by addition of InCl3 or ZnCl2, forms RR/RR-[(L4)In2Cl4](μ-OEt) (RR/RR-4) and RR/RR-[(L4)Zn2Cl2](μ-OEt) (RR/RR-5)

(Figure 2-8). Low yielding recrystallization from toluene affords pure complex RR/RR-4. The NMR spectra of complexes 4 and 5 indicate a high degree of symmetry (Figure C-9 and Figure C-12).

Complex 4 exhibits a sharp singlet for the aromatic protons at 6.82 ppm as well as the two distinct methyl groups born by the terminal amines at 2.81 ppm and 2.40 ppm. A set of doublets is observed for the diastereotopic methylene protons Ar-CH2-NH. Finally, the alkoxide peaks are diagnostic of the desired product, with a triplet for CH3-CH2-O- at 1.49 ppm and two diastereotopic resonances for CH3-CH2-O

69 at 4.35-4.39 ppm, 4.44-4.48 ppm as often observed for these alkoxide bridged dinuclear species. The

1H NMR spectrum of complex 5 is similar, with signals at slightly lower chemical shifts. Synthesis of complex 5 was attempted from the mixture of diastereomers rac-L7, but due to the complex mixture of compounds obtained (Figure B-1) we only pursued our investigation with RR/RR-L7, as mentioned above.

Figure 2-8: One pot syntheses of RR/RR-[(L7)In2Cl4](μ-OEt) (4) and RR/RR-[(L7)Zn2Cl2](μ-OEt) (5).

Single crystals of pure complex RR/RR-4 could not be obtained. In order to successfully obtain a single crystal of this compound, we used a ligand mixture that was synthesized from a poorly resolved

N,N-dimethylcyclohexyldiamine (90% RR and 10% SS). The complex crystallizes as a racemate in a centrosymmetric space group and the structure shown is SS/SS-4 (Δ,Δ). Note that pure RR/RR-4 is used for all reactivity studies. The crystals were obtained from a saturated solution of CHCl3 and diethyl ether at room temperature. The molecular structure of 4, obtained by single crystal X-ray crystallography

(Figure 2-9 and Table E-3), shows six coordinate distorted octahedral indium centers bridged through

70 the phenolic and alkoxide oxygens. The chlorides coordinated to each indium center point in different direction to retain C2 symmetry along the O1-Ar-Me axis. Each arm is highly twisted to accommodate the presence of the chlorides. The In-N distances for the central nitrogen atoms are shorter (2.226 and

2.235 Å) than the terminal ones (2.388 and 2.380 Å).

Figure 2-9: Solid state structure of complex SS/SS-4 (Δ,Δ) obtained by single crystal X-ray diffraction. Thermal ellipsoids are set at 50% probability and H atoms and solvent removed for clarity. Bond lengths available in appendix (Table E-3).

Although it is possible to form several conformers of the complex based on the orientation of the chelating diamine rings (Figure 2-10), as observed in solid state structures reported for analogous zinc complexes bearing achiral ligands,152, 154 no such conformers are observed in our case. The introduction

71 of the cyclohexyl rings requires the two amino substituents to be in equatorial positions and thus only one stable conformation can be obtained as opposed to the ethylene backbone (Figure 2-10).

Figure 2-10: Structure of the different chiral conformations of the chelating ring arising from cyclohexyldiamine and ethylenediamine (δ = right handed, λ = left handed as defined for chelating ring). Note that a RR-cyclohexyldiamine chelating ring will always display a λ configuration while an ethylenediamine chelating ring can be λ or δ.

The formation of two diastereomers, however, cannot be ruled out as proposed for complexes 1, 2 and 3. Note that, the two enantiomers proposed for 1, 2 and 3 (Figure 2-4) become diastereomers for

RR/RR-4 due to the chirality of the ligand backbone - the enantiomers of each diastereomers being the diastereomers of SS/SS-4 (Figure 2-11). The NMR signatures of complex 4 and 5 suggest that only one diastereomer is obtained. From a theoretical point of view, the presence of a second diastereomer should give rise to a second NMR signal. It is also possible that the two signals are overlapping.

Figure 2-11: Structure of the different isomers arising from the dinucleating structure of RR/RR-1 and SS/SS-1 and their relationship (Δ = right handed, Λ = left handed – viewed from a plane above the metal center containing the three atoms with highest CIP numbers - as defined for octacoordinated complexes202).

The two step methodology can only be used to generate zinc complex 5; although the indium analogue was observed it cannot be isolated using this route and an intractable mixture of compounds is obtained.

Deprotonation of RR/RR-L7 with KO(t-Bu) followed by salt metathesis with ZnCl2 affords RR/RR-

[(L4)Zn2Cl2](μ-Cl) (RR/RR-6) in 56% isolated yield (Figure 2-12). Further reaction of 6 with two equivalents of NaOEt affords 5 in 39% yield resulting in an overall isolated yield of 22%. Interestingly, complex 5 exhibits a triplet in addition to a doublet for the diastereotopic protons Ar-CH2-NH at 4.11 and 3.65 ppm which suggests the presence of different diastereomers or conformers in solution (Figure

C-13 and Figure C-14).

Figure 2-12: Synthesis of RR/RR-[(L7)Zn2Cl2](μ-OEt) (2) in a stepwise reaction.

The chiral imine RR/RR-L5 can be used as a dinucleating platform for zinc, but not indium and does not support metal alkoxide formation. Deprotonation of the imine proligand RR/RR-L5 followed by salt-

1 metathesis with ZnCl2 forms RR/RR-[(L5)Zn2Cl2](μ-Cl) (RR/RR-7) in 56% yield (Figure 2-13). The H and 13C{1H} NMR spectra show only one set of signals (Figure C-15 and Figure C-16). As observed for complexes 4 and 5, complex 7 is highly symmetrical, showing only one singlet at 8.29 ppm corresponding to imine N−H and another at 7.14 ppm corresponding to the aromatic protons. Only two singlets for the N−CH3 protons are observed at 2.63 and 2.43 ppm.

X-ray quality crystals of RR/RR-7 were grown from a saturated solution of THF and diethyl ether at room temperature and analyzed by single crystal X-ray crystallography (Figure 2-14). The solid state structure shows five-coordinate zinc centers bridged by the phenolic oxygen and a chloride ligand. The geometry of the zinc center is between trigonal bipyramidal and square pyramidal. The terminal chlorides occupy the apical positions at each zinc center. The unit cell contains two diastereomers, 7a and 7b, which differ in the orientation of the terminal chloride ligands (Figure 2-15 and Figure E-1).

Such diastereomers were also observed for the amino zinc systems bearing achiral ligand backbones.154

The N,N’ chelating ring shows a similar conformation in the two diastereomers, as explained above. C2 symmetry is retained through the O1-Ar-Me axis.

Figure 2-13: Attempted synthesis of zinc alkoxide complex with imine ligand RR/RR-L5.

Formation of the bridging alkoxide complex through further reactivity of 7 or through a one-pot reaction is not possible (Figure 2-13). Reaction of RR/RR-7 with various equivalents of NaOEt leads to a mixture of compounds including unreacted 7 (Figure B-2). Attempted recrystallization from a saturated solution of THF and diethyl ether leads to the isolation of the hydroxyl-bridged byproduct

RR/RR-[(L2)Zn2Cl2](μ-OH) (RR/RR-8). Complex 8 was characterized by single crystal X-ray diffraction

(Figure 2-14). The solid state structures of 7 and 8 show similarly distorted square pyramidal geometry around the zinc centers (Table E-1). The one-pot reaction of RR/RR-L5 and four equivalents of NaOEt, followed by addition of ZnCl2, forms pure RR/RR-7; mixtures containing the desired alkoxide complex are not observed.

Figure 2-14: Solid state structures of complexes RR/RR-7 (top) and RR/RR-8 (bottom) obtained by single crystal X-ray diffraction. Thermal ellipsoids are set at 50% probability and H atoms and solvent are removed for clarity. The unit cell for complex 7 contains two distinct diastereomers, 7a and 7b, differing in the position of the chloride ligands around the zinc centers, only 7b is depicted here. Bond lengths available in appendix (Table E-1).

The analogous indium complexes bearing imine proligand RR/RR-L5 cannot be synthesized under these conditions. The crystal structure of complex 4 shows that the ligand arms must be twisted to accommodate the chloride present on each indium center. Presumably, the rigid backbone in the imine ligand does not allow such a thermodynamically stable structure and prevents the formation of the desired complex (Figure 2-15).

Figure 2-15: Structure of the zinc diastereomers 7a and 7b and their respective enantiomers (top), and hypothetic structure of the indium analogue (bottom), (C = clockwise, A = anticlockwise - viewed from a plane above the metal center containing three of atoms with the highest CIP priority numbers - as defined for pentacoordinated complexes202).

2.4. Synthesis indium alkoxide complexes bearing achiral ligand backbone.

In an attempt to draw comparisons with the previously reported zinc complexes I-241-243,152, 154, 179 we attempted the synthesis of indium complexes with proligands L1-3. Only L3 was successful. Achiral indium alkoxide complexes can be prepared by both stepwise and one-pot routes (Figure 2-2 and Figure

2-16). Deprotonation of L3 followed by salt metathesis with InCl3 forms [(L3)In2Cl4](μ-Cl) (1) with

33% isolated yield. Further reaction with two equivalents of NaOEt forms [(L3)In2Cl4](μ-OEt) (2) in

21% yield (overall yield of 7%). Alternatively, a one-pot reaction can directly access complex 2 in 34% isolated yield.

Figure 2-16: One-pot synthesis of complex 2.

As observed with analogous zinc complexes,154 the 1H and 13C{1H} NMR spectra of complexes 1 and

2 show that they are highly symmetric (Figure C-17 and Figure C-18). Complex 1 exhibits sharp singlets for the aromatic protons at 7.05 ppm and the N-CH3 protons of the central amines at 3.02 ppm. The diastereotopic protons of Ar-CH2-NH appear as a set of doublets while the N-(CH3)2 protons appear as two broad singlets at 2.60 and 2.87 ppm. Complex 2 exhibits a similar 1H NMR signature, with the two broad singlets for the N-(CH3)2 protons as observed at 2.86 and 3.02 ppm (CDCl3, 25 °C). Variable

1 temperature H NMR spectra of complex 2 (toluene-d8, −35 - 75 °C) show that these two broad signals sharpen to two singlets at 2.29 and 2.63 ppm at −35 °C (Figure 2-17). Coalescence is observed at 40 78

°C, although peak overlap prevents calculation of the exchange rate. One sharp singlet is observed at

2.49 ppm at 65 °C. This fluxionality strongly suggests that the terminal amines are labile. Thus a possible equilibrium between the two enantiomers is possible if both terminal amines undergo decoordination

(Figure 2-18). A more complex mechanism also involving equilibrium between the two isomers, by N- inversion at the central amine, would require the dissociation of both terminal and internal amines and would lead to observation of equivalent benzylic CH2-N protons; this is not observed.

1 Figure 2-17: H VT NMR (400 MHz, toluene-d8) of complex 2 from −35 to 75 °C. Chemical shifts were obtained using methylene protons of toluene-d8 as a reference (δ = 2.09 ppm). The peaks of interest (blue) correspond to the protons of the methyl group located on the terminal amine. The shifts observed were reversible.

Figure 2-18: Possible mechanism of equilibrium between the two enantiomers of complexes 1, 2 and 3.

Single crystals of complex 2 suitable for X-ray analysis were grown from a saturated solution in toluene (Figure 2-3). Attempts to crystalize complex 1 from a saturated solution of acetonitrile result in the isolation of the hydrolysis complex [(L3)In2Cl4](μ-OH) (3), as determined by single crystal X-ray diffraction (Figure 2-3). The solid state structures of complexes 2 and 3 are similar to that of 4, with six-coordinate distorted octahedral indium centers bridged through the phenolic and alkoxy or hydroxy oxygens. Bond lengths and angles around indium centers for 2 and 3 are comparable (Table E-3). As observed for complex 2, the relative positions of the chloride ligands and the arms are critical to retaining

C2 symmetry along the O1-Ar-t-Bu axis and affording the thermodynamically stable complex.

Interestingly, the zinc analogues display only sharp singlets for the methyl protons of the terminal amine

(regardless the central moiety) and only split to different singlet at very low temperatures (-75 °C).152,

154 Thus, these achiral indium species display a lower degree of flexibility than the reported zinc analogues, perhaps due to the greater steric congestion created by the larger indium cation.154

2.5. Polymerization of lactide.

Complexes [(L3)In2Cl4](μ-OEt) (2), RR/RR-[(L7)In2Cl4](μ-OEt) (RR/RR-4) and RR/RR-

[(L7)Zn2Cl2](μ-OEt) (RR/RR-5) show higher selectivity and lower activity as catalysts for the ring- opening polymerization of rac-LA compared to analogous mononuclear indium118, 180-182 and mononuclear and dinucleating zinc88, 152, 154 complexes (Table 2-1). Polymerization of 200 equivalents of rac-LA catalyzed with 4 (~2 mM in CH2Cl2) reaches 82% conversion after 11 days; complex 2 reaches

97% conversion after 8 days. Both complexes 2 and 4 yield highly heterotactic PLA with Pr values of

0.89 and 0.87, respectively (Figure D-1).

Polymerization of meso-LA by complex 2 results in atactic PLA (Pr = 0.46, Figure D-2), which is consistent with a chain end control mechanism. Complex 5 reaches 95% conversion after 5 days under identical reaction conditions, yielding PLA with a heterotactic bias (Pr = 0.64). In contrast, the achiral zinc complex C polymerizes 200 equivalents of rac-LA in less than 1 minute to yield atactic PLA.154

Reactions of complex 2 at higher temperatures, 40 °C (CH2Cl2), 100 °C (toluene), and 140 °C (neat in melt), leads to higher activity but inferior selectivity and dispersity as expected (Table 2-1, entries 7-

9). Interestingly, complex 2 is much less active in THF, reaching only 49% conversion after 7 days.

This suggests that THF competes with the coordination of lactide at the indium center prior to ring opening polymerization, emphasizing the delicate steric effects at the sterically congested indium centers in these complexes. Similar effects are observed with added alcohol. In contrast to their dimeric indium counterparts, the activity of complexes 2 and 4 is decreased with addition of alcohol (Table 2-1, entries

3 and 10).183-185 Steric bulk and less flexibility of complex 4 is likely decreasing the chain transfer reaction rate.

Table 2-1: Polymerization of lactide by initiators 2, 4, and 5.

a T Conv. Mntheo MnGPC b d Entry Cat Sol. Time [LA]/[cat] Ɖ Pr (°C) (%) (Da)b (Da)c 17b e 1 C CH2Cl2 R.T. 1 min 200 92 26570 23800 1.06 - 2 4 CH2Cl2 R.T. 11 d 200 82 23640 17380 1.02 0.89 f 3 4 CH2Cl2 40 7 d 1000 < 10 - - - - 4 5 CH2Cl2 R.T. 5 d 200 95 27430 24290 1.04 0.64 5 2 CH2Cl2 R.T. 8 d 200 97 28000 17080 1.06 0.87 6 2 THF R.T. 7 d 200 49 - - - 0.88 7 2 CH2Cl2 40 5 d 200 95 27430 15800 1.03 0.87 8 2 C7H8 100 1 d 200 94 27130 10740 1.64 0.62 9 2 - 140 3 h 200 98 28310 26080 1.80 0.60 f 10 2 CH2Cl2 40 7 d 1000 30 - - - - 11 2 CH2Cl2 R.T. 8 d 200 98 28890 20670 1.17 0.46

Polymerization of rac-LA (entries 1-10) and meso-LA (entry 11). [LnM2(µ-OEt)] = 5 mM (entry 1), a 1 b [LnM2(µ-OEt)] = 2 mM (entries 2-11). Monomer conversion determined by H NMR spectroscopy. c Calculated from [LA]o/[initiator] × LA conversion × MLA (144.13) + MEtOH (46.07). Determined by GPC measurements in THF. d Calculated from the 1H{1H} NMR spectra and Bernoullian statistics.27 e Atactic PLA. f Immortal polymerization of 1000 equivalents of rac-LA with 5 equivalents of benzyl alcohol as chain transfer agent, [LA]0:[BnOH]:[ LnM2(µ-OEt)] = 1000:5:1. Polymerization of 20 equivalents of rac-LA with complex 2 can be carried out at room temperature to generate low molecular weight polymer for end group analysis with matrix-assisted laser desorption/ionization-time-of-flight (MALDI-TOF) mass spectroscopy (Figure D-3). The resulting polymers exhibit one distribution of peaks, separated by m/z = 144 Da, corresponding to

+ [H(C6H8O4)n(OC2H5)Na] . No peaks at m/z = 72 Da, indicative of transesterification, are observed. The same results are observed in toluene at 100 °C (Figure D-4). Thus, it appears that the deterioration in stereoselectivity and dispersity observed at higher temperatures does not arise from transesterification.

2.6. Conclusions and perspectives

Complexes 2 and 4 were designed to prevent dissociation of the two indium centers during polymerization, therefore forcing the ROP to proceed via a tandem mechanism as we reported for dinuclear indium system I-45.181 However, the new dinucleating indium systems are fundamentally different as they exhibit rigid octahedral structure at the indium centers, resulting in heavily congested active sites. In contrast most of the indium catalysts in the literature have a square pyramidal or square planar structure around the indium centers and therefore exhibit a free coordination site available for lactide coordination.68, 172, 180, 188-194 In catalysts with high steric encumbrance, the catalysts dissociate during polymerization and liberate a site for monomer coordination.116, 118, 195 Complex I-45 is an unusual exception to this case, as the asymmetrical dimer conformation allows greater steric flexibility at the indium active site.196

The decreased activity of the chiral indium (4) and zinc (5) complexes, as well as the achiral indium complex (2) may be attributed to the higher energy barrier to lactide coordination. The decrease of activity in THF or in the presence of alcohols, which can directly compete with lactide for a coordination site, is strong evidence to support this hypothesis. The possible ease of ligand decoordination for complex I-241 compared to complex 5, as described previously for related mononuclear zinc complexes, supports the relative differences in the reactivity of the zinc species.197 This flexibility can also be invoked in comparing the greater reactivity of complex 2 compared to complex 4. Designing new dinuclear indium catalysts will require a delicate steric balance to allow isolation of thermodynamically stable complexes but also to create a favorable environment for lactide polymerization. For a more active dinucleating indium catalyst, it is important to take into account the final geometry of the indium center.

Pursuing the work on this class of dinucleating complexes will require to free one coordination site.

For this two different strategies can be used 1) design of an asymmetric proligand, 2) synthesis of an heterobimetallic complex (Figure 2-19). Pentadentate proligands, as described in this chapter, suffer from being two sterically demanding to allow efficient ROP of lactide. The removal of the two terminal amines is not an option as it will most probably lead to a dimeric complex with a similar structure than

I-134-136 - described in the first chapter. An asymmetric tetradentate proligand, on the other side, might allow dinucleation of the metal center without aggregation or dimerization phenomena. Finally, one can imagine the synthesis of an heterobimetallic complex bearing one zinc and one indium center, however the synthesis of these complexes can be challenging due to the formation of the homobimetallic complexes being more thermodynamically favored.203

Figure 2-19: Structure of a dinucleating indium complex bearing asymmetric ligand backbone (A) and a dinucleating heterobimetallic indium zinc complex (B).

2.7. Experimental procedures.

General procedures. All air and/or water sensitive reactions were carried out under N2 in an MBraun glovebox. Bruker Avance 600 MHz, 400 MHz or 300 MHz spectrometers were used to record the 1H

NMR, 13C{1H} NMR spectra and 1H{1H} NMR spectra. 1H NMR chemical shifts are given in ppm

13 1 versus residual protons in deuterated solvents as follows:  7.27 for CDCl3. C{ H} NMR chemical

13 shifts are given in ppm versus residual C in solvents as follows:  77.00 for CDCl3. Diffraction measurements for X-ray crystallography were made on Bruker X8 APEX II and Bruker APEX DUO diffractometers with graphite monochromated Mo-Kα radiation. The structures were solved by direct methods and refined by full-matrix least-squares using the SHELXTL crystallographic software of the

Bruker-AXS. Unless specified, all non-hydrogen atoms were refined with anisotropic displacement parameters, and all hydrogen atoms were constrained to geometrically calculated positions but were not refined. EA CHN analysis was performed using a Carlo Erba EA1108 elemental analyzer. The elemental composition of an unknown sample was determined by using a calibration factor. The calibration factor was determined by analyzing a suitable certified organic standard (OAS) of a known elemental composition. Molecular weights were determined by GPC-LLS using an Agilent liquid chromatograph equipped with an Agilent 1200 series pump and autosampler, three Phenogel 5 μm Narrow Bore columns

(4.6 × 300 mm with 500 Å, 103 Å and 104 Å pore size), a Wyatt Optilab differential refractometer, Wyatt tristar miniDAWN (laser light scattering detector) and a Wyatt ViscoStar viscometer. The column temperature was set at 40 °C. A flow rate of 0.5 mL/min was used and samples were dissolved in THF

(ca. 2 mg/mL), and a dn/dc value of 0.042 mL/g was used. Narrow molecular weight polystyrene standards were used for system calibration purposes.

Materials. Toluene, diethyl ether, hexane, and tetrahydrofuran were degassed and dried using alumina columns in a solvent purification system. THF was further dried over sodium/benzophenone 85 and vacuum transferred to a Straus flask where it was degassed prior to use. In addition, CH3CN, CHCl3 and CH2Cl2 were dried over CaH2 and vacuum transferred to a Straus flask where they were degassed prior to use. Deuterated chloroform (CDCl3) was dried over CaH2 and vacuum transferred to a Straus flask and then degassed through a series of freeze-pump-thaw cycles. Trifluoroacetic acid was purchased from Alfa Aesar and dried over activated 4 Å molecular sieves and stored under N2 prior to use. InCl3 was obtained from Strem Chemicals and ZnCl2 from Alfa Aesar and used without further purification.

N,N,N’-trimethylethylenediamine, N,N-dimethylethylenediamine, potassium tert-butoxide, (±)-trans- diaminocyclohexane, p-cresol and 4-tert-butylphenol were obtained from Alfa Aesar and used without further purification. NaOEt was obtained from Alfa Aesar, dissolved in dry ethanol, stirred for 16h, precipitated from solution with hexanes, filtered, washed with hexanes and dryed at 50 °C under vacuum.

Lactide samples were obtained from Purac Biomaterials and recrystallized several times from hot, dry toluene and dried under vacuum prior to use. Synthesis of (R,R), (S,S) and (±)-N,N-dimethyl-trans-1,2- diaminocyclohexane was performed according to literature procedures198-201 from (±)-trans- diaminocyclohexane and distilled at 70 °C under reduced pressure prior to use.

Synthesis of (±)-L4. Freshly distilled (±)-N,N-dimethyl-trans-1,2-diaminocyclohexane (1.38 g, 9.70 mmol) was added to a solution of 4-tert-butyl-2,6-diformylphenol (0.500 g, 2.42 mmol) in methanol (20 mL). The solution was stirred at room temperature for 19 h. The methanol was removed in vacuo at room temperature. Excess (±)-N,N-dimethyl-trans-1,2-diaminocyclohexane was distilled from the residual mixture at 70 °C under dynamic vacuum. The product is obtained as a dark orange oil (0.979 g,

89%) and used without further purification. Attempts at further purification were unsuccessful. 1H NMR

(300 MHz, CDCl3, 25 °C): δ 8.53 (2H, br s, N=CH), 7.63 (2H, br s, Ar-H), 3.28 (2H, m, CH-N), 2.66

(2H, m, CH-N), 2.30 (12H, s, N-(CH3)2), 1.33 (9H, s, Ar-(CH3)3).

Synthesis of (±)-L5. (±)-L5 was prepared and purified in an analogous manner to (±)-L4 from 4- methyl-2,6-diformylphenol (0.500 g, 3.05 mmol) (yield: 1.07 g, 85%). The compound was used without

1 further purification. Attempts at further purification were unsuccessful. H NMR (300 MHz, CDCl3, 25

°C): δ 8.49 (2H, br s, N=CH), 7.43 (2H, br s, Ar-H), 3.25 (2H, m, CH-N), 2.63 (2H, m, CH-N), 2.27

+ + (12H, s, N(CH3)2), 2.22 (3H, s, Ar-CH3). EI-LRMS: calc. [M] 412.32, found [M] 412.

Synthesis of (±)-L6. NaBH4 (0.326 g, 8.61 mmol) was added in small portions to a solution of L4

(0.979 g, 2.15 mmol) in methanol (50 mL). The solution was stirred at room temperature for 16 h. The methanol was removed in vacuo at room temperature. The resulting yellow residue was redissolved in

CH2Cl2 (10 mL) and distilled water (10 mL). Organics were extracted with more CH2Cl2 (2×10 mL), combined, dried over MgSO4 and filtered. The resulting mixture was concentrated to 10 mL and filtered through a small plug of alumina. The filtrate was pumped to dryness yielding a clear yellow oil (0.819 g, 83%). The compound was used without further purification. Attempts at further purification were unsuccessful. The product was confirmed by 1H NMR spectroscopy to be a mixture of diastereomers.

1 H NMR (300 MHz, CDCl3, 25 °C): δ 7.02 (s, 2H, Ar-H), 7.00 (s, 2H, Ar-H), 4.00-3.90 (4H, overlapping doublets, NH-CH2-Ar), 3.80-3.70 (4H, overlapping doublets, NH-CH2-Ar), 2.16 (12H, s, N-(CH3)2),

+ + 2.15 (12H, s, N-(CH3)2), 1.28 (18H, s, Ar-(CH3)3). EI-LRMS: calc. [M] 458.40, found [M] 458.

Synthesis of (±)-L7. (±)-L7 was prepared and purified in an analogous manner to (±)-L6 from (±)-L5

(1.07 g, 2.59 mmol) (yield: 0.864 g, 80%). The compound was used without further purification.

Attempts at further purification were unsuccessful. The product was confirmed by 1H NMR spectroscopy

1 to be a mixture of diastereomers. H NMR (300 MHz, CDCl3, 25 °C): δ 6.85 (s, 2H, Ar-H), 6.83 (s, 2H,

Ar-H), 4.03-3.90 (4H, overlapping doublets, N-CH2-Ar), 3.80-3.70 (4H, overlapping doublets, N-CH2-

Ar), 2.24 (3H, s, Ar-CH3), 2.23 (3H, s, Ar-CH3), 2.17 (12H, s, N-(CH3)2), 2.16 (12H, s, N-(CH3)2). EI-

LRMS: calc. [M]+ 416.35, found [M]+ 416.

Synthesis of RR/RR-L7. RR/RR-L7 was prepared and purified in an analogous manner to (±)-L7;

1 RR/RR-L5 was formed in situ. H NMR (400 MHz, CDCl3, 25 °C): δ 6.82 (s, 2H, Ar-H), 3.91 (2H, d,

JHH = 14 Hz, N-CH2-Ar), 3.71 (2H, d, JHH = 14 Hz, N-CH2-Ar), 2.35-2.25 (2H, m, CH-N), 2.22 (3H, s,

Ar-CH3), 2.16 (12H, s, N-(CH3)2).

Synthesis of RR/RR-[(L7)In2Cl4](μ-OEt) (4). Sodium ethoxide (0.209 g, 3.07 mmol) was transferred to a solution of proligand RR/RR-L7 (0.320 g, 0.769 mmol) in toluene (10 mL). The solution was stirred at room temperature for 8 h, after which indium trichloride was added (0.340 g, 1.54 mmol).

The solution was stirred at room temperature for 16 h. Then the solution was filtered through glass fiber to obtain a clear pale yellow solution. The mixture was concentrated to 5 mL and place in freezer at −35

°C for 16 h causing crystals to form. The supernatant solution was decanted and the crystals were washed with cold toluene (2×5 mL) and stirred for 30 min with hexanes, then dried. A second crop can be obtained from the supernatant in a similar manner. The product was isolated as a white powder (0.134 g, 21%). Crystals suitable for X-ray analysis were grown from a saturated solution in CHCl3 and diethyl

1 ether at room temperature. H NMR (400 MHz, CDCl3, 25 °C): δ 6.82 (2H, s, Ar-H), 5.54 (2H, dd, JHH

= 14 Hz, Ar-CH2-N), 4.51-4.41 (1H, m, O-CH2-CH3), 4.40-4.30 (1H, m, O-CH2-CH3), 3.87 (2H, dd, JHH

= 14 Hz, Ar-CH2-N), 3.27 (2H, t, JHH = 11 Hz, N-CH), 2.81 (6H, s, N-(CH3)2 ), 2.61 (2H, d), 2.45 (2H, d, CH-N), 2.40 (6H, s, N-(CH3)2), 2.33 (2H, d), 2.23 (3H, s, Ar-CH3), 1.92-1.80 (6H, m), 1.48 (3H, t,

13 1 JHH = 7 Hz, O-CH2-CH3), 1.44-1.00 (8H, m). C{ H} NMR (100 MHz, CDCl3, 25 °C): δ 160.26, 134.56,

127.86, 122.41, 64.02, 61.74, 54.05, 50.34, 44.09, 38.36, 31.05, 24.78, 24.31, 20.73, 20.01, 19.30. Anal.

Calc. for C27H48Cl4In2N4O2: C, 38.97; H, 5.81; N, 6.73. Found: C, 38.62; H, 5.51; N, 6.19.

Synthesis of RR/RR-[(L7)Zn2Cl2](μ-OEt) (5). Sodium ethoxide (0.133 g, 1.97 mmol) was transferred to a solution of proligand RR/RR-L7 (0.205 g, 0.492 mmol) in toluene (10 mL). The solution was stirred at room temperature for 8 h, after which zinc dichloride was added (0.134 g, 0.983 mmol).

The solution was stirred at room temperature for 16 h. Then the solvent was removed in vacuo at room temperature. The residual solid was redissolved in 5 mL of CH2Cl2 and filter through celite. The resulting mixture was further concentrated to 2 mL, and ether was adding causing a solid to precipitate. The precipitate was allowed to settle and the solution was decanted; the wet precipitate was washed with hexanes 3 times and dried to afford complex 5 as an off-white powder (0.183 g, 56%). 1H NMR (400

MHz, CDCl3, 25 °C): δ 6.74 (2H, s, Ar-H), 4.44 (2H, dd, JHH = 13, 3 Hz, Ar-CH2-N), 4.23-4.15 (1H, m,

O-CH2-CH3), 4.06-3.99 (1H, m, O-CH2-CH3), 3.68 (2H, dd, JHH = 13, 4 Hz, Ar-CH2-N), 2.67 (6H, s, N-

(CH3)2 ), 2.55 (2H, dt, JHH = 11, 3 Hz, CH-N), 2.29 (2H, d), 2.24 (2H, m, CH-N), 2.19 (9H, singlets overlapping, N-(CH3)2 and Ar-CH3), 1.93 (2H, d), 1.81 (2H, d), 1.72 (2H, d), 1.35 (3H, t, JHH = 7 Hz,

13 1 O-CH2-CH3), 1.28-0.92 (10H, m). C{ H} NMR (100 MHz, CDCl3, 25 °C): δ 160.43, 131.68, 123.86,

121.52, 69.90, 60.74, 51.81, 47.65, 44.77, 37.98, 30.61, 24.61, 24.32, 21.25, 20.10. Anal. Calc. for

C27H48Cl2Zn2N4O2: C, 48.96; H, 7.30; N, 8.46. Found: C, 48.99; H, 7.29; N, 8.07.

Synthesis of RR/RR-[(L7)Zn2Cl2](μ-Cl) (6). Potassium tert-butoxide (0.077 g, 0.68 mmol) was transferred to a solution of proligand RR/RR-L7 (0.295 g, 0.709 mmol) in toluene (10 mL). The solution was stirred at room temperature for 16 h. Then the toluene was removed in vacuo at room temperature yielding a yellow solid residue. The potassium salt was redissolved in toluene (5 mL) and zinc dichloride

(0.193 g, 1.42 mmol) was transferred to this solution using toluene (5 mL). The solution was stirred at room temperature for 16 h. Then the toluene was removed in vacuo at room temperature. The residual solid was redissolved in CH2Cl2 (5 mL), and filtered through celite. The resulting mixture was concentrated to 2 mL and diethyl ether was added causing a precipitate to form. The precipitate was

89 allowed to settle and the solution was decanted; the wet precipitate was washed with hexanes 3 times

1 and dried to afford complex 2 as an off-white powder (0.258 g, 56%). H NMR (400 MHz, CDCl3, 25

°C): δ 6.79 (2H, s, Ar-H), 4.11 (2H, app t, JHH = 12 Hz, Ar-CH2-N), 3.65 (2H, d, JHH = 11 Hz, Ar-CH2-

N), 2.64 (2H, m), 2.62 (6H, s, N-(CH3)2), 2.49 (6H, s, N-(CH3)2), 2.41-2.51 (4H, m), 2.17 (3H, s, Ar-

13 1 CH3), 1.98 (2H, d), 1.90-1.82 (4H, m), 1.49 (2H, t), 1.32-1.10 (8H, m). C{ H} NMR (150 MHz,

CDCl3, 25 °C): δ 158.97, 130.17, 125.28, 124.50, 85.66, 57.38, 51.98, 43.39, 36.68, 31.66, 24.67, 24.23,

21.06, 19.72. Anal. Calc. for C25H43Cl3Zn2N4O: C, 46.00; H, 6.64; N, 8.58. Found: C, 45.66; H, 6.51;

N, 8.43.

Alternative synthesis of RR/RR-[(L7)Zn2Cl2](μ-OEt) (5). Sodium ethoxide (0.021 g, 0.31 mmol) was transferred to a solution of complex 6 (0.100 g, 0.153 mmol) in toluene (10 mL). The solution was stirred at room temperature for 16 h. Then the toluene was removed in vacuo at room temperature yielding a yellow solid residue. The residual solid was redissolved in CH2Cl2 (5 mL), and filtered through celite. Solvent was removed in vacuo and the resulting solid was washed with hexanes 3 times to afford complex 5 as an off-white powder (0.040 g, 40%).

Synthesis of RR/RR-[(L5)Zn2Cl2](μ-Cl) (7). Potassium tert-butoxide (0.177 g, 0.158 mmol) was transferred to a solution of proligand RR/RR-L5 (0.733 g, 1.61 mmol) in toluene (5 mL). The solution was stirred at room temperature for 16 h. Then the toluene was removed under vacuo at room temperature yielding an orange solid residue. The solid was redissolved in hexane and filtered through celite. Hexane was removed in vacuo to afford an orange powder (0.653 g, 90%). The potassium salt

(0.316 g, 0.701 mmol) was redissolved in toluene (5 mL) and zinc dichloride (0.191 g, 1.40 mmol) was transferred to this solution using toluene (5 mL). The solution was stirred at room temperature for 16 h.

Then the toluene was removed in vacuo at room temperature. The residual solid was redissolved in

CH2Cl2 (5 mL), and filtered through celite. Solvent was removed in vacuo and the residual solid was washed with hexanes 3 times and dried to afford complex 7 as an orange powder (0.290 g, 64%, i.e. total yield: 58%). Crystals suitable for X-ray analysis were grown from a saturated solution in THF and ether

1 at room temperature. H NMR (600 MHz, CDCl3, 25 °C): δ 8.29 (2H, s, Ar-CH=N), 7.14 (2H, s, Ar-H),

3.41 (2H, t, JHH = 9 Hz, CH-N), 2.92 (2H, t, JHH = 9 Hz, CH-N), 2.63 (6H, s, N-(CH3)2), 2.48-2.46 (2H,

m), 2.43 (6H, s, N-(CH3)2), 2.24 (3H, s, Ar-H), 2.05 (2H, d), 1.96-1.90 (4H, m), 1.30-1.40 (8H, m).

13 1 C{ H} NMR (150 MHz, CDCl3, 25 °C): δ 166.31, 165.03, 140.57, 124.47, 121.55, 65.85, 60.03, 42.98,

35.95, 29.37, 24.32, 23.88, 20.98, 19.32. Anal. Calc. for C25H39Cl3Zn2N4O: C, 46.29; H, 6.06; N, 8.64.

Found: C, 46.02; H, 6.04; N, 8.33.

One pot synthesi of [(L3)In2Cl4](μ-OEt) (2). Sodium ethoxide (0.626 g, 9.20 mmol) was transferred to a solution of proligands L3 (0.871 g, 2.30 mmol) in hexane (15 mL). The solution was stirred at room temperature for 5 h, then indium trichloride was added (1.018 g, 4.603 mmol) and the mixture was stirred for 16 h at room temperature causing a yellow solid to precipitate. The solid was allowed to settle, and the hexane was decanted. The residual wet solid was dried, dissolved in CH2Cl2 (15 mL) and filtered through celite. The mixture was concentrated to 5 mL causing an off-white solid to precipitate. The

CH2Cl2 was decanted and the resulting solid was dried to afford pure complex 2 (0.620 g, 34%). Residual solutions were combined and concentrated further causing more precipitate to form, 1H NMR revealed the resulting solid to be a mixture of complex 1 and 2 in addition to unknown by-products. 1H NMR

(600 MHz, CDCl3, 25 °C): δ 7.03 (2H, s, Ar-H), 5.36 (2H, d, JHH = 18 Hz, Ar-CH2-N), 4.44-4.39 (1H, m, OCH2CH3), 4.34-4.29 (1H, m, OCH2CH3), 3.51 (2H, t, JHH = 12 Hz, NCH2CH2N), 3.25 (2H, d, JHH

= 18 Hz, Ar-CH2-N), 3.04 (2H, t, JHH = 12 Hz, NCH2CH2N), 3.02 (6H, s, N(CH3)), 2.86 (6H, br s, N-

(CH3)2), 2.57 (6H, br s, N(CH3)2), 2.43 (2H, d, JHH = 12 Hz, NCH2CH2N), 2.04 (2H, d, JHH = 12 Hz,

13 1 NCH2CH2N), 1.46 (3H, t, OCH2CH3), 1.28 (9H, s, Ar-(CH3)3). C{ H} NMR (150 MHz, CDCl3, 25

°C): δ 158.93, 141.96, 131.61, 122.03, 63.46, 61.53, 55.41, 50.82, 47.96, 47.92, 45.60, 33.82, 31.38,

19.14. Anal. Calc. for C24H46Cl4In2N4O2: C, 36.30; H, 5.84; N, 7.06. Found: C, 36.61; H, 5.86; N, 7.01.

Representative polymerization of rac-lactide. Rac-LA or meso-LA (200 equiv) in CH2Cl2 or THF was added to a solution of complex (5 mg) in CH2Cl2 or THF to obtain a 2mM concentration of catalyst.

The mixture was allowed to stir at room temperature in a vial or at higher temperature in a vacuum- sealed bomb. For immortal polymerization, BnOH (5 equiv) was added to the rac-LA solution prior addition to the complex. The solvent was then removed in vacuo and a small portion of the crude

1 1 1 polymer was tested for conversion and tacticity via H and H{ H} NMR spectroscopy (25 °C, CDCl3).

The remaining crude polymer was redissolved in a minimum of dichloromethane (1-2 mL). Methanol

(2-5 mL) was then added to this solution causing precipitation of the polymer only for the polymer samples obtained in CH2Cl2. For those obtained in THF, the crude polymers were soluble in this CH2Cl2-

MeOH mixture and could not be purified. The solution was allowed to settle and the supernatant solution was removed. This process was repeated 2 more times, and the resulting polymer was dried under vacuum. The polymer was tested for the presence of remaining catalyst or monomer using 1H NMR spectroscopy before being tested for molecular weight and PDI using GPC in THF.

References

1. Barnes, D. K., Nature 2002, 416, 808-809. 2. Barnes, D. K.; Galgani, F.; Thompson, R. C.; Barlaz, M., Philos. T. Roy. Soc. B 2009, 364, 1985- 1998. 3. Al-Salem, S.; Lettieri, P.; Baeyens, J., Waste Manage. 2009, 29, 2625-2643. 4. Gross, R. A.; Kalra, B., Science 2002, 297, 803-807. 5. Mohanty, A.; Misra, M.; Drzal, L., J. Polym. Environ. 2002, 10, 19-26. 6.“Plastics – the facts 2015” (PlasticsEurope, Brussels, Belgium, 2015) http://www.plasticseurope.org/Document/plastics---the-facts-2015.aspx (Accessed in March 2016). 7. Thompson, R. C.; Moore, C. J.; Vom Saal, F. S.; Swan, S. H., Philos. T. Roy. Soc. B 2009, 364, 2153-2166. 8. Thompson, R. C.; Olsen, Y.; Mitchell, R. P.; Davis, A.; Rowland, S. J.; John, A. W.; McGonigle, D.; Russell, A. E., Science 2004, 304, 838-838. 9. Gielen, D.; Boshell, F.; Saygin, D., Nat. Mater. 2016, 15, 117-120. 10. Aeschelmann, F.; Carus, M., Ind. Biotechnol. 2015, 11, 154-159. 11. Amass, W.; Amass, A.; Tighe, B., Polym. Int. 1998, 47, 89-144. 12. Vink, E. T.; Rábago, K.; Glassner, D. A.; Springs, B.; O'Connor, R. P.; Kolstad, J.; Gruber, P. R., Macromol. Biosci. 2004, 4, 551-564. 13. Lovett, J., "Corbion Purac successfully develops PLA resin from second generation feedstocks". In http://www.corbion.com/media/press-releases?newsId=1955535: 2015. (Accessed in March 2016). 14. http://www.natureworksllc.com/News-and-Events/Press-Releases/2014/10-30-14-DOE-Grant- to-NatureWorks-to-Transform-Biogas-into-Lactic-Acid-for-Ingeo. (Accessed in March 2016). 15. Auras, R. A.; Lim, L.-T.; Selke, S. E.; Tsuji, H., Poly (lactic acid): synthesis, structures, properties, processing, and applications. John Wiley & Sons: 2011; Vol. 10. 16. Achmad, F.; Yamane, K.; Quan, S.; Kokugan, T., Chem. Eng. J. 2009, 151, 342-350. 17. Hartmann, M., High molecular weight polylactic acid polymers. In Biopolymers from renewable resources, Springer: 1998; pp 367-411. 18. Platel, R. H.; Hodgson, L. M.; Williams, C. K., Polym. Rev. 2008, 48, 11-63. 19. Wu, J.; Yu, T.-L.; Chen, C.-T.; Lin, C.-C., Coord. Chem. Rev. 2006, 250, 602-626. 20. Garlotta, D., J. Polym. Environ. 2001, 9, 63-84. 21. Tsui, A.; Wright, Z. C.; Frank, C. W., Annu. Rev. Chem. Biomol. Eng. 2013, 4, 143-170. 22. Lü, J.-M.; Wang, X.; Marin-Muller, C.; Wang, H.; Lin, P. H.; Yao, Q.; Chen, C., Expert Rev. Mol. Diagn. 2014. 23. Behera, B., J. Med. Pharma. Innov. 2013, 1, 1-5. 24. Baker, G. L.; Vogel, E. B.; Smith III, M. R., Polym. Rev. 2008, 48, 64-84. 25. Drumright, R. E.; Gruber, P. R.; Henton, D. E., Adv. Mater. 2000, 12, 1841-1846. 26. Auras, R.; Harte, B.; Selke, S., Macromol. Biosci. 2004, 4, 835-864. 27. Kiesewetter, M. K.; Shin, E. J.; Hedrick, J. L.; Waymouth, R. M., Macromolecules 2010, 43, 2093-2107. 28. Dittrich, V. W.; Schulz, R. C., Angew. Makromol. Chem. 1971, 15, 109-126. 29. Arlman, E.; Cossee, P., J. Catal. 1964, 3, 99-104. 30. Kricheldorf, H. R.; Berl, M.; Scharnagl, N., Macromolecules 1988, 21, 286-293. 31.Brunelle, D. J., Hanser Publishers, 1993 1993, 361. 32. Endo, M.; Aida, T.; Inoue, S., Macromolecules 1987, 20, 2982-2988.

33. Inoue, S., J. Polym. Sci. Pol. Chem. 2000, 38, 2861-2871. 34 Montaudo, G.; Montaudo, M. S.; Puglisi, C.; Samperi, F.; Spassky, N.; LeBorgne, A.; Wisniewski, M., Macromolecules 1996, 29, 6461-6465. 35. Thakur, K. A.; Kean, R. T.; Zell, M. T.; Padden, B. E.; Munson, E. J., Chem. Commun. 1998, 1913-1914. 36. Zell, M. T.; Padden, B. E.; Paterick, A. J.; Thakur, K. A.; Kean, R. T.; Hillmyer, M. A.; Munson, E. J., Macromolecules 2002, 35, 7700-7707. 37. Ovitt, T. M.; Coates, G. W., J. Am. Chem. Soc. 1999, 121, 4072-4073. 38. Ovitt, T. M.; Coates, G. W., J. Am. Chem. Soc. 2002, 124, 1316-1326. 39. Ovitt, T. M.; Coates, G. W., J. Polym. Sci. Pol. Chem.2000, 38, 4686-4692. 40. Kuruppalil, Z., Int. J. Eng. Res. Innov. 2011, 3, 59-69. 41. Kyrikou, I.; Briassoulis, D., J. Polym. Environ. 2007, 15, 125-150. 42. Vink, E. T.; Rabago, K. R.; Glassner, D. A.; Gruber, P. R., Polym. Degrad. Stab. 2003, 80, 403- 419. 43. Vink, E. T.; Glassner, D. A.; Kolstad, J. J.; Wooley, R. J.; O’Connor, R. P., Ind. Biotechnol. 2007, 3, 58-81. 44. Davis, G.; Song, J., Ind. Crops Prod. 2006, 23, 147-161. 45. Hopewell, J.; Dvorak, R.; Kosior, E., Philos. T. Roy. Soc. B 2009, 364, 2115-2126. 46. Shen, J. J. 2011, http://nature.berkeley.edu/classes/es196/projects/2011final/ShenJ _2011.pdf (Accessed in March 2016). 47.Groot, W. J.; Borén, T., Int. J. Life Cycle Ass. 2010, 15, 970-984. 48. Gironi, F.; Piemonte, V., Environ. Prog. Sustain. Energy 2011, 30, 459-468. 49. Papong, S.; Malakul, P.; Trungkavashirakun, R.; Wenunun, P.; Chom-in, T.; Nithitanakul, M.; Sarobol, E., J. Clean. Prod. 2014, 65, 539-550. 50. Madival, S.; Auras, R.; Singh, S. P.; Narayan, R., J. Clean. Prod. 2009, 17, 1183-1194. 51. Yates, M. R.; Barlow, C. Y., Resour. Conserv. Recycl. 2013, 78, 54-66. 52. Iffland, K.; Carus, M.; de Bie, F.; Diels, L.; van Haveren, J.; Willems, P.; Ravenstijn, J.; Vink, E.; Wagemann, K., 2015. 53. Associates, F., LCI summary for PLA and PET 12-once water bottles. In http://www.petresin.org/pdf/FranklinPETPLAlifecycleanal_12-oz.pdf, 2007. (Accessed in March 2016). 54. Hong, M.; Chen, E. Y.-X., Nat. Chem. 2016, 8, 42-49. 55. Chiellini, E.; Corti, A., Oxo-biodegradable Plastics: Who They Are and to What They Serve— Present Status and Future Perspectives. In Polyolefin Compounds and Materials, Springer: 2016; pp 341-354. 56. Jedrzejas, M. J.; Setlow, P., Chem. Rev. 2001, 101, 607-618. 57. Mitic, N.; Smith, S. J.; Neves, A.; Guddat, L. W.; Gahan, L. R.; Schenk, G., Chem. Rev. 2006, 106, 3338-3363. 58. Delferro, M.; Marks, T. J., Chem. Rev. 2011, 111, 2450-2485. 59. McInnis, J. P.; Delferro, M.; Marks, T. J., Acc. Chem. Res. 2014, 47, 2545-2557. 60. Romain, C.; Thevenon, A.; Saini, P. K.; Williams, C. K., 2015. 61. Klaus, S.; Lehenmeier, M. W.; Anderson, C. E.; Rieger, B., Coord. Chem. Rev. 2011, 255, 1460- 1479. 62. O'Keefe, B. J.; Monnier, S. M.; Hillmyer, M. A.; Tolman, W. B., J. Am. Chem. Soc. 2001, 123, 339-340. 63. O'Keefe, B. J.; Breyfogle, L. E.; Hillmyer, M. A.; Tolman, W. B., J. Am. Chem. Soc. 2002, 124, 4384-4393.

64. Pietrangelo, A.; Hillmyer, M. A.; Tolman, W. B., Chem. Commun. 2009, 2736-2737. 65. Pietrangelo, A.; Knight, S. C.; Gupta, A. K.; Yao, L. J.; Hillmyer, M. A.; Tolman, W. B., J. Am. Chem. Soc. 2010, 132, 11649-11657. 66. Fliedel, C.; Vila‐Viçosa, D.; Calhorda, M. J.; Dagorne, S.; Avilés, T., ChemCatChem 2014, 6, 1357-1367. 67. Wang, L.; Fadlallah, S.; Bellini, C.; Orione, C.; Dorcet, V.; Carpentier, J.-F.; Sarazin, Y., Organometallics 2015, 34, 1321-1327. 68. Blake, M. P.; Schwarz, A. D.; Mountford, P., Organometallics 2011, 30, 1202-1214. 69. Carmalt, C. J.; King, S. J., Coord. Chem. Rev. 2006, 250, 682-709. 70. Chamberlain, B. M.; Cheng, M.; Moore, D. R.; Ovitt, T. M.; Lobkovsky, E. B.; Coates, G. W., J. Am. Chem. Soc. 2001, 123, 3229-3238. 71. Chisholm, M. H.; Gallucci, J.; Phomphrai, K., Inorg. Chem. 2002, 41, 2785-2794. 72. Drouin, F.; Oguadinma, P. O.; Whitehorne, T. J.; Prud’homme, R. E.; Schaper, F., Organometallics 2010, 29, 2139-2147. 73. Drouin, F.; Whitehorne, T. J.; Schaper, F., Dalton Trans. 2011, 40, 1396-1400. 74. Whitehorne, T. J.; Schaper, F., Chem. Commun. 2012, 48, 10334-10336. 75. Whitehorne, T. J.; Schaper, F., Inorg. Chem. 2013, 52, 13612-13622. 76. Whitehorne, T. J.; Schaper, F., Can. J. Chem. 2013, 92, 206-214. 77. Chen, H.-L.; Chuang, H.-J.; Huang, B.-H.; Lin, C.-C., Inorg. Chem. Commun. 2013, 35, 247-251. 78. Di Iulio, C.; Middleton, M.; Kociok‐Köhn, G.; Jones, M. D.; Johnson, A. L., Eur. J. Inorg. Chem. 2013, 2013, 1541-1554. 79. Crespo-Velasco, N.-T.; Guerrero-Ram, L.-G.; Flores-Alamo, M.; Muñoz-Hernández, M.-Á., Main Group Chem. 2015, 14, 141-157. 80. Bouyahyi, M.; Roisnel, T.; Carpentier, J.-F., Organometallics 2009, 29, 491-500. 81. Otero, A.; Fernández-Baeza, J.; Sánchez-Barba, L. F.; Tejeda, J.; Honrado, M.; Garcés, A. s.; Lara-Sánchez, A.; Rodríguez, A. M., Organometallics 2012, 31, 4191-4202. 82. Grunova, E.; Roisnel, T.; Carpentier, J.-F., Dalton Trans. 2009, 9010-9019. 83. Fortun, S.; Daneshmand, P.; Schaper, F., Angew. Chem. Int. Ed. Engl. 2015, 127, 13873-13876. 84. Fliedel, C.; Rosa, V.; Alves, F. M.; Martins, A. M.; Avilés, T.; Dagorne, S., Dalton Trans. 2015, 44, 12376-12387. 85. Huang, B.-H.; Lin, C.-N.; Hsueh, M.-L.; Athar, T.; Lin, C.-C., Polymer 2006, 47, 6622-6629. 86. Shueh, M.-L.; Wang, Y.-S.; Huang, B.-H.; Kuo, C.-Y.; Lin, C.-C., Macromolecules 2004, 37, 5155-5162. 87. Horeglad, P.; Kruk, P.; Pécaut, J., Organometallics 2010, 29, 3729-3734. 88. Williams, C. K.; Breyfogle, L. E.; Choi, S. K.; Nam, W.; Young, V. G.; Hillmyer, M. A.; Tolman, W. B., J. Am. Chem. Soc. 2003, 125, 11350-11359. 89. Douglas, A. F.; Patrick, B. O.; Mehrkhodavandi, P., Angew. Chem. Int. Ed. Engl. 2008, 120, 2322-2325. 90. Acosta-Ramírez, A.; Douglas, A. F.; Yu, I.; Patrick, B. O.; Diaconescu, P. L.; Mehrkhodavandi, P., Inorg. Chem. 2010, 49, 5444-5452. 91. Xu, C.; Yu, I.; Mehrkhodavandi, P., Chem. Commun. 2012, 48, 6806-6808. 92. Yu, I.; Acosta-Ramírez, A.; Mehrkhodavandi, P., J. Am. Chem. Soc. 2012, 134, 12758-12773. 93. Osten, K. M.; Yu, I.; Duffy, I. R.; Lagaditis, P. O.; Yu, J. C.-C.; Wallis, C. J.; Mehrkhodavandi, P., Dalton Trans. 2012, 41, 8123-8134. 94. Osten, K. M.; Aluthge, D. C.; Patrick, B. O.; Mehrkhodavandi, P., Inorg. Chem. 2014, 53, 9897- 9906.

95. Yu, I.; Ebrahimi, T.; Hatzikiriakos, S.; Mehrkhodavandi, P., Dalton Trans. 2015, 44, 14248- 14254. 96. Osten, K.; Aluthge, D.; Mehrkhodavandi, P., Dalton Trans. 2015, 44, 6126-6139. 97. Fang, J.; Yu, I.; Mehrkhodavandi, P.; Maron, L., Organometallics 2013, 32, 6950-6956. 98. Wu, J.-C.; Huang, B.-H.; Hsueh, M.-L.; Lai, S.-L.; Lin, C.-C., Polymer 2005, 46, 9784-9792. 99. Chen, H.-Y.; Tang, H.-Y.; Lin, C.-C., Macromolecules 2006, 39, 3745-3752. 100. Hung, W. C.; Huang, Y.; Lin, C. C., J. Polym. Sci. Pol. Chem. 2008, 46, 6466-6476. 101. Hung, W.-C.; Lin, C.-C., Inorg. Chem. 2008, 48, 728-734. 102. Darensbourg, D. J.; Karroonnirun, O., Inorg. Chem. 2010, 49, 2360-2371. 103. Mandal, D.; Chakraborty, D.; Ramkumar, V.; Chand, D. K., RSC Adv. 2016, 6, 21706-21718. 104. Kang, Y. Y.; Park, H.-R.; Lee, M. H.; An, J.; Kim, Y.; Lee, J., Polyhedron 2015, 95, 24-29. 105. Tang, H.-Y.; Chen, H.-Y.; Huang, J.-H.; Lin, C.-C., Macromolecules 2007, 40, 8855-8860. 106. Doherty, S.; Errington, R. J.; Housley, N.; Clegg, W., Organometallics 2004, 23, 2382-2388. 107. Chuang, H. J.; Chen, H. L.; Huang, B. H.; Tsai, T. E.; Huang, P. L.; Liao, T. T.; Lin, C. C., J. Polym. Sci. Pol. Chem. 2013, 51, 1185-1196. 108. Chuang, H. J.; Chen, H. L.; Ye, J. L.; Chen, Z. Y.; Huang, P. L.; Liao, T. T.; Tsai, T. E.; Lin, C. C., J. Polym. Sci. Pol. Chem. 2013, 51, 696-707. 109. Huang, Y.; Hung, W. C.; Liao, M. Y.; Tsai, T. E.; Peng, Y. L.; Lin, C. C., J. Polym. Sci. Pol. Chem. 2009, 47, 2318-2329. 110. Hild, F.; Neehaul, N.; Bier, F.; Wirsum, M.; Gourlaouen, C.; Dagorne, S., Organometallics 2013, 32, 587-598. 111. Takashima, Y.; Nakayama, Y.; Watanabe, K.; Itono, T.; Ueyama, N.; Nakamura, A.; Yasuda, H.; Harada, A.; Okuda, J., Macromolecules 2002, 35, 7538-7544. 112. Zhao, N.; Wang, Q.; Hou, G.; Song, H.; Zi, G., Inorg. Chem. Commun. 2014, 41, 6-10. 113. Russell, S. K.; Gamble, C. L.; Gibbins, K. J.; Juhl, K. C.; Mitchell, W. S.; Tumas, A. J.; Hofmeister, G. E., Macromolecules 2005, 38, 10336-10340. 114. Hayes, C. E.; Sarazin, Y.; Katz, M. J.; Carpentier, J.-F.; Leznoff, D. B., Organometallics 2013, 32, 1183-1192. 115. Sinenkov, M.; Kirillov, E.; Roisnel, T.; Fukin, G.; Trifonov, A.; Carpentier, J.-F., Organometallics 2011, 30, 5509-5523. 116. Aluthge, D. C.; Yan, E. X.; Ahn, J. M.; Mehrkhodavandi, P., Inorg. Chem. 2014, 53, 6828-6836. 117. Grunova, E.; Kirillov, E.; Roisnel, T.; Carpentier, J.-F., Organometallics 2008, 27, 5691-5698. 118. Aluthge, D. C.; Patrick, B. O.; Mehrkhodavandi, P., Chem. Commun. 2013, 49, 4295-4297. 119. Aluthge, D.; Ahn, J.; Mehrkhodavandi, P., Chem. Sci. 2015, 6, 5284-5292. 120. Platel, R. H.; Hodgson, L. M.; White, A. J.; Williams, C. K., Organometallics 2007, 26, 4955- 4963. 121. Platel, R. H.; White, A. J.; Williams, C. K., Chem. Commun. 2009, 4115-4117. 122. Platel, R. H.; White, A. J.; Williams, C. K., Inorg. Chem. 2011, 50, 7718-7728. 123. Ma, H.; Spaniol, T. P.; Okuda, J., Angew. Chem. Int. Ed. Engl. 2006, 118, 7982-7985. 124. Peckermann, I.; Kapelski, A.; Spaniol, T. P.; Okuda, J., Inorg. Chem. 2009, 48, 5526-5534. 125. Kapelski, A.; Okuda, J., J. Polym. Sci. Pol. Chem. 2013, 51, 4983-4991. 126. Bonnet, F.; Cowley, A. R.; Mountford, P., Inorg. Chem. 2005, 44, 9046-9055. 127. Dyer, H. E.; Huijser, S.; Susperregui, N.; Bonnet, F.; Schwarz, A. D.; Duchateau, R.; Maron, L.; Mountford, P., Organometallics 2010, 29, 3602-3621. 128. Dugah, D. T.; Skelton, B. W.; Delbridge, E. E., Dalton Trans. 2009, 1436-1445.

129. Bouyahyi, M.; Ajellal, N.; Kirillov, E.; Thomas, C. M.; Carpentier, J. F., Chem. Eur. J. 2011, 17, 1872-1883. 130. Nie, K.; Feng, T.; Song, F.; Zhang, Y.; Sun, H.; Yuan, D.; Yao, Y.; Shen, Q., Sci. China Chem. 2014, 57, 1106-1116. 131. Hu, M.; Wang, M.; Zhang, P.; Wang, L.; Zhu, F.; Sun, L., Inorg. Chem. Commun. 2010, 13, 968-971. 132. Hu, M.; Wang, M.; Zhu, H.; Zhang, L.; Zhang, H.; Sun, L., Dalton Trans. 2010, 39, 4440-4446. 133. Su, C.-K.; Chuang, H.-J.; Li, C.-Y.; Yu, C.-Y.; Ko, B.-T.; Chen, J.-D.; Chen, M.-J., Organometallics 2014, 33, 7091-7100. 134. Kim, Y.; Jnaneshwara, G.; Verkade, J. G., Inorg. Chem. 2003, 42, 1437-1447. 135. Kim, Y.; Verkade, J. G. In Living polymerization of lactide using titanium alkoxide catalysts, Macromolecular Symposia, 2005; Wiley Online Library: 2005; pp 105-118. 136. Chamberlain, B. M.; Sun, Y.; Hagadorn, J. R.; Hemmesch, E. W.; Young, V. G.; Pink, M.; Hillmyer, M. A.; Tolman, W. B., Macromolecules 1999, 32, 2400-2402. 137. Bukhaltsev, E.; Frish, L.; Cohen, Y.; Vigalok, A., Org. lett. 2005, 7, 5123-5126. 138. Honrado, M.; Otero, A.; Fernández-Baeza, J.; Sánchez-Barba, L. F.; Garcés, A.; Lara-Sánchez, A. n.; Rodrı́guez, A. M., Organometallics 2014, 33, 1859-1866. 139. Thibault, M.-H.; Fontaine, F.-G., Dalton Trans. 2010, 39, 5688-5697. 140. Yu, X.-F.; Wang, Z.-X., Dalton Trans. 2013, 42, 3860-3868. 141. Wang, Y.; Ma, H., Chem. Commun. 2012, 48, 6729-6731. 142. Normand, M.; Roisnel, T.; Carpentier, J.-F.; Kirillov, E., Chem. Commun. 2013, 49, 11692- 11694. 143. Maudoux, N.; Roisnel, T.; Dorcet, V.; Carpentier, J. F.; Sarazin, Y., Chem. Eur. J. 2014, 20, 6131-6147. 144. Pang, X.; Duan, R.; Li, X.; Chen, X., Polym. Chem. 2014, 5, 3894-3900. 145. Pang, X.; Duan, R.; Li, X.; Gao, B.; Sun, Z.; Wang, X.; Chen, X., RSC Adv. 2014, 4, 22561- 22566. 146. Sun, Z.; Duan, R.; Zhang, H.; Pang, X.; Wang, X.; Chen, X., J. Renew. Mat. 2015, 3, 82-90. 147. Li, W.; Zhang, Z.; Yao, Y.; Zhang, Y.; Shen, Q., Organometallics 2012, 31, 3499-3511. 148. Li, M.; Hong, J.; Chen, Z.; Zhou, X.; Zhang, L., Dalton Trans. 2013, 42, 8288-8297. 149. Chi, S.-H.; Luo, Y.-J.; Chen, J., Synth. React. Inorg. M. 2014, 44, 611-615. 150. Sun, S.; Nie, K.; Tan, Y.; Zhao, B.; Zhang, Y.; Shen, Q.; Yao, Y., Dalton Trans. 2013, 42, 2870- 2878. 151. Garcés, A.; Sánchez-Barba, L. F.; Fernández-Baeza, J.; Otero, A.; Honrado, M.; Lara-Sánchez, A.; Rodríguez, A. M., Inorg. Chem. 2013, 52, 12691-12701. 152. Williams, C. K.; Brooks, N. R.; Hillmyer, M. A.; Tolman, W. B., Chem. Commun. 2002, 2132- 2133. 153. Breyfogle, L. E.; Williams, C. K.; Young Jr, V. G.; Hillmyer, M. A.; Tolman, W. B., Dalton Trans. 2006, 928-936. 154. Knight, P. D.; White, A. J. P.; Williams, C. K., Inorg. Chem. 2008, 47, 11711-11719. 155. Lian, B.; Thomas, C. M.; Casagrande, O. L.; Lehmann, C. W.; Roisnel, T.; Carpentier, J.-F., Inorg. Chem. 2007, 46, 328-340. 156. Sun, Y.; Cui, Y.; Xiong, J.; Dai, Z.; Tang, N.; Wu, J., Dalton Trans. 2015, 44, 16383-16391. 157. Kong, W.-L.; Wang, Z.-X., Dalton Trans. 2014, 43, 9126-9135. 158. Zelikoff, A. L.; Kopilov, J.; Goldberg, I.; Coates, G. W.; Kol, M., Chem. Commun. 2009, 6804- 6806.

159. Saha, T. K.; Ramkumar, V.; Chakraborty, D., Inorg. Chem. 2011, 50, 2720-2722. 160. Tsai, C.-Y.; Du, H.-C.; Chang, J.-C.; Huang, B.-H.; Ko, B.-T.; Lin, C.-C., RSC Adv. 2014, 4, 14527-14537. 161. Tseng, H.-C.; Chen, H.-Y.; Huang, Y.-T.; Lu, W.-Y.; Chang, Y.-L.; Chiang, M. Y.; Lai, Y.-C.; Chen, H.-Y., Inorg. Chem. 2016. 162. He, J.-X.; Duan, Y.-L.; Kou, X.; Zhang, Y.-Z.; Wang, W.; Yang, Y.; Huang, Y., Inorg. Chem. Commun. 2015, 61, 144-148. 163. Maudoux, N.; Roisnel, T.; Carpentier, J.-F.; Sarazin, Y., Organometallics 2014, 33, 5740-5748. 164. Normand, M.; Kirillov, E.; Roisnel, T.; Carpentier, J.-F., Organometallics 2012, 31, 1448-1457. 165. Char, J.; Brulé, E.; Gros, P. C.; Rager, M.-N.; Guérineau, V.; Thomas, C. M., J. Organomet. Chem. 2015, 796, 47-52. 166. Jin, W.-J.; Ding, L.-Q.; Chu, Z.; Chen, L.-L.; Lü, X.-Q.; Zheng, X.-Y.; Song, J.-R.; Fan, D.-D., J. Mol. Catal. A Chem. 2011, 337, 25-32. 167. Xiao, G.; Yan, B.; Ma, R.; Jin, W. J.; Lü, X. Q.; Ding, L. Q.; Zeng, C.; Chen, L. L.; Bao, F., Polym. Chem. 2011, 2, 659-664. 168. Ding, L.; Jin, W.; Chu, Z.; Chen, L.; Lü, X.; Yuan, G.; Song, J.; Fan, D.; Bao, F., Inorg. Chem. Commun. 2011, 14, 1274-1278. 169. Chen, H.-Y.; Liu, M.-Y.; Sutar, A. K.; Lin, C.-C., Inorg. Chem. 2009, 49, 665-674. 170. Hsieh,I.-P.; Huang, C.-H.; Lee, H. M.; Kuo, P.-C.; Huang, J.-H.; Lee, H.-I.; Cheng, J.-T.; Lee, G.-H., Inorganica Chim. Acta 2006, 359, 497-504. 171. Buffet, J.-C.; Okuda, J.; Arnold, P. L., Inorg. Chem. 2009, 49, 419-426. 172. Kalita, L.; Walawalkar, M. G.; Murugavel, R., Inorganica Chim. Acta 2011, 377, 105-110. 173. Normand, M.; Kirillov, E.; Roisnel, T.; Carpentier, J.-F. o., Organometallics 2011, 31, 1448- 1457. 174. Dagorne, S.; Normand, M.; Kirillov, E.; Carpentier, J.-F., Coord. Chem. Rev. 2013, 257, 1869- 1886. 175. Normand, M.; Dorcet, V.; Kirillov, E.; Carpentier, J.-F., Organometallics 2013, 32, 1694-1709. 176. Pal, M. K.; Kushwah, N.; Manna, D.; Wadawale, A.; Sudarsan, V.; Ghanty, T. K.; Jain, V. K., Organometallics 2013, 32, 104-111. 177. Quan, S. M.; Diaconescu, P. L., Chem. Commun. 2015, 51, 9643-9646. 178. Lindoy, L. F.; Meehan, G. V.; Svenstrup, N., Synthesis 1998, 1998, 1029-1032. 179. Breyfogle, L. E.; Williams, C. K.; Young, V. G.; Hillmyer, M. A.; Tolman, W. B., Dalton Trans. 2006, 928-936. 180. Douglas, A. F.; Patrick, B. O.; Mehrkhodavandi, P., Angew. Chem. Int. Ed. Engl. 2008, 47, 2290-2293. 181. Yu, I.; Acosta-Ramirez, A.; Mehrkhodavandi, P., J. Am. Chem. Soc. 2012, 134, 12758-12773. 182. Aluthge, D. C.; Ahn, J. M.; Mehrkhodavandi, P., Chem. Sci. 2015, 6, 5284-5292. 183. Xu, C.; Yu, I.; Mehrkhodavandi, P., Chem. Commun. 2012, 48, 6806-6808. 184. Ebrahimi, T.; Hatzikiriakos, S. G.; Mehrkhodavandi, P., Macromolecules 2015, 48, 6672-6681. 185. Yu, I.; Ebrahimi, T.; Hatzikiriakos, S. G.; Mehrkhodavandi, P., Dalton Trans. 2015, 44, 14248- 14254. 186. Cheng, M.; Attygalle, A. B.; Lobkovsky, E. B.; Coates, G. W., J. Am. Chem. Soc. 1999, 121, 11583-11584. 187. Ovitt, T. M.; Coates, G. W., J. Am. Chem. Soc. 1999, 121, 4072-4073. 188. Dagorne, S.; Normand, M.; Kirillov, E.; Carpentier, J. F., Coord. Chem. Rev. 2013, 257, 1869- 1886.

189. Normand, M.; Kirillov, E.; Roisnel, T.; Carpentier, J. F., Organometallics 2012, 31, 1448-1457. 190. Normand, M.; Roisnel, T.; Carpentier, J. F.; Kirillov, E., Chem. Commun. 2013, 49, 11692- 11694. 191. Hsieh, I. P.; Huang, C. H.; Lee, H. M.; Kuo, P. C.; Huang, J. H.; Lee, H. I.; Cheng, J. T.; Lee, G. H., Inorganica Chim. Acta 2006, 359, 497-504. 192. Broderick, E. M.; Guo, N.; Vogel, C. S.; Xu, C.; Sutter, J.; Miller, J. T.; Meyer, K.; Mehrkhodavandi, P.; Diaconescu, P. L., J. Am. Chem. Soc. 2011, 133, 9278-9281. 193. Hu, M. G.; Wang, M.; Zhang, P. L.; Wang, L.; Zhu, F. J.; Sun, L. C., Inorg. Chem. Commun. 2010, 13, 968-971. 194. Pal, M. K.; Kushwah, N. P.; Manna, D.; Wadawale, A. P.; Sudarsan, V.; Ghanty, T. K.; Jain, V. K., Organometallics 2013, 32, 104-111. 195. Peckermann, I.; Kapelski, A.; Spaniol, T. P.; Okuda, J., Inorg. Chem. 2009, 48, 5526-5534. 196. Fang, J.; Yu, I.; Mehrkhodavandi, P.; Maron, L., Organometallics 2013, 32, 6950-6956. 197. Labourdette, G.; Lee, D. J.; Patrick, B. O.; Ezhova, M. B.; Mehrkhodavandi, P., Organometallics 2009, 28, 1309-1319. 198. Larrow, J. F.; Jacobsen, E. N.; Gao, Y.; Hong, Y. P.; Nie, X. Y.; Zepp, C. M., J. Org. Chem. 1994, 59, 1939-1942. 199. Lee, D. W.; Ha, H. J.; Lee, W. K., Synth. Commun. 2007, 37, 737-742. 200. Luo, S. Z.; Xu, H.; Li, J. Y.; Zhang, L.; Cheng, J. P., J. Am. Chem. Soc. 2007, 129, 3074-3075. 201. Amarasinghe, N. R.; Turner, P.; Todd, M. H., Adv. Synth. Catal. 2012, 354, 2954-2958. 202. Sloan, T. E. Top. Stereochem. 2009, 12, 1. 203. Saini, P. K.; Romain, C.; Williams, C. K. Chem. Commun. 2014, 50(32), 4164-4167.

Appendices

Appendix A. Methodology for the calculation of tacticity

Table A- 1: Bernoullian equations of probability obtained for the different tetrad present in PLA.70

rac-LA meso-LA

Tetrad Bernouillan Distribution of Bernouillan Distribution of Probability Atactic PLA Probability Atactic PLA 푃 푃 mmm 푃 2 + 푟 푚 0.375 0 0 푚 2 푃 푃 mmr 푟 푚 0.125 0 0 2 푃 푃 rmm 푟 푚 0.125 0 0 2

푃 2 푃 2 + 푃 푃 rmr 푟 0.125 푚 푟 푚 0.250 2 2

푃 2 + 푃 푃 푃 2 mrm 푟 푟 푚 0.250 푚 0.125 2 2 푃 푃 rrm 0 0 푟 푚 0.125 2 푃 푃 mrr 0 0 푟 푚 0.125 2 푃 푃 rrr 0 0 푃 2 + 푟 푚 0.375 푟 2 Nomenclature in the literature:

-(R,R)- linkage = meso (m) or iso (i) -(R,S)- linkage = racemic (r) or syndio (s) Therefore:  For rac-LA: o Pm = Pi and relates to isotacticity o Pr = Ps and relates to heterotacticity

 For meso-LA: o Pm = Pi and relates to heterotacticity o Pr = Ps and relates to syndiotacticity

Note that Pr relates to different tacticity depending on which monomer is used.

100

Determination of tacticity from 1H{1H} NMR:

x z x y z

 For rac-LA: o x  [rmr], y  [mmr] corresponding to the areas of the two peaks in the 5.20-5.25 ppm range and z  [mmm] + [rmm] + [mrm] corresponding to the area of the overlapping peaks below 5.20 ppm 푥 푦 푧 o A = x + y + z, and therefore , and represent the normalized areas for each section 퐴 퐴 퐴 2 0.5 푥 푃푟 2푥 o = [rmr] = and thus Pr = ( ) 퐴 2 퐴 푦 푃푟푃푚 2푦 o = [mmr] = and thus Pm = 퐴 2 퐴푃푟

 For meso-LA: o x  [rmr] corresponding to the area of the peak above 5.20 ppm and z  [rrr] + [rrm] + [mrm] + [mrr] corresponding to the area of the overlapping peaks below 5.20 ppm o Note that no y is required as no peak should arise from this region of the spectrum 푥 푧 o A = x + z, and therefore and represent the normalized areas for each section 퐴 퐴 푥 푃 2 + 푃 푃 o = [rmr] = 푚 푟 푚 퐴 2 푧 푃 푃 푃 푃 푃 푃 푃 2 o = [rrr] + [rrm] + [mrr] + [mrm] = (푃 2 + 푟 푚) + ( 푟 푚) + ( 푟 푚) + ( 푚 ) 퐴 푟 2 2 2 2 o The system obtained has two equations and two unknowns and thus can be solved (using excel) to get Pm and Pr

Note that these equations assume that no transesterification reaction is occurring. Note that for rac-LA, both Pm and Pr value can be calculated independently, thus the sum of the two will not necessarily be equal to 1 (but 0.95-1.05) depending on how Bernoullian is the system. For meso- LA, on the other side, the calculation of Pm is made relative to Pr (and vice versa) and therefore the sum of the two must be equal to exactly 1.

101

Appendix B. Experimental section – Additional information

B-1. Synthesis of 4-tert-butyl-2,6-diformylphenol

This synthesis was adapted from the literature.154, 178 A Schlenk flask was charged with 4-tert- butylphenol (15.089 g, 100.4 mmol) and hexamethylenetetramine (28.190 g, 201.1 mmol). Anhydrous trifluoroacetic acid (200 mL) was transferred to the flask under N2 using a plastic cannula. The solution was refluxed under N2 at 120 °C for 20 h. The mixture was cooled, then poured into 4M HCl (400 mL) and stirred for approximately 10 minutes. The solution was extracted with CH2Cl2 (2 x 400 mL) and the organics were collected and washed with 4M HCl (2 x 400 mL), water (400 mL) and brine (400 mL).

The organics were collected and dried over MgSO4, filtered and pumped to dryness in vacuo yielding the crude compound as a bright yellow coloured solid. The solid was redissolved in CH2Cl2 and vacuum filtered through a small silica plug, washing with CH2Cl2 until no more colour was removed from the silica. The CH2Cl2 filtrates were pumped to dryness in vacuo yielding a bright yellow solid (16.15 g,

78%). The solid was dissolved in a minimum of hot cyclohexane and filtered. Crystals formed as the solution cooled to room temperature. The solution was vacuum filtered yielding the first portion of the product as bright yellow crystals. The filtrate was concentrated in vacuo causing precipitation of a second portion of the product, which was collected via vacuum filtration. This process was repeated again

102 yielding a third portion of the product. The products were combined and dried under vacuum yielding

1 the purified compound as bright yellow crystals (10.91 g, 53%). H NMR (300 MHz, CDCl3, 25 °C): δ

11.49 (1H, s, OH), 10.26 (2H, s, CHO), 7.99 (2H, s, Ar-H), 1.36 (9H, s, C(CH3)3).

B-2. Synthesis of 4-methyl-2,6-diformylphenol

This synthesis was adapted from the literature154, 178. 4-methyl-2,6-diformylphenol was prepared and purified in an analogous manner to 4-tert-butyl-2,6-diformylphenol from p-cresol (5.000 g, 46.2 mmol)

1 and recrystallized twice from cyclohexane (yield: 3.120 g, 41%). H NMR (300 MHz, CDCl3, 25 °C): δ

11.45 (1H, s, OH), 10.22 (2H, s, CHO), 7.77 (2H, s, Ar-H), 2.39 (3H, s, CH3).

103

B-3. Additional information concerning the synthesis of (±)-4-(tert-butyl)-2,6-bis(((2-

(dimethylamino)cyclohexyl)imino)methyl)phenol (rac-L4)

The product was confirmed by 1H NMR spectroscopy to be a mixture of the bis-imine compound

(bis-L4) and the mono-imine compound (mono-L4) when 2 or 3 equivalents of diamine starting material was used (with a ratio of respectively 1:1 and 1:10, bis:mono). Resolution of every peak for the different compounds in the 1H NMR spectrum of the mixture was possible by independent synthesis of the mono-

L1 using 1 equivalent of (±)-N,N-dimethyl-trans-1,2-diaminocyclohexane, a selection of diagnostic

1 peaks follows. H NMR (300 MHz, CDCl3, 25 °C): δ for mono-L4: 10.54 (1H, s, CHO), 8.24 (1H, s,

4 4 N=CH), 7.92 (1H, d, JHH = 3 Hz, Ar-H), 7.44 (1H, d, JHH = 3 Hz, Ar-H), 2.28 (6H, s, N(CH3)2), 1.31

(9H, s, C(CH3)3); for (±)-NMe2N 2.23 (6H, s, N(CH3)2).

B-4. Additional information concerning the synthesis of synthesis of (±)-4-(methyl)-2,6- bis(((2-(dimethylamino)cyclohexyl)imino)methyl)phenol (rac-L5)

104

As observed for the synthesis of rac-L4, the presence of the mono-substituted imine can be found when lower equivalency of starting (±)-N,N-dimethyl-trans-1,2-diaminocyclohexane are used. A

1 selection of diagnostic peaks follows. H NMR (300 MHz, CDCl3, 25 °C): δ for mono-L5: 10.54 (1H, s,

4 4 CHO), 8.24 (1H, s, N=CH), 7.92 (1H, d, JHH = 3 Hz, Ar-H), 7.44 (1H, d, JHH = 3 Hz, Ar-H), 2.28 (6H, s, N(CH3)2), 1.31 (9H, s, C(CH3)3); for (±)-NMe2N 2.23 (6H, s, N(CH3)2).

B-5. Synthesis of enantiopure proligands L5 and L7

Proligands RR/RR-L5, SS/SS-L5, RR/RR-L7 and SS/SS-L7 were synthesized in analogous manner as it was previously described for the racemic counter parts. Imines analogues were made in situ.

B-6. Synthesis of 4-(tert-butyl)-2,6-bis(((2-(dimethylamino)ethyl)imino)methyl)phenol (L1)

105

This synthesis was adapted from the literature.154 To a solution of 4-tert-butyl-2,6-diformylphenol

(3.925 g, 19.0 mmol) in methanol (50 mL) was added N,N-dimethylethylenediamine (4.5 mL, 41 mmol).

The solution was stirred at room temperature for 19 h, then pumped to dryness yielding the crude compound as a thick, orange oil (6.593 g, 100%). The compound was used without further purification.

1 3 H NMR (600 MHz, CDCl3, 25 °C): δ 8.61 (1H, br s, N=CH), 7.69 (1H, br s, Ar-H), 3.76 (4H, t, JHH =

3 6 Hz, NCH2CH2N), 2.66 (4H, t, JHH = 6 Hz, NCH2CH2N), 2.33 (12H, s, N(CH3)2), 1.32 (9H, s,

13 1 C(CH3)3). C{ H} NMR (150 MHz, CDCl3, 25 °C): δ 159.08, 140.69, 59.83, 45.57, 33.89, 31.16. Anal.

Calc. for C20H34N4O: C, 69.32; H, 9.89; N, 16.17. Found: C, 68.99; H, 9.70; N, 16.13.

B-7. Synthesis of 4-(tert-butyl)-2,6-bis(((2-(dimethylamino)ethyl)amino)methyl)phenol (L2)

106

154 This synthesis was adapted from the literature . A solution of L2 (0.588 g, 1.28 mmol) in methanol

(20 mL) was cooled in an ice bath to 0 °C then NaBH4 (0.194 g, 5.13 mmol) was added to the solution in small portions. Note that alternatively, the imine L5 does not need to be isolated (as described above) and the NaBH4 may be added in situ after the formation of the imine is complete. The solution was stirred at room temperature for 16 h, then the solvent was removed in vacuo. The resulting residue was redissolved in CH2Cl2 (~5 mL) and the solution was vacuum filtered through a small plug of alumina.

The alumina plug was washed with CH2Cl2 (5x10 mL) then methanol (~100 mL) until no more colored was removed into the filtrate. The solution was pumped to dryness and the residue was redissolved in

CH2Cl2 and filtered through celite. The filtrate was pumped to dryness yielding a pale yellow oil (0.39

1 g, 87%). The compound was used without further purification. H NMR (600 MHz, CDCl3, 25 °C): δ

3 3 7.01 (2H, s, Ar-H), 3.89 (4H, s, Ar-CH2-N), 2.73 (4H, t, JHH = 6 Hz, NCH2CH2N), 2.44 (4H, t, JHH =

13 1 6 Hz, NCH2CH2N), 2.21 (12H, s, N(CH3)2), 1.27 (9H, s, C(CH3)3). C{ H} NMR (150 MHz, CDCl3,

25 °C): δ 154.15, 140.95, 124.86, 123.68, 58.66, 51.29, 46.36, 45.47, 33.86, 31.57. Anal. Calc. for

C20H38N4O: C, 68.53; H, 10.93; N, 15.98. Found: C, 67.78; H, 10.68; N, 16.27.

107

B-8. Synthesis of 4-(tert-butyl)-2,6-bis(((2-(dimethylamino)ethyl)(methyl)amino) methyl)phenol (L3)

This synthesis was adapted from the literature152, 154. To a solution of 4-tert-butylphenol (5.251 g,

34.96 mmol) and N,N,N’-trimethylethylenediamine (10.0 mL, 76.9 mmol) in ethanol (100 mL) was added paraformaldehyde (2.205 g, 73.43 mmol). The solution was refluxed at 80 °C for 3 days, then it was cooled and the solvent was removed in vacuo to yield the crude product as a thick, pale orange coloured oil (13.87 g). The crude oil was redissolved in CH2Cl2 (~10 mL) and vacuum filtered through a small alumina plug. The alumina plug was washed with CH2Cl2 (200 mL), then the filtrate was collected and pumped to dryness in vacuo yielding a yellow oil. The oil was dissolved in hexane, and the supernatant solution was decanted from the insoluble portion of the oil, which was discarded. The hexane supernatant solution was pumped to dryness in vacuo yielding a pale yellow oil (9.52 g, 72%).

1 The compound was used without further purification. H NMR (600 MHz, CDCl3, 25 °C): δ 7.02 (2H,

3 3 s, Ar-H), 3.59 (4H, s, Ar-CH2-N), 2.55 (4H, t, JHH = 6 Hz, NCH2CH2N), 2.48 (4H, t, JHH = 6 Hz,

13 1 NCH2CH2N), 2.27 (6H, s, N(CH3)), 2.21 (12H, s, N(CH3)2), 1.26 (9H, s, C(CH3)3). C{ H} NMR (150

MHz, CDCl3, 25 °C): δ 153.75, 140.68, 125.48, 122.57, 58.66, 57.11, 54.83, 45.66, 42.31, 31.55, 33.80,

31.55. Anal. Calc. for C22H42N4O: C, 69.79; H, 11.18; N, 14.80. Found: C, 69.80; H, 11.21; N, 15.03.

108

B-9. Synthesis of rac-6

To a solution of proligand rac-L7 (0.499 g, 1.19 mmol) in toluene (5 mL) was transferred potassium tert-butoxide (0.132 g, 1.17 mmol, 0.98 equiv) using toluene (5 mL). The solution was stirred at room temperature for 16 h, after which it was pumped to dryness yielding a yellow solid residue (0.471 g,

87%). The potassium salt (0.100 g, 0.22 mmol) was redissolved in THF (5 mL) and zinc dichloride

(0.097 g, 0.44 mmol, 2 equiv) was transferred to this solution using THF (5 mL). The solution was stirred at room temperature for 16 h, after which the solvent was removed in vacuo. The residual solid was redissolved in CH2Cl2 (5 mL), and filtered through celite. Solvent was removed in vacuo, and residual solid washed with ether 3 times. The resulting solid was proved by 1H NMR spectroscopy to be a mixture of product, probably diastereoisomers. Isolation of one of these compounds was not possible.

RR/RR-[(L7)Zn2Cl2](μ-Cl) (6)

rac-6

1 Figure B-1: H NMR (400MHz, RT, CDCl3) of complex RR/RR-6 compared to rac-6 obtained from proligands respectively RR/RR-L7 and rac-L7.

109

B-10. Attempted synthesis of RR/RR-[(L5)Zn2Cl2](μ-OEt)

To a solution of complex 7 (0.050 g, 0.08 mmol) in toluene (5 mL) was transferred sodium ethoxide

(0.010 g, 0.15 mmol, 2 equiv) using toluene (5 mL). The solution was stirred at room temperature for

16 h, after which it was pumped to dryness yielding a orange solid residue. The residual solid was redissolved in CH2Cl2 (5 mL), and filtered through celite. Solvent was removed in vacuo and the resulting solid was washed with hexanes 3 times. The resulting solid was proved by 1H NMR to be a mixture of product, probably diastereoisomers in addition to starting complex 7. Isolation of one of these compounds was not possible. Crystals suitable for X-ray analysis were grown from a saturated solution in THF and ether at room temperature; however single crystal X-ray crystallography of these crystals revealed them to be the related hydroxide-bridged complex 8.

RR/RR-[(L5)Zn2Cl2](μ-Cl) (7)

RR/RR-[(L5)Zn2Cl2](μ-Cl) (7) + 2 NaOEt

RR/RR-[(L5)Zn2Cl2](μ-OH) (8)

1 Figure B-2: H NMR (400MHz, RT, CDCl3) of respectively complex 7, mixture of products from the attempted alkoxylation and complex 8.

110

B-11. MALDI-TOF Mass Spectroscopy on PLA Oligomers.

In a 20 mL scintillation vial 2 (4.4 mg, 0.0055 mmol) was dissolved in 1 mL of CH2CL2, and rac-LA

(0.016 g, 0.011 mmol, 20 equiv) was added using 1 mL of CH2CL2. The mixture was allowed to stir at room temperature for 16 h. Thereafter the mixture was quenched with 1 mL of wet hexanes. The solvent was evaporated under vacuum, and the polymer was further dried for 16 h prior analysis. Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometric analysis was performed on a Bruker Autoflex MALDI-TOF equipped with a nitrogen laser (337 nm). The accelerating potential of the Bruker instrument was 19.5 kV. The polymer samples were dissolved in THF (ca. 1 g/mL). The concentration of a cationization agent, sodium trifluoroacetate, in THF was 1 mM. The matrix used was trans-2-[3-(4-tert-butylphenyl)2-methyl-2-propenylidene]malononitrile (DCTB) at a concentration of 20 mg/mL. A sample solution was prepared by mixing polymer, matrix, and salt in a volume ratio of 5:5:1, respectively. The mixed solution was spotted on a stainless steel target and then left to dry at room temperature. The spectra were collected in a linear mode.

111

Appendix C. Characterization of compounds in solution; 1H and 13C{1H} NMR spectra

1 Figure C-1: H NMR spectrum (400 MHz, CDCl3, 25 °C) of rac-L4.

112

1 Figure C-2: H NMR spectrum (400 MHz, CDCl3, 25 °C) of RR/RR-L5.

13 1 Figure C-3: C{ H} NMR spectrum (100 MHz, CDCl3, 25 °C) of (RR/RR)-L2.

113

1 Figure C-4: H NMR spectrum (400 MHz, CDCl3, 25 °C) of rac-L6.

13 1 Figure C-5: C{ H} NMR spectrum (150 MHz, CDCl3, 25 °C) of rac-L3.

114

1 Figure C-6: H NMR spectrum (400 MHz, CDCl3, 25 °C) of RR/RR-L7.

13 1 Figure C-7: C{ H} NMR spectrum (100 MHz, CDCl3, 25 °C) of (RR/RR)-L4.

115

1 Figure C-8: H NMR spectra (300 MHz, CDCl3, 25 °C) of rac-L6 and crossover experiments associated of the SS/SS-L6 + RR/RR-L6 and rac-L6 + RR/RR-L6.

116

1 Figure C-9: H NMR spectrum (400 MHz, CDCl3, 25 °C) of RR/RR-[(L7)In2Cl4](μ-OEt) (4).

13 1 Figure C-10: C{ H} NMR spectrum (100 MHz, CDCl3, 25 °C) of RR/RR-[(L7)In2Cl4](μ-OEt) (4).

117

1 Figure C-11: H NMR spectrum (400 MHz, CDCl3, 25 °C) of RR/RR-[(L7)Zn2Cl2](μ-OEt) (5).

13 1 Figure C-12: C{ H} NMR spectrum (100 MHz, CDCl3, 25 °C) of RR/RR-[(L7)Zn2Cl2](μ-OEt) (5).

118

1 Figure C-13: H NMR spectrum (400 MHz, CDCl3, 25 °C) of RR/RR-[(L7)Zn2Cl2](μ-Cl) (6).

13 1 Figure C-14: C{ H} NMR spectrum (150 MHz, CDCl3, 25 °C) of RR/RR-[(L7)Zn2Cl2](μ-Cl) (6).

119

1 Figure C-15: H NMR spectrum (600 MHz, CDCl3, 25 °C) of RR/RR-[(L5)Zn2Cl2](μ-Cl) (7).

13 1 Figure C-16: C{ H} NMR spectrum (150 MHz, CDCl3, 25 °C) of RR/RR-[(L5)Zn2Cl2](μ-Cl) (7).

120

Ar-C(CH3)3

NCH3

N(CH3)2 Ar-H 2 3 N-CH2CH2-N OCH CH N-CH2-Ar

N-CH2CH2-N OCH2CH3

7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 Chemical Shift (ppm) 1 Figure C-17: H NMR spectrum (600 MHz, CDCl3, 25 °C) of [(L3)In2Cl4](μ-OEt) (2).

13 1 Figure C-18: C{ H} NMR spectrum (150 MHz, CDCl3, 25 °C) of [(L3)In2Cl4](μ-OEt) (2).

121

Appendix D. Polymer characterization

1 1 Figure D-1: H{ H} NMR of heterotactic PLA (600 MHz, CDCl3, 25 °C) obtained from rac-LA using complex 2 in CH2Cl2 at room temperature (Table 2-1, Entry 2).

1 1 Figure D-2: H{ H} NMR of atactic PLA (600 MHz, CDCl3, 25 °C) obtained from meso-LA using complex 4 in CH2Cl2 at room temperature (Table 2-1, Entry 11).

122

Figure D-3: MALDI-TOF mass spectra of PLA oligomers obtained using complex 2 in CH2Cl2 at room temperature with 20 equivalents of lactide.

Figure D-4: MALDI-TOF mass spectra of PLA oligomers obtained using complex 2 at 100 °C in toluene with 40 equivalents of lactide.

123

Appendix E. Solid state structure and crystallographic parameters

Figure E-1: Solid state structures of the two diastereoisomers present in complex 7, respectively 7a (left) and 7b (right). Unit cell obtained by single crystal X-ray diffraction. Thermal ellipsoids are set at 50% probability and H atoms and solvent removed for clarity. The two distinct diastereoisomers differ from the orientation of the chloride around the zinc centers.

124

Table E-1: Selected bond lengths and angles for complexes 7a, 7b, and 8.

Complex 7a Complex 7b Complex 8

Zn1-Cl1 2.259(3) 2.277(3) 2.2657(5)

Zn1-Cl3/O2 2.355(2) 2.371(3) 2.0001(15)

Zn1-N4 2.220(9) 2.206(8) 2.1631(16)

Zn1-N3 2.090(8) 2.064(7) 2.1393(16)

Bond lengths Zn1-O1 2.130(6) 2.103(6) 2.0531(14)

(Å) Zn2-Cl2 2.274(3) 2.269(3) 2.2830(5)

Zn2-Cl3/O2 2.469(2) 2.476(3) 1.9658(14)

Zn2-N1 2.168(8) 2.180(8) 2.1586(17)

Zn2-O1 2.015(6) 2.044(6) 2.0766(14)

Zn2-N2 2.104(8) 2.067(8) 2.1055(17)

Cl3/O2-Zn1-Cl1 105.89(11) 110.47(10) 109.59(5)

Cl3-O2-Zn1-N4 98.2(2) 99.0(2) 93.60(6)

Cl3/O2-Zn1-N3 130.8(2) 118.2(2) 144.87(6)

Bond angles (°) Cl3/O2-Zn1-O1 80.83(16) 81.64(19) 79.44(6)

N4-Zn1-Cl1 102.7(3) 97.3(2) 109.08(4)

N3-Zn1-Cl1 122.5(2) 131.0(2) 105.37(4)

N3-Zn1-N4 80.2(3) 81.0(3) 78.03(6)

125

O1-Zn1-Cl1 97.8(2) 97.8(2) 113.14(4)

O1-Zn1-N4 158.9(3) 163.6(3) 137.10(6)

O1-Zn1-N3 84.6(3) 84.3(3) 83.90(6)

Cl3/O2-Zn2-Cl2 103.20(10) 100.92(10) 113.60(5)

Cl3/O2-Zn2-N1 94.8(2) 95.9(2) 99.00(6)

Cl3/O2-Zn2-O1 80.31(18) 80.27(18) 79.66(6)

Cl3/O2-Zn2-N2 150.3(2) 157.4(2) 145.94(6)

N1-Zn2-Cl2 104.6(2) 111.4(3) 102.17(4)

O1-Zn2-Cl2 115.2(2) 118.4(2) 106.31(4)

O1-Zn2-N1 140.0(3) 129.9(3) 149.44(6)

O1-Zn2-N2 86.4(3) 84.3(3) 83.90(6)

N2-Zn2-Cl2 106.4(2) 100.9(2) 99.55(4)

N2-Zn2-N1 78.6(3) 81.5(3) 80.64(6)

Zn2-Cl3/O2-Zn1 88.29(7) 86.51(8) 103.23(7)

Zn2-O1-Zn1 108.3(2) 106.5(2) 97.66(6)

126

Table E-2: Selected crystallographic parameters for complexes 4 and 5.

Complex 4 Complex 5

Empirical formula C25H39N4OCl3Zn2 C29H47Cl2N4O3Zn2

Fw 648.69 658.30

T (K) 296.15 296.15

a (Å) 11.1530(11) 21.841(2)

b (Å) 11.3575(11) 12.0898(12)

c (Å) 45.662(4) 12.3600(12)

α (°) 90 90

β (°) 90 99.878(2)

γ (°) 90 90

volume (Å3) 5784.0(10) 3215.3(6)

Z 8 4

crystal system orthorhombic Monoclinic

space group P212121 C2

3 dcalc (g/cm ) 1.490 1.360

μ (MoKα) (mm-1) 1.961 1.687

2θ max (deg) 50.81 66.622

total no. of reflections 10099 108171

no. of indep reflections (Rint) 10099 12345 (0.0372)

2 residuals (refined on F , all data): R1; wR2 0.0853; 0.0931 0.0261; 0.0565

GOF 1.047 1.034

2 a b residuals (refined on F ): R1 ; wR2 0.0549; 0.0867 0.0225; 0.0550 a b 2 2 2 2 2 1/2 R1 = Σ ||Fo| - |Fc|| / Σ |Fo|; wR2 = [Σ(w(Fo - Fc ) )/Σw(Fo ) ] .

127

Table E-3: Selected bond lengths and angles for complexes 2, 3, and 4.

Complex 4 Complex 2 Complex 3

In2-N1 2.385(6) 2.353(5) 2.3689(17)

In1-N4 2.378(6) 2.382(4) 2.3904(18)

In2-N2 2.219(6) 2.284(6) 2.2847(17)

In1-N3 2.235(6) 2.264(5) 2.2933(19)

In2-Cl4 2.4941(18) 2.4601(15) 2.4761(6)

Bond lengths In1-Cl2 2.4793(16) 2.4683(15) 2.5055(6)

(Å) In2-Cl3 2.4627(18) 2.4547(18) 2.4599(6)

In1-Cl1 2.4691(16) 2.4431(17) 2.4524(6)

In2-O1 2.246(5) 2.219(4) 2.2569(15)

In1-O1 2.248(4) 2.231(3) 2.2441(14)

In2-O2 2.102(5) 2.110(4) 2.0947(16)

In1-O2 2.107(5) 2.097(4) 2.0938(17)

In1-O1-In2 101.42(18) 102.01(13) 101.08(5)

In1-O2-In2 111.4(2) 110.60(16) 112.13(7) Bond angles (°) O1-In2-O2 73.64(18) 73.69(13) 73.17(6)

O1-In1-O2 73.50(18) 73.69(14) 73.46(6)

128

O1-In2-N1 87.21(19) 87.14(16) 86.25(6)

O1-In1-N4 86.80(19) 85.27(13) 85.97(6)

O2-In2-N1 105.0(2) 103.5(2) 101.06(6)

O2-In1-N4 105.2(2) 104.71(15) 105.74(7)

O1-In2-N2 85.95(19) 86.44(16) 86.53(6)

O1-In1-N3 86.13(18) 86.80(14) 86.32(6)

O2-In2-N2 159.4(2) 159.86(18) 159.70(6)

O2-In1-N3 159.4(2) 160.00(15) 159.19(6)

N1-In2-N2 76.1(2) 78.3(2) 77.44(6)

N3-In1-N4 76.2(2) 77.40(16) 77.04(7)

N1-In2-Cl4 163.68(16) 166.54(19) 166.75(4)

N4-In1-Cl2 163.51(16) 166.38(12) 165.29(5)

N2-In2-Cl3 103.07(16) 100.38(16) 99.47(5)

N3-In1-Cl1 102.24(15) 100.28(11) 98.77(5)

O1-In2-Cl3 168.55(12) 170.12(10) 170.91(4)

O1-In1-Cl1 168.43(12) 169.48(9) 170.42(4)

O2-In2-Cl4 91.27(16) 89.94(11) 92.19(5)

O2-In1-Cl2 91.27(16) 88.86(11) 88.93(5)

129

Table E-4: Selected crystallographic parameters for complexes 1, 7, and 8.

Complex 4 Complex 2 Complex 3

Empirical formula C27H48Cl4In2N4 C24H44Cl5In2N5 C24H45Cl4In2N5O2

O2 O2

Fw 832.13 1856.18 807.09

T (K) 296.15 296.15 296.15

a (Å) 9.6695(10) 11.307(5) 15.310(2)

b (Å) 13.0329(13) 18.369(5) 11.2460(17)

c (Å) 18.2092(19) 22.010(5) 19.389(3)

α (°) 107.933(2) 84.491(5) 90

β (°) 92.617(2) 83.099(5) 104.472(4)

γ (°) 97.399(2) 83.917(5) 90

volume (Å3) 2156.2(4) 4497(3) 3232.4(8)

Z 4 2 4

crystal system triclinic triclinic monoclinic

space group P-1 P-1 P 21/c

3 dcalc (g/cm ) 2.563 1.371 1.658

μ (MoKα) (cm-1) 2.682 12.94 17.87

2θmax (deg) 55.232 58.106 72.24

total no. of reflections 79607 93165 73072

no. of indep reflections (Rint) 9971 (0.0232) 23586 (0.0520) 15357 (0.0714)

2 residuals (refined on F , all data):R1; 0.0816; 0.3332 0.1027; 0.1441 0.0694; 0.0685

GOF 3.144 1.022 1.000

2 a b residuals (refined on F ): R1 ; wR2 0.0775; 0.3221 0.0571; 0.1194 0.0355; 0.0605 a b 2 2 2 2 2 1/2 R1 = Σ ||Fo| - |Fc|| / Σ |Fo|; wR2 = [Σ(w(Fo - Fc ) )/Σw(Fo ) ] .

130