Olefins from Carbonyls

Home , Julia olefination, McMurry reaction, Peterson olefination, Wittig reaction

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1962

Olefins from carbonyls

Development of new phosphorus-based cross- coupling reactions

NICOLAS DANIELE D'IMPERIO

ACTA UNIVERSITATIS UPSALIENSIS ISSN 1651-6214 ISBN 978-91-513-1001-5 UPPSALA urn:nbn:se:uu:diva-419124 2020 Dissertation presented at Uppsala University to be publicly examined in Siegbahnsalen, Ångströmlaboratoriet, Lägerhyddsvägen 1, Uppsala, Friday, 30 October 2020 at 10:15 for the degree of Doctor of Philosophy. The examination will be conducted in English. Faculty examiner: Professor Carlos Romero-Nieto (University of Heidelberg, Organisch-Chemisches Institut).

Abstract D'Imperio, N. D. 2020. Olefins from carbonyls. Development of new phosphorus-based cross- coupling reactions. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1962. 110 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-513-1001-5.

Olefins, compounds containing C=C double bonds, are omnipresent in nature and serve as useful starting materials for various chemical modifications. Olefins are of crucial importance in Medicinal Chemistry and are also present in essential objects like dyes, polymers and plastics. Thus, developing methodologies for synthesizing olefins is at the heart of Organic Chemistry. In this regard, many reactions have been developed over the years for the production of olefins from different starting materials. To date only one reaction, namely the McMurry coupling, is available for constructing olefins from two carbonyls. This reaction is frequently applied in an academic context, but suffers many drawbacks that limit its wider use. This thesis offers innovative phosphorus-based methodologies for coupling two carbonyls into olefins. All the methods presented herein are based on a one-pot sequence in which a first carbonyl is transformed into a phosphaalkene (P=C double bond containing molecule) which, upon activation, reacts with a second carbonyl with formation of desired alkenes. The first two chapters of this thesis give a general overview on literature protocols for the formation of olefins, along with a comparison between P=C and C=C double bonds containing molecules. The third chapter is dedicated to the presentation of a new potential phosphorus-based coupling reagent. The studies presented in this chapter set the basis for the development of a new cross-coupling reaction of aldehydes to olefins. In the fourth chapter a new method for the one-pot synthesis of disubstituted alkenes from aldehydes is presented. The reactivity of phosphaalkene intermediates proved to be crucial in determining the reaction scope of the process. In the fifth chapter, a closer look into the E-Z stereoselectivity of the protocol is described. The following two chapters deal with more reactive phosphaalkenes. Studies on their chemical properties showed to be fundamental for developing an unprecedent cross-coupling of ketones and aldehydes to trisubstituted alkenes. In summary, this thesis represents the development of new phosphorus-based cross-couplings of carbonyls to olefins. Protocols for the stereoselective synthesis of disubstituted alkenes from two aldehydes, and trisubstituted olefins from ketones and aldehydes are presented. These innovative methodologies offer precious alternatives to the McMurry coupling.

Keywords: olefins, carbonyl olefination, phosphaalkenes, stereoselective synthesis, Wittig, Horner-Wadsworth-Emmons, McMurry

Nicolas Daniele D'Imperio, Department of Chemistry - Ångström, Synthetic Molecular Chemistry, 523, Uppsala University, SE-751 20 Uppsala, Sweden.

ISSN 1651-6214 ISBN 978-91-513-1001-5 urn:nbn:se:uu:diva-419124 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-419124)

To my unique family

“The person who follows the crowd will usually go no further than the crowd. The person who walks alone is likely to find himself in places no one has ever seen before.”

Albert Einstein

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I D’Imperio, N., Arkhypchuk, A., (2019) Reactivity patterns of benzhydryl(mesityl)phosphane oxide – a potential intermediate in carbonyl-carbonyl coupling reactions? Phosphorus, Sulfur Sil- icon Relat. Elem., 194: 575-579. DOI: 10.1080/10426507.2018.1543305. II D’Imperio, N., Arkhypchuk, A., Ott, S. (2018) One-pot intermolecular reductive cross-coupling of deactivated aldehydes to un- symmetrically 1,2-disubstituted alkenes. Org. Lett., 20: 5086- 5089. DOI: 10.1021/acs.orglett.8b01754. III D’Imperio, N., Arkhypchuk, A., Ott, S. (2020) E,Z-Selectivity in the reductive cross-coupling of two benzaldehydes to stilbenes under substrate control. Org. Biomol. Chem., 18: 6171-6179. DOI: 10.1039/d0ob01139h. IV D’Imperio, N., Arkhypchuk, A., Mai, J., Ott, S. (2019) Triphen- yphosphaalkenes in chemical equilibria. Eur. J. Inorg. Chem., 1562-1566. DOI: 10.1002/ejic.201801322. V Arkhypchuk, A., D’Imperio, N., Ott, S., (2019) Triarylalkenes from the site-selective reductive cross-coupling of benzophe- nones and aldehydes. Chem. Commun., 55: 6030-6033. DOI: 10.1039/c9cc02972a.

Reprints were made with permission from the respective publishers.

Contribution report

Paper I Performed half of the synthetic work and characterizations of the synthesized products. Contributed to the discussion of the project and to the writing of the manuscript and the supporting information.

Paper II Performed most of the synthetic work and the characterizations. Contributed to the design of the project. Wrote the manuscript and the supporting information with feedback from S. Ott.

Paper III Performed most of the synthetic work and the characterizations. Wrote the manuscript and the supporting information with feedback from S. Ott.

Paper IV Performed a major part of the work. Major contribution to the in- terpretation of the results. Contributed to the design of the project. Wrote the manuscript and the supporting information with feedback from S. Ott.

Paper V Contributed to the design and discussion of the project and, partially, in the writing of the manuscript.

Contents

1. Introduction ...... 11 2. Background ...... 13 2.1 Formation of C=C bonds ...... 13 2.1.1 Carbonyl olefinations ...... 14 2.1.2 Wittig and related phosphorus olefination reactions ...... 15 2.1.3 Phosphorus free olefination reactions ...... 22 2.2 Phosphorus as the Carbon copy ...... 26 2.2.1 Comparison of P=C and C=C double bonds ...... 26 2.2.2 Stabilization of phosphaalkenes ...... 27 2.2.3 Synthetic pathways towards phosphaalkenes ...... 29 2.2.4 Reactivity and applications of phosphaalkenes ...... 31 3. Reactivity patterns of a potential new olefinating reagent (Paper I) ...... 34 3.1 Alkenes from two carbonyl compounds – the McMurry reaction .... 34 3.2 Novel methodology for aldehyde-aldehyde couplings via phosphaalkene intermediates ...... 35 3.3 Reducing the size of the phosphorus protecting group ...... 37 3.4 A new potential coupling reagent - preparation and reactivity ...... 38 3.4.1 Planning a new broader cross-coupling reaction ...... 38 3.4.2 Synthesis of benzhydryl(mesityl)phosphine oxide ...... 39 3.4.3 Reactivity study of benzhydryl(mesityl)phosphine oxide ...... 40 3.4.4 Conclusions and outlook ...... 42 4. Development of a new reductive cross-coupling of deactivated aldehydes (Paper II) ...... 43 4.1 Modification of the previous method ...... 43 4.2 The phospha-Peterson reaction ...... 45 4.3 Phosphaalkenes as electrophiles ...... 48 4.4 HWE-type olefinating reagents ...... 49 4.5 Optimization of the phospha-Peterson reaction ...... 50 4.6 Bridging low- and high-valent phosphorus chemistry ...... 53 4.7 Study of the olefination step ...... 55 4.8 Development of the new one-pot procedure ...... 58 4.9 Substrate scope of the new method ...... 60 4.10 Advantages and limitations ...... 63 4.11 Conclusions ...... 65

5. E-Z selectivity of the new cross-coupling procedure (Paper III) ...... 66 5.1 Introduction ...... 66 5.2 Synthesis of E- and Z- alkenes; results and discussion ...... 67 5.3 Conclusions ...... 70 6. Reducing the steric bulk on phosphaalkenes – studies on the stability of P-Ph-phosphaalkenes (Paper IV) ...... 72 6.1 Introduction ...... 72 6.2 Kinetic stabilization of phosphaalkenes ...... 73 6.3 Development of a new reliable synthesis of P-Ph-phosphaalkenes .. 74 6.4 Investigations on the chemical equilibria of P-Ph-phosphaalkenes .. 76 6.5 Conclusions ...... 80 7. Trisubstituted alkenes from the cross-coupling of ketones and aldehydes (Paper V) ...... 82 7.1 Reducing the size of the phosphorus protecting group for increased reactivity ...... 82 7.2 Developing a new cross-coupling reaction ...... 83 7.3 Substrate scope ...... 85 7.4 Advantages and limitations ...... 89 7.5 Conclusions ...... 90 8. Outlook ...... 91 9. Conclusions and summary ...... 93 Svensk sammanfattning ...... 96 Acknowledgements ...... 99 References ...... 103

Abbreviations

AcOH Acetic acid Ad Adamantyl Ar Aryl or aromatic Cp Cyclopentadienyl ligand C6D6 Deuterated benzene DABCO 1,4-diazabicyclo[2.2.2]octane DBU 1,8-diazabicyclo[5.4.0]undec-7-ene DME Dimethoxyethane DMF Dimethylformamide Dmp 2,6-dimesitylphenyl DMSO Dimethyl sulfoxide EDG Electron donating group Et Ethyl EtOH Ethanol EWG Electron withdrawing group Het Heteroaryl or heterocyclic HOMO Highest occupied molecular orbital HWE Horner-Wadsworth-Emmons IP Ionization potential KHMDS Potassium hexamethyldisilylamide LDA Lithium diisopropylamide Ln n number of ligands LUMO Lowest unoccupied molecular orbital Me Methyl MeOH Methanol Mes Mesityl (2,4,6-trimethylphenyl) Mes* Supermesityl (2,4,6-tri(tert-butyl)phenyl) NaHMDS Sodium hexamethyldisilylamide NMR Nuclear magnetic resonance OLEDS Organic light emitting diodes OPA Oxaphosphetane PG Protecting group Ph Phenyl SN2 Nucleophilic substitution of second type TBAOH Tetrabutylammonium hydroxide tBuOK Potassium tertbutoxide

THF Tetrahdyrofuran Tip Triisopropylphenyl TM Transition metal TMS Trimethylsilyl TMSCl Trimethylsilyl chloride TMSOTf Trimethylsilyl triflate TS Transition state

1. Introduction

Phosphaalkenes (P=C) and olefins (C=C) are of great importance in chemistry. Compounds containing P=C double bonds have been known for several decades.[1] In the early days of their discoveries, phosphaalkenes were mainly interesting as a new type of low-coordinate phosphorus compounds. With the development of their chemistry, more and more interest has grown for this new class of molecules. Several approaches have been developed for their synthesis, together with a detailed knowledge on their chemical reactivity.[2] Nowadays, phosphaalkenes are employed in various applications. Firstly, they show interesting reactivity and serve thus as starting point for further chemical transformations.[3] Secondly, phosphaalkenes had a considerable impact in polymer chemistry,[4] coordination chemistry[3c, 5] and organic electronics.[6] Thus, syntheses, applications and studies on the stability of these interesting systems are of crucial importance in modern organophosphorus chemistry. In this regard, this thesis describes new entries for the synthesis of phosphaalkenes. Compounds that were thought to be unstable until a few years ago, have been synthesized in this thesis by new and optimized procedures. Moreover, these species have been utilized as intermediates in new synthetic protocols for the unprecedented formation of olefins from two carbonyls. Olefins, which are compounds featuring C=C double bonds, are ubiquitous in Nature and provide useful starting materials for various chemical modifications.[7] Alkenes are found in natural products like lipids and vitamins and important pharmaceutical compounds (Figure 1.1).[8]

Figure 1.1 Selected examples of biologically important molecules containing C=C double bonds.

For example, the anticancer drug Tamoxifen,[9] the antioxidants Resveratrol[10] and Pterostilbene[11] and the drug Rosuvastatin,[12] have all a characteristic C=C double bond in common. Alkenes are not only important in Medicinal Chemistry, but are also present in pigments, dyes and plastics. Additionally, olefins are crucial components in modern optoelectronics devices.[13] Considering the enormous importance of olefins, the discovery of new methodologies for the synthesis of C=C double bonds containing compounds is at the heart of Organic Chemistry. Regarding the development of new protocols for the synthesis of C=C double bonds, this thesis represents new possibilities for the formation of alkenes from feedstock materials. More specifically, this thesis is devoted to the reductive cross-coupling of two different carbonyl groups into dissimilarly substituted olefins. To date, only one method is known for the coupling of two carbonyls into alkenes, namely the McMurry reaction.[14] This reaction is frequently used, in particular in an academic context, however exhibits certain drawbacks that limit its wider use. Thus, alternatives as presented in this thesis can be of high significance for the way organic chemists devise multi-step sequences to complex organic compounds in the future.

12 2. Background

The first part of this chapter introduces various available methods for the formation of C=C double bonds. Particular attention is devoted to carbonyl olefination methodologies, in which alkenes are formed from aldehydes and/or ketones. In this regard, phosphorus-mediated olefination reactions are described in detail, along with an overview on non-phosphorus-based olefination methods. The second part of this chapter is focused on the chemistry of phosphaalkenes. C=C and P=C double bonds are compared, along with details regarding synthesis, stability and applications of phosphaalkenes.

2.1 Formation of C=C bonds The formation of C=C double bonds has a tremendous importance on fundamental and industrial organic synthesis.[7a, 7b, 15] The C=C double bond is not only important as a structural motif, but serves also as a functional group for further modifications. As a result, several reactions have been developed for the synthesis of C=C double bonds, as shown in Figure 2.1.

McMurry Cope reaction Wittig

Aldol Horner-Wadsworth-Emmons

Knoevenagel C = C Julia and Julia-Kocienski bond forming reactions

Stobbe Peterson

Olefin metathesis Tebbe and Petasis Takai Figure 2.1 List of name reactions for the synthesis of C=C double bonds.[16]

13 Many different methods are available for the synthesis of alkenes. In several cases, carbonyl moieties are transformed into C=C double bonds via olefination reactions.[15b] Classical examples of this kind of reactions are: Wittig, Horner-Wadsworth-Emmons, Julia, Peterson and McMurry reactions. Among these, it should be pointed out that the McMurry coupling is the only procedure in which two carbonyl groups are directly coupled to form an olefin.

2.1.1 Carbonyl olefinations In an olefination reaction, a carbonyl group is transformed into an alkene with the formation of a new C=C double bond.[15, 17] In order to perform such transformation, the carbonyl compound is reacted with an “olefinating reagent”, as shown in Scheme 2.1.

Scheme 2.1 General reaction scheme for the olefination of a carbonyl group mediated by an olefinating reagent.

Depending on the type of olefinating reagent, a plethora of methods with different stereoselectivity, reactivity and functional group tolerance have been developed.[15b] Phosphorus-based olefinating reagents are the most studied and are intimately linked with the names of Wittig,[18] Horner, Wadsworth and Emmons (HWE olefination).[19] Other important reactions are the sulphur- based Julia olefination[20] and the silicon-mediated Peterson reactions.[21] The corresponding olefinating reagents for each of these reactions are shown in Figure 2.2.

Figure 2.2 Organo main-group olefinating reagents.

Besides olefinating reagents based on main group elements, some transition metal mediated olefination reactions have been developed. Among the various available transition metals, titanium has been extensively exploited to trans- form C=O into C=C double bonds.[22] In 1978, Tebbe reported a one-carbon homologation (methylation) of carbonyl compounds using a titanium-alumin- ium complex.[23] A similar reactivity has been shown by Petasis,[24] and now methylation of carbonyl groups via titanium intermediates is known as the Tebbe-Petasis olefination (Scheme 2.2 ).[16] Such a reaction has been applied not only to aldehydes and ketones, but also to esters, lactones and amides.[25]

Scheme 2.2 General representation of methylation of carbonyl groups via Tebbe and Petasis reagents.[16] Cp = cyclopentadiene.

Another important titanium-based reaction is the McMurry coupling.[14, 22a, 26] As previously mentioned, the McMurry reaction is the only method that directly couples two carbonyl groups to form C=C double bonds. The reaction is mediated by low-valent titanium species in which the metal functions as both electron donor and oxygen acceptor (Scheme 2.3).

Scheme 2.3 General McMurry coupling leading to the formation of mixtures of alkene products.

Due to its radical mechanism, the McMurry reaction results in mixtures of symmetrical and unsymmetrical products when two different carbonyl groups with similar reactivity are reacted.[15b]

2.1.2 Wittig and related phosphorus olefination reactions The Wittig reaction: In 1953, Wittig and Geissler reported the first olefination of carbonyls with phosphonium ylides to form alkenes and triphenylphosphine oxide in almost quantitative yield (Scheme 2.4).[18a]

Scheme 2.4 First reported olefination of benzophenone with a phosphonium ylide.[18a]

After their first report, Wittig recognized the importance of this reaction and developed a systematic study about the formation of olefins from phosphonium ylides.[18b, 27] As a result of his commitment and discoveries in this field, Wittig was awarded the Nobel Prize in chemistry in 1979 and the reaction of phosphorus ylides with carbonyls to form alkenes has taken his name.[16] Since its discovery, the Wittig reaction has become one the most important and widely used methods for the synthesis of alkenes.[28]

15 Structure of phosphonium ylides: Phosphonium ylides can be represented as the two limiting resonance structures shown in Figure 2.3.

Figure 2.3 Representation of phosphonium ylides as ylide and ylene resonance structures.

1H, 13C and 31P NMR spectroscopy data are consistent with the dipolar ylide structure with only a small contribution of the ylene form,[29] and theoretical calculations also support these data.[30]

Phosphonium ylides – types, preparation and reactivity: Depending on the substituent attached to the α-carbon, phosphonium ylides are categorized as stabilized, semi-stabilized and non-stabilized ylides, as depicted in Figure 2.4.[28a]

Figure 2.4 General classification of phosphonium ylides based on the R1 substituent.[28a] EWG = electron-withdrawing group.

The preparation of triphenylphosphonium ylides (X = Y = Z = Ph in Figure 2.4), the most frequently used, is shown in Scheme 2.5. At first, the corresponding phosphonium salts are made by the reaction of air-stable and inex- pensive triphenylphosphine and alkyl halides, which should be reactive towards SN2 displacement (Scheme 2.5, first reaction). Deprotonation of the latter finally generates the desired triphenylphosphonium ylides (Scheme 2.5, second part.)

Scheme 2.5 SN2 reaction between triphenylphosphine and a general alkyl halide leading to the formation of a phosphonium salt with subsequent synthesis of the corresponding ylide.

16 The reactivity of the ylides is highly dependent on the substituent R1 in Figure 2.4. Indeed, the three types have completely different chemical behavior and show distinctive stereochemical outcomes in their reaction with carbonyl compounds. In the case of non-stabilized (R1 = alkyl) and semi-stabilized (R1 = aryl, vinyl, halo, alkoxy) ylides, the reactivity is very high. Since these ylides are unstable towards oxygen and moisture, they cannot be isolated and are made in-situ at -78 °C under inert conditions. Because their corresponding phosphonium salts are only weakly acidic, strong bases must be used in the formation of non-stabilized and semi-stabilized ylides. Such bases are for in- t stance LDA, n-BuLi, NaNH2, NaHMDS or BuOK. Once formed, these compounds are reactive at low temperature towards both ketones and aldehydes, thus elaborated olefinic structures can be obtained. In contrast, stabilized ylides (R1 = EWG) are less reactive and can be isolated and purified. The acidity of their corresponding phosphonium salt is higher, therefore stabilized ylides can be prepared with weaker bases such as NaOH.[17] Because of their stability, these ylides are unreactive towards ketones, and elevated temperatures are required in their reaction with aldehydes.

Mechanism and stereochemistry of the Wittig reaction: Since its discovery, the Wittig reaction has been intensively studied and its mechanism has been the focus of several debates.[28a, 31] In a general reaction scheme (Scheme 2.6), the first step is the generation of the ylide followed by its reaction with a carbonyl compound, leading to the formation of phosphine oxide and alkene products.

X Ph base Ph 1 1 Ph P CH2R Ph P CHR Ph - HX Ph

O O P R2 Ph Ph Ph

R1 R2 R1 R1 R2 R1 R2 R2 ZE + ZE R1 = Alkyl R1 = Aryl, vinyl, halo, alkoxy R1 = EWG (non-stabilized ylide, Li salt-free ) (semi-stabilized ylide) (stabilized ylide) Scheme 2.6 General reaction scheme and stereochemical outcome of the Wittig olefination depending on the R1 substituent of the ylide. Adapted from[32].

Regarding the stereochemistry of the reaction, it is mainly determined by the nature of the ylide. As presented in Scheme 2.6, non-stabilized ylides under Li salt-free conditions lead preferentially to Z-alkenes. A mixture of E- and Z-

17 olefins is obtained using semi-stabilized ylides, whereas E-alkenes can be obtained with stabilized ylides.[17, 28a] Other variables, like the type of base used to form the ylide, solvent, and temperature might also affect the stereochemical outcome of the reaction.[16] The initially proposed mechanism involved a nucleophilic addition of the ylide to the carbonyl group with the generation of a dipolar intermediate (betaine), followed by the formation of a four-membered oxaphosphetane ring (OPA) which then collapses to the phosphine oxide and alkene products (Path a in Scheme 2.7).[18b, 28a, 33] An alternative and modern mechanism, under Li salt-free conditions, proposes instead direct formation of the oxaphosphetane intermediate via [2+2] cycloaddition of the ylide and the carbonyl compound (Path b in Scheme 2.7).[34] Formation of the OPA is followed by a stereospecific syn-cycloreversion with formation of the two products.

Scheme 2.7 Proposed mechanisms of the Wittig reaction.[31a, 32]

Several computational studies support the presence of OPA intermediates.[35] Moreover, OPA have been detected by low-temperature 31P, 1H and 13C NMR spectroscopic analyses in reactions of non-stabilized ylides and carbonyl compounds.[36] Thus, it is now accepted and categorically established that in all Li salt-free Wittig reactions betaines are not formed, and OPA are the only true intermediates of the reaction.[31a] There are no spectroscopic evidences of betaines under Li salt-free conditions[34b] and these intermediates have only been observed under special conditions.[37] Regarding the presence of betaine intermediates and the stereochemical outcome of the reaction, it is important to mention the Schlosser modification of the Wittig reaction.[16, 38] The Schlosser modification leads to E-alkenes when non-stabilized ylides are reacted with carbonyl groups in the presence of Li bases (e.g. n-BuLi, LiHMDS etc). When phosphonium salts are reacted with non-Li bases (e.g. NaHMDS, NaOH, tBuOK etc), the so formed non- stabilized ylides react via a [2+2] cycloaddition with carbonyl groups forming OPA intermediates. In these conditions, only Z-alkenes are formed. In contrast to this mechanism, in the Schlosser modification non-stabilized ylides are

18 formed with Li bases, and the reaction is believed to occur via Li-betaine intermediates forming E-alkenes in high stereoselectivity (Scheme 2.8).

Ph Ph Ph R1 O P 1 1 Li salts Ph R Ph P CHR R2 Ph R2 Li O R2 E trans-Li-betaine Scheme 2.8 General scheme of the Schlosser reaction.

Observations and restrictions of the Wittig reaction: As already mentioned, the Wittig reaction is probably the most widely used method to form alkenes. Since its discovery, the reaction has gained importance on industry level in the formation of various olefins. For instance, a crucial step in the synthesis of Vitamin A (estimated production 3,000 tons per year) is a Wittig reaction.[39] Thus, its impact on fundamental and industrial chemistry cannot be overemphasized, but it is interesting to observe that the Wittig reaction has never been utilized to reductively couple two carbonyl compounds. Indeed, in a retro-synthetical analysis, the Wittig reaction represents the coupling of an alkyl halide and a carbonyl group, as depicted in Scheme 2.9.

Scheme 2.9 Retro-synthetical analysis of the Wittig reaction.

One of the main drawbacks of the Wittig reaction is the poor nucleophilicity of phosphonium ylides. Because of the low reactivity of these species, hindered ketones are unreactive in Wittig reactions. Therefore, tetrasubstituted alkenes cannot be synthesized with this olefination method, and the McMurry reaction still remains one of the best protocols to synthesize hindered olefins. Moreover, a stoichiometric amount of phosphorus by-product is generated along with the desired olefin, and a fully catalytic version of the Wittig reaction has not been discovered yet. Furthermore, the triphenylphosphine oxide is not water soluble, and its separation from the alkenes can be sometimes tedious.[40] Regarding these limitations, an important related olefination method is the Horner-Wadsworth-Emmons reaction.[16]

The Horner-Wadsworth-Emmons (HWE) reaction: In 1958, Horner discovered a Wittig-type reaction of phosphine oxides to form olefins.[41] Then, in the early 1960s, Wadsworth and Emmons systematically studied the reactivity of phosphonate carbanions for the preparation of alkenes.[19a] The stereoselective olefination of aldehydes and ketones with phosphonate carbanions is now known as the Horner-Wadsworth-Emmons (HWE)

19 reaction.[15b, 16, 19b, 42] Scheme 2.10 represents a general scheme of the HWE olefination reaction.

2 3 4 O R base O R2 R R R2 R3 O R1 P R1 P R1 P O R1 R1 R4 R1 phosphate R1 = aryl, alkyl, O-aryl, O-alkyl 2 R = aryl, alkyl, EWG (COR, CO2R, CN, SO2R) R3, R4 = H, aryl, alkyl Scheme 2.10 General scheme of the HWE olefination reaction.

When R1 in Scheme 2.10 is an aryl group, the reaction is referred as Horner- Wittig olefination (phosphine oxides and carbonyls).[43] Instead, phosphonates are used in typical HWE reactions (R1 = O-alkyl, Scheme 2.10). In classical HWE conditions, the phosphonates carbanions are stabilized by electron-withdrawing groups (R2 = EWG, Scheme 2.10).[42] Thus, HWE reactions are typically employed in the synthesis of E-α,β-unsaturated esters, as shown in Scheme 2.11.

O O 1 O OR base O EtO P R2 R2 OR1 OEt stabilized phosphonate Scheme 2.11 Classical HWE reaction of stabilized phosphonate species leading to the formation of E-α,β-unsaturated esters.

The starting phosphonate reagents are prepared by the Michaelis-Arbuzov reaction of phosphites and organic halides (Scheme 2.12).[44]

Scheme 2.12 Preparation of stabilized phosphonates by means of the Michaelis-Ar- buzov reaction.[44]

One of the main advantages of the HWE reaction over the Wittig is that phosphonates are more nucleophilic than their corresponding phosphonium ylides.[42] Due to this higher reactivity, both aldehydes and ketones can be used, under milder reaction conditions, in the synthesis of di- and trisubstituted alkenes. Additionally, also sterically hindered ketones, unreactive in the Wittig reaction, show reactivity under HWE conditions.[16] It should be mentioned that tetrasubstituted olefins cannot be synthesized via typical HWE reactions either.

20 Another advantage of the HWE reaction is that di-alkyl phosphate by-products (R1 = O-alkyl, Scheme 2.10) are water soluble, and their separation from alkene products is easily achieved by simple aqueous extraction. The detailed mechanism of the reaction is represented in Scheme 2.13.[19b]

O k EtO EWG cis O P slow EtO EWG EtO P EWG R1 EtO O O 1 R R1 Z minor cis-betaine O EWG cis-OPA O EtO P EtO P O OEt OEt phosphate O ktrans O EtO EWG O P faster EtO EWG EtO P EWG 1 R1 EtO R 1 O fast O R R1 E major trans-betaine trans-OPA Scheme 2.13 Accepted mechanism of the HWE reaction.[16]

The first step is the nucleophilic attack of the phosphonate anion to the carbonyl group with formation of betaine intermediates. These anionic species are in equilibrium with the corresponding oxaphosphetanes (OPA). Elimina- tion of the phosphate byproducts, with concomitant formation of the desired olefinic products, is irreversible. Due to the irreversibility of the final step, the stereoselectivity of the reaction depends on the reversible addition of the olefinating agent to the carbonyl. Because the elimination of the E-alkene is faster, the reaction results in the preferential formation of E-olefins.[16, 45] En- hanced E-selectivity is achieved with bulky EWGs and elevated temperatures.[46] Various modifications of classic HWE reagents and conditions have been presented over the years to achieve Z-stereoselectivity.[47] Among these, Still-Gennari,[48] Ando[49] and Corey-Kwiatkowski[50] reagents are the most widely used and are shown in Figure 2.5.

Figure 2.5 Molecular structures of modified HWE reagents for Z-selective olefinations.

The main disadvantage of the HWE reaction is that phosphonates require EWG groups for the stabilization of their corresponding carbanions and for the final step to occur. In case of non-stabilized phosphonates (R1 = aryl or alkyl, Scheme 2.10) the reaction can stop at the betaine intermediate yielding β-hydroxyphosphonates.[15b, 42, 51]

21 In conclusion, the HWE reaction represents a powerful variation of the Wit- tig olefination. Phosphonate reagents employed in the HWE protocol are more reactive than phosphonium ylides, so ketones can be used for constructing C=C double bonds. Moreover, the phosphate by-products are easily removed via aqueous work-up. As shown in Scheme 2.9 for the Wittig reaction, it should be remembered that the HWE reaction represents, in a retro-synthetical analysis, the coupling of an organic halide with a carbonyl compound. Fur- thermore, also in the HWE reaction an equimolar amount of phosphorus by- product is generated during the reaction.

2.1.3 Phosphorus free olefination reactions Various non-phosphorus mediated olefination reactions were briefly introduced in section 2.1.1. In this section, a more detailed look into sulphur- and silicon-based olefination methods (Julia and Peterson reactions, respectively) is presented.

Sulphur-based olefination – the Julia reaction: In 1973, Julia reported a new olefination reaction based on the use of phenylsulfones as olefinating reagents.[20a] A few years later, Lythgoe and Ko- cienski explored the scope and limitations of this method,[20b, 52] that today is known as Julia-Lythgoe and Julia-Kocienski olefination reactions.[16, 53] In the first version (Julia-Lythgoe variation), the reaction was divided in three steps (Path a, Scheme 2.14): nucleophilic addition of a sulfonyl stabilized carbanion to a carbonyl group, acylation of the formed β-hydroxysulfone intermediate followed by elimination, under reductive conditions, to form the desired alkene.[20a] After initial formation of the β-hydroxysulfone intermediate, acylation of the substrate is required prior to the reductive elimination step. This final part of the reaction can be performed with different reducing [54] agents such as Li or Na in ammonia, alkali (Hg) amalgam, LiAlH4 or SmI2.

Scheme 2.14 Classic Julia-Lythgoe (Path a, Ar = Ph) and modified one-pot Julia-Ko- cienski (Path b, Ar = Het.) olefination reactions.[16]

The classic procedure is quite tedious to perform, and usually the modified Julia-Kocienski variation is preferred.[16] In this second generation reaction, the sulfone reagents bear heterocycle moieties instead of the phenyl ring used in the first variation.[55] The role of the heterocycle ring is to provide a non- reductive one-pot mechanism for the elimination step from the β-hydroxysulfone intermediate.[56] The hetero-substituted intermediate is very labile and it rapidly undergoes fragmentation to SO2 and alkene products via Smiles-type rearrangement (Path b, Scheme 2.14).[53, 56-57] Benzothiazole (BT) and 1-phenyl-1H-tetrazole (PT) sulfones are the most widely used, although other het- erocycles have also been developed.[55-56, 58] The main advantages of these procedures are high versatility, wide functional group tolerance, and very high degree of E-stereoselection.[16] Due to the great E-stereoselectivity, the Julia reaction has been widely applied in the synthesis of various E-alkene products.[59]

Silicon-based olefination – the Peterson reaction: In 1968, Peterson showed that α-trimethylsilyl organometallic compounds can react with carbonyl groups forming the corresponding alkene.[21a] Nowadays, the formation of alkenes from the reaction of aldehydes or ketones with α-silyl carbanions is known as the Peterson olefination,[16, 21b, 60] the general reaction pathway of which is reported in Scheme 2.15.

Scheme 2.15 General reaction pathway of the Peterson olefination.

23 The stability of β-hydroxysilanes depends on the electronic nature of the R2 substituent in Scheme 2.15: with electron-withdrawing groups (EWG), the stability of the β-hydroxysilanes is low and these intermediates undergo a rapid and spontaneous elimination to form E- and Z-alkenes, with E-products being slightly favored over the Z. When starting α-silyl carbanions are substituted with electron-donating groups (R2 = EDG in Scheme 2.15), the corresponding β-hydroxysilanes are stable and isolable. Thus, the two diastereoisomers can be purified and subsequently converted into their corresponding alkenes, via basic or acidic elimination,[61] with complete control of the stereochemical outcome of the reaction.[60a, 62] Besides purification of β-hydroxysilanes intermediates by means of column chromatography, many approaches have also been developed for their stereoselective synthesis.[63] As stated, diastereomer- ically pure β-hydroxysilanes can undergo stereoselective elimination to the alkene products under basic or acidic conditions. In general, when β-hydroxysilanes are treated with bases (e.g. NaH, KH, tBuOK), stereospecific syn-elimination occurs, whereas stereospecific anti-elimination is observed by treating β-hydroxysilanes with Lewis acids (e.g. AcOH, H2SO4, BF3- [60b, 64] OEt2). The mechanisms under basic and acidic conditions are presented in Scheme 2.16 and Scheme 2.17, respectively.[16, 21b]

Scheme 2.16 The Peterson elimination under basic condition starting from diastereomeric pure syn β-hydroxysilane.[16]

Concerning the mechanism under basic conditions (Scheme 2.16), two different variations are feasible. In a stepwise fashion (upper part of Scheme 2.16), a [1,3] silyl shift is followed by elimination of a silyloxide moiety with concomitant formation of a pure isomeric alkene. In the concerted pathway (lower part of Scheme 2.16), a Wittig-type four membered ring is formed which de- 1 1 composes, via [2+2] cycloreversion, to alkene and (R )3Si-O-Si(R )3 products. In this case, the driving force of the reaction is the formation of the strong Si- O bond. Regarding the mechanism under acidic conditions (Scheme 2.17), stereospecific anti-elimination occurs through protonation of the hydroxyl group

24 and simultaneous dehydration and desilylation to yield isomerically pure alkenes.

Scheme 2.17 The Peterson elimination under acidic conditions.[16]

In conclusion, the Peterson olefination is a useful method to convert carbonyl compounds into pure E- and Z-alkenes. The reaction offers high stereoselectivity and reactivity, allowing both ketones[65] and aldehydes[66] to be employed in the synthesis of di- and trisubstituted olefins. Due to the high nucleophilicity of the α-silyl carbanions, Peterson reagents are more reactive than phosphonium ylides used in the Wittig reaction.[21b, 67] Furthermore, the silicon based by-products of the reaction are usually volatile (disiloxane) and are thus easier to remove compare to triphenylphosphine oxide generated in typical Wittig reactions. However, the low availability of some α-silyl carbanions limits its general use.[68]

25 2.2 Phosphorus as the Carbon copy Looking at the periodic table, one might expect many differences in chemical behavior, structural properties, and reactivity between carbon and phosphorus, respectively located in groups 14 and 15. However, these two elements are linked by a diagonal relationship and, when phosphorus is in a low-coordinate state, the two have many similarities.[2-3] Due to this connection, phosphorus is often referred in the literature as a “carbon copy”[3b] or “carbon photo- copy”.[69] One similarity between P and C lies in their ability to accept and release electrons. The two atoms have analogous first ionization potentials (IP, IPP = [70] 1011.8 kJ/mol and IPC = 1086.5 kJ/mol). Furthermore, the two elements have similar valence orbital ionization energies: -18.8 eV (3s) and -10.1 eV (3p) for phosphorus, -19.4 eV (3s) and -10.6 eV (3p) for carbon.[71] Another crucial comparison involves their electronegativities: their effective π-electronegativities are nearly similar, whereas the σ-electronegativity of phosphorus is slightly lower than that of carbon (2.1 for P vs. 2.5 for C, according to the Pauling scale).[72] This last point has a crucial impact on the chemistry of phosphaalkenes, which will be presented in the following sections.

2.2.1 Comparison of P=C and C=C double bonds Starting from the late 1940s, when the concept of the “classic double bond rule” was introduced, it was believed that elements of the third row in the periodic table could not form stable compounds with (p-p) π bonds.[73] How- ever, a great development in main group chemistry has proven this theory to be wrong. Phosphaalkenes are compounds containing phosphorus-carbon double bond (P=C) in which the phosphorus atom has a valency of three (λ3) and a coordination number of two (σ2).[2] The first synthesis of phosphaalkenes was reported by Becker in 1976.[1] Since this first article, many other examples of low-coordinate main group compounds have been realized. Concerning the field of low-coordinate phosphorus chemistry, compounds like diphosphene (P=P),[74] phosphaalkyne (C≡P),[75] silaphosphene (Si=P)[76] and phos- phaallene (P=C=C)[77] have also been synthesized. The focus of this thesis is on phosphaalkenes and their chemistry. Due to the previously described similarities between phosphorus and carbon, the P=C unit resembles the C=C double bond. Thus, phosphaalkenes can also be described as “heavy olefins”. Regarding the polarity of P=C double bonds, due to the small difference in the effective π-electronegativities of P and C, the (3p-2p) π-bond in phosphaethene (HP=CH2) is apolar like the (2p-2p) π-bond of ethene (H2C=CH2). In- stead, because the σ-electronegativity of phosphorus is lower than that of carbon, the σ component of the P=C double bond is highly polarized (Pδ+-Cδ-).[78]

26 Concerning the reactivity of P=C and C=C units, a closer look to the energy of their highest occupied molecular orbital (HOMO) is suggested (Figure 2.6).

[79] Figure 2.6 Comparison of HOMO energy levels between H2C=CH2 and HP=CH2.

As shown in Figure 2.6, the HOMO levels of the two molecules are close to each other. The HOMO of ethene is slightly lower in energy than the π-MO of phosphaethene. As a consequence, the π-systems in phosphaalkenes are expected to have higher reactivity than these in alkenes.[3d] Moreover, because the lone pair of P (nP) is quite high in energy, it also contributes to the reactivity of P=C double bonds.[3b] Another important aspect regarding the higher reactivity of the P=C unit is shown by thermodynamic data which report a lower strength for the P=C double bond vs. the C=C double bond (calculated dissociation energies: 45 kcal/mol for P=C and 65 kcal/mol for the C=C double bond).[80] As a result of these similarities in the energy levels of their HO- MOs, phosphaalkenes and alkenes exhibit a comparable general reactivity. For instance, phosphaalkenes take part in many standard olefin reactions, such as hydrogenations, polymerizations, epoxidations, cycloadditions and other (more details in section 2.2.4).[2, 3b] Despite the similarities described above, there are also some important differences between phosphaalkenes and alkenes. P=C double bonds, compared to C=C double bonds, contain a phosphorus atom which has a reactive lone pair. This moiety is available as an alternative binding site for transition metals[3c] and also for oxidation of the P(III) center to a P(V) center.[2] Moreover, the energy of the lowest unoccupied molecular orbital (LUMO) in phosphaalkenes is considerably lower compared to that of alkenes. As a result, the HOMO-LUMO gap in phosphaalkenes is quite narrow.[3d, 81] This detail is the origin of many recent and interesting applications of phosphaalkenes in the field of polymer science, organic- and opto-electronics (more detailed in section 2.2.4).[4a, 6a, 69, 81a, 82]

2.2.2 Stabilization of phosphaalkenes As described above, phosphaalkenes have a reactive P=C double bond in their backbone which makes these systems prone to decomposition or unwanted polymerization side reactions, if precaution is not taken. Two strategies are available to protect and stabilize the fragile P=C units, namely thermodynamic and kinetic stabilization of P=C double bonds.[83]

27 Thermodynamic stabilization is achieved as a result of delocalization/con- jugation of P=C units into cyclic aromatic systems. For instance, phospha- benzene and 2,4,6-tri-phenylphosphabenzene have been synthesized taking advantage of this type of stabilization.[84] Kinetic stabilization increases the energy level of the transition state towards the decomposition of P=C double bonds. Such stability is achieved when complexation and/or steric hindrance at the phosphorus atom are pro- vided.[3a] Regarding stabilization via complexation, phosphaalkenes can form stable complexes with various transition metals in different ways.[3b, 3c, 5a, 85] The most common complexation is through the phosphorus lone pair, leaving the P=C double bond untouched and available for further reactions.[3c, 86] Other possibilities are complexation via the P=C double bond along with the phosphorus lone pair.[3c, 5a, 87] Pertinent to this thesis is the kinetic stabilization of acyclic transition metal- free phosphaalkenes. In order to stabilize fragile acyclic P=C double bonds, sterically demanding groups can be installed on either one or both termini of the P=C unit. Many examples of protecting groups have been reported over the last decades and are summarized in an excellent review by Joshifuji.[88] Among these examples, the most established protecting group is “Mes*” or “super mesityl” (2,4,6-tri-tert-butylphenyl).[74, 89] Other groups, whose molecular structures are presented in Figure 2.7, are 2,4,6-triisopropylphenyl (Tip), 2,4,6-trimethylphenyl (Mes or mesityl), phenyl (Ph), 2,6-dimesitylphenyl (Dmp), adamantyl, and tris(trimethylsilyl)methyl.[88, 90]

tBu tBu

tBu

Mes* Tip Mes Ph

SiMe Me3Si 3 SiMe3

dmp adamantyl tris(trimethylsilyl)methyl Figure 2.7 Molecular structures of commonly used bulky protecting groups for the kinetic stabilization of phosphaalkenes.[88] This thesis is focalized on the use of Mes and Ph protecting groups, highlighted in the box.

In general, the stability of phosphaalkenes decreases with reduced steric demand of the protecting group and follows the order Mes* > Tip > Mes > Ph. A more detailed discussion is given in sections 3.3 and 6.2.

28 2.2.3 Synthetic pathways towards phosphaalkenes Many procedures are currently available for the synthesis of phosphaalkenes.[2-3, 91] A selection of the most important methods is reported in Scheme 2.18.

SiMe3 PR 1 2 SiMe SiMe R COR 3 R1COR2 3 K P TMLn PR B R SiMe3 R1COCl

RMgX R1COR2 SiMe3 J R1 PC PR C Li R2 PC R R1 H SiMe base cat. CO2 3 I P C PR D 1 R R SiMe3

H 1 2 Cl R COR HCX3 H PR PR E H Cl R1CX base H 2 Cl H PR P C R2 H R R1 G F Scheme 2.18 General synthetic pathways towards phosphaalkenes. Adapted from[32].

Becker et al. used a condensation reaction of bis(trimethylsilyl)phosphine (RPTMS2) with acyl chloride, followed by a 1,3-silyl shift (route A in Scheme 2.18), to synthesize the first stable phosphaalkene.[1] Due to the similarity of this reaction with the Peterson olefination (see section 2.1.3), the formation of phosphaalkenes starting from silylated phosphines and carbonyl compounds is presently known as the phospha-Peterson reaction (more details in section 4.2).[91b, 92] The reaction of di- and mono-silylated phosphines with carbonyl compounds represents a very useful method for forming phosphaalkenes. In- deed, RPTMS2 can also form P=C double bonds upon reaction with aldehydes and ketones (route B in Scheme 2.18). Such a reaction can be catalyzed by [91b] [93] bases (e.g. KOH) or mediated by Lewis acids (e.g. AlCl3) , or it can also be performed with lithiated mono-silylated phosphines (route C in).[77b, 92a, 94] The generality of the silatropic shift in the phospha-Peterson reaction can be observed in the reaction of RPTMS2 with small molecules like CO2 (pathway D).[95] Another classical preparation of phosphaalkenes is based on the 1,2-elimination of HX (dehydrohalogenation) from an appropriate precursor (methods E, F and G).[91d, 96] P-P-dichlorophosphines are typical starting material that

29 can react with a haloform or tetrahalomethane generating C-C-dihalo substituted phosphaalkenes (E).[97] Halophosphines containing an acidic α-proton can react with bases (e.g. Et3N, DBU, DABCO) forming P=C double bonds via a dehydrohalogenation reaction (route F).[98] Another possibility is the in- I sertion of carbenes (R CX2) to primary phosphines under basic conditions (G).[99] Phosphaalkenes can also be prepared by the reaction of primary phosphines and carbonyl derivatives (route H)[100] as imines are generated by combining primary amines and appropriate carbonyl groups. As the reaction generates water as a by-product, it can be performed with drying agents such as [2] P4O10 and CaO/CaCl2. Upon treatment with a catalytic amount of base, secondary vinylphosphines can form P=C double bonds through a 1,3-proton shift (method I).[101] Grignard reagents can perform a nucleophilic attack to phosphaalkynes to generate metal substituted P=C double bonds in which the nucleophilic R moiety of the organometallic reagent is now attached to the phosphorus atom. The so formed intermediates can further react with various electrophiles generating the desired phosphaalkenes (J).[102] Lastly, transition metal terminal phosphinidene complexes can generate P=C double bonds upon reaction with different carbonyl groups (route K).[103] This reaction can be seen as the phosphorus version of the Tebbe olefination (discussed in section 2.1.1). As the Peterson olefination has found its phosphorus version, namely phospha-Peterson reaction, the Wittig and HWE reactions also have their phosphorus variations, respectively called “phospha-Wittig” and “phospha-HWE” reactions.[3d, 104] General schemes of these reactions, along with the comparison of classic Wittig and HWE olefinations, are shown in Scheme 2.19.

Scheme 2.19 Phosphorus variations of Wittig and HWE olefinations towards phosphaalkenes. Left: transition metal-coordinated (top) and metal-free (middle) phospha-Wittig reactions; classic Wittig olefination from phosphonium ylide (bottom). Right: transition metal-coordinated (top) and metal-free (middle) phospha-HWE reactions; classic HWE olefination with phosphonate reagents (bottom).

30 In the phospha-Wittig reaction, a phospha-ylide (or phosphoranylidene phosphine, left part of Scheme 2.19) reagent is reacted with aldehydes (but not ketones) to form the corresponding phosphaalkene.[104] In initial reports, metal- coordinated (W, Mo or Fe) phospha-Wittig reagents were used to stabilize both the starting material and the final phosphaalkene products.[105] A few years later, the group of Protasiewicz reported the first metal-free phospha- Wittig reaction.[106] Following the discovery of the phospha-Wittig reaction, metal-coordinated phosphanyl phosphonates (right part of Scheme 2.19) were used in the formation of P=C double bonds by a method that is now referred to phospha- HWE reaction.[104] Since these reagents are more reactive than that used in the phospha-Wittig reaction, ketones could also be employed for the synthesis of trisubstituted phosphaalkenes.[3a, 105b, 105d] Metal free variations of the phospha- HWE reaction are now possible as well.[89a, 107]

2.2.4 Reactivity and applications of phosphaalkenes As already mentioned, phosphorus and carbon have many similarities, and indeed phosphaalkenes possess analogous reactivity to that of olefins.[3a, 3d] Fur- thermore, phosphaalkenes can take part in classic reactions of alkenes such as hydrogenations, epoxidations, E/Z photoisomerizations,[108] and cycloadditions[3e, 109] among others.[2-3, 5a, 85a] On the other hand, some differences exist regarding the chemistry of phosphaalkenes and olefins. These divergences arise, for instance, from the presence of a lone pair on the phosphorus atom, from the different oxidation states of the phosphorus center and from the polarity of the P=C double bond. A general overview of the peculiar reactivity of phosphaalkenes is presented in Scheme 2.20.

X R2 X Nu H H Nu PC or R P PCR1 or PCR1 R R1 X R R2 R R2 oxidation 1,2-addition X = O, S [X] NuH

R2 PC R R1 R3Li TMLn MeOH

R1 [L MT] R2 n 3 PC R PC H R R1 R R2 TM coordination polymerization Scheme 2.20 Overview of some typical reactions of phosphaalkenes.

As discussed in section 2.2, due to the difference in electronegativity between phosphorus and carbon, the P=C double bond is normally polarized as Pδ+Cδ-

31 .[3a, 3b, 3d, 110] Therefore, phosphaalkenes can be attacked by several polar reagents with cleavage of the π-P=C double bond (more details in section 4.3). Protic reagents (NuH) like hydrogen halides, alcohols, amines and thiols can thus perform 1,2-additions onto P=C double bonds (Scheme 2.20).[2, 91d, 111] Depending on the substitution at the phosphorus and carbon atoms of the P=C double bond, these systems can be described as “classically” and “inversely” polarized phosphaalkenes.[91c] As a result, a nucleophile can attack both the phosphorus atom (Pδ+ in classic phosphaalkenes) or the carbon atom (Pδ- in inversely polarized phosphaalkenes), as shown in Scheme 2.20 (detailed in section 4.3). Another consequence of the polarity of the P=C double bond is the reactivity of phosphaalkenes in polymerization reactions. Small alkyl lithium bases can add into the P=C double bond and act as initiators for anionic polymerizations of sterically unprotected phosphaalkenes (more details in section 4.3).[4a, 4b, 112] Moreover, phosphaalkenes can coordinate transition metals in many ways, not only via the lone pair but also through the P=C double bond.[3c, 5a] This is a classical method to protect unstable phosphaalkenes from degradative side- reactions such as self-oligo- and polymerizations.[105a, 113] Lastly, because the phosphorus atom has different possible oxidation states, phosphaalkenes can also react with oxidants like sulphur, oxygen and ozone.[111] Moving from σ2λ3 phosphaalkenes, σ3λ5 metaphosphinates and σ2λ5 metaphosphates species can be formed when P=C double bonds are sub- jected to oxidation reactions.[89d, 114] Interesting to note is that metaphosphinates have been used in several occasions in Wittig-type olefination reactions, as shown in Scheme 2.21.[115]

Scheme 2.21 Wittig-type olefination from an unstable metaphosphinate species.[115]

Metaphosphinates are unstable and generated in-situ with a flash-photolysis of the corresponding phosphoryl-substituted diazo compound. Once formed, they react with carbonyl groups via a [2+2] cycloaddition forming a stable oxaphosphetane (OPA) which collapses to the olefin and the metaphosphate products upon thermal activation.[114, 116] Phosphaalkenes do not only show a wide range of reactivity, but can also be employed in various applications, such as organic-electronics, polymer and coordination chemistry etc.

32 The groups of Protasiewicz and Gates exploited the phospha-Wittig and phospha-Peterson reactions to form P=C double bonds which produced phosphorus containing polymers upon anionic treatment.[4, 117] Due to the ability of phosphaalkenes to coordinate metals, metal-coordinated phosphaalkenes have been applied as ligands in various catalytic reactions.[3c, 5b] Moreover, phosphaalkenes were also recently used as ligands in the stabilization of gold nanoparticles.[118] Another interesting field for application of phosphaalkenes is in organic- electronics.[6b, 119] Indeed, incorporation of phosphorus into the π-conjugated framework of organic molecules results in alteration of the properties of the materials.[6a, 82] The small HOMO-LUMO gap of the P=C double bond can be used to tune the optical properties of phosphorus-based materials. Due to this unique effect, phosphaalkenes have become valuable building blocks of new materials with applications in organic electronic devices such as OLEDs, semi-conductors, and photovoltaic.[6c, 6d]

33 3. Reactivity patterns of a potential new olefinating reagent (Paper I)

This chapter is dedicated to the reactivity study of a new potential reagent in the field of olefination reactions. An introduction to existing methods regarding the coupling of two carbonyl groups, such as the McMurry reaction and a method previously developed in our lab are presented in this chapter. Building on this first proof-of-concept study, a new broader reaction scope is demonstrated herein, along with the synthesis and application of a new potential coupling reagent.

3.1 Alkenes from two carbonyl compounds – the McMurry reaction The first reports on a reductive coupling of carbonyl compounds to alkenes mediated by low-valent titanium reagents were made by Mukayama and Tyr- [15b, 26a] lik. TiCl4-Zn and TiCl3-Zn were respectively used by the two groups as coupling reagents.[120] One year later (1974), McMurry reported a reductive [14] coupling of carbonyls into olefins using TiCl3 and LiAlH4. As a result of McMurry's development in establishing this process,[121] the coupling of two carbonyl compounds to alkenes mediated by low-valent titanium reagents is known as the McMurry reaction.[16] Since its discovery, it has been used in a wide range of applications,[15b, 16, 26b, 45, 122] such as in the preparation of the drug 4-hydroxytamoxifen[123] and in macrocyclization of various natural products.[124] In most cases, it is used for reductive homo-couplings of aldehydes and ketones to form olefins; diols can also be formed, depending on the reaction conditions.[26a, 125] The reaction proceeds via titanapinacol intermediates with a radical mechanism (Scheme 3.1).

Scheme 3.1 General reaction scheme of the “low-valent” promoted intermolecular homo-coupling of two carbonyls to isomeric alkenes.[16]

34 The mechanism of the McMurry reaction is not yet entirely clear, as it is not certain which factors determine the stereoselectivity of the procedure. Gener- ally, the reaction has poor stereoselectivity with E-isomers being slightly favored over the Z. The driving force of the reaction is the formation of TiO2, and many variations have been developed for tuning the reactivity of initial Ti0 species.[45] The most common way of preparing the low-valent titanium reducing agents is the reduction of TiCl3 with a zinc-copper couple (Zn-Cu) in DME.[121] The best feature of the McMurry reaction is that sterically hindered olefins, which cannot be synthesized by other means (e.g. Wittig type olefination reactions), are formed in high yield and even sterically hindered tetrasubstituted alkenes can be prepared with this method. On the other hand, the reaction has some severe drawbacks. Functional groups that are prone to reduction are incompatible with the harsh conditions required by the McMurry reaction. Examples of such groups are allylic and benzylic alcohols,[126] epox- ides,[127] oximes,[128] sulfides[129] and nitro compounds.[130] Moreover, the McMurry reaction is not ideal for preparing unsymmetrical alkenes (see for example Scheme 2.3) due to an intrinsic lack of selectivity and it usually leads to mixtures of symmetrical and unsymmetrical alkenes in the coupling of two different carbonyl groups.[131] Because of this limitation, the McMurry reaction is more often used for homocoupling reactions. A strategy to obtain desired unsymmetrical olefins is to use an excess of one component to achieve high conversion of the more valuable reactant.[121] Another strategy to selectively form unsymmetrical alkenes is to use at least one diaryl ketone. [26a, 131] In conclusion, the McMurry reaction is a powerful method to reductively couple two carbonyls to symmetrical alkenes, especially when sterically hindered and/or strained olefins are to be synthesized.[132] However, the reaction is often described as “tricky” because of its intolerance to certain functional groups and difficulties in preparation of the active Ti0 species.[16]

3.2 Novel methodology for aldehyde-aldehyde couplings via phosphaalkene intermediates An unprecedented one-pot reductive coupling of two aldehydes to unsymmetrical E-alkenes has recently been developed in our lab.[133] The new reaction is free of transition metals and proceeds via a one-pot protocol under ambient temperature within minutes in good overall yields. The reported procedure can be mechanistically divided into three steps, as depicted in Scheme 3.2.[133]

Scheme 3.2 General reaction scheme of the new one-pot reductive cross-coupling of two aldehydes to E-1,2-disubstituted unsymmetrical alkenes.[133]

In the first step (phosphaalkene formation in Scheme 3.2), the phospha-HWE reaction[89a] between a lithiated phosphanyl-phosphonate and various aldehydes affords λ3σ2 P-Mes*-phosphaalkene intermediates. Interesting to notice is that phosphaalkene formation is concomitant with a change in polarity (Um- polung) of the carbon center from δ+ to δ-. Because of the electronic distribution of the P=C double bond, with the P atom bearing partial positive charge (see section 2.2),[110] phosphaalkene intermediates can be attacked by a nucleophilic base like TBAOH, forming the corresponding phosphinites.[111] The second step (phosphaalkene activation in Scheme 3.2) represents “the bridging” between the chemistry of low-coordinate phosphaalkenes (PIII) and the chemistry of high-valent (PV) phosphine oxides. The newly formed PIII phosphinites are known to tautomerize[134] to their corresponding PV phosphine oxides which bear resemblance to classic Horner-Wittig olefinating reagents (see section 4.4). In the third and final step (alkene formation in Scheme 3.2), λ5σ4 phosphine oxides react with a second carbonyl to form desired olefinic products. In comparison to the McMurry reaction, the method described herein has several advantageous features. The reaction doesn’t require refluxing conditions, transition metals aren’t used, and desired alkenes are synthesized within a few minutes compare to several hours that are usually required under typical McMurry conditions.[14, 122] Most importantly, in contrast to the McMurry reaction, this new methodology proceeds under an ionic mechanism which allows for the selective synthesis of unsymmetrical alkenes, without the formation of the undesired symmetrical products.[133] It should be pointed out that the above developed method represents the first intermolecular reductive ionic coupling of two different aldehydes into dissimilarly substituted alkenes. However, like many reactions in organic chemistry, also this methodology has some intrinsic drawbacks. Because of the steric hindrance of the bulky Mes* protecting group, the reaction has a very limited substrate scope. Firstly, only couplings between two aldehydes are feasible, and no ketones can be used in the synthesis of elaborated olefins. Second, aldehydes must be electron-poor. Therefore, only aldehydes bearing EWG groups can be employed

36 in the new method. P-Mes*-phosphaalkenes made from electron-rich aldehydes are not reactive towards any base, preventing the formation of the crucial λ5σ4 olefinating reagents. Furthermore, the second aldehyde has to be selected with care, since strongly electron-donating carbonyls are not suitable for the reaction.[133] Lastly, the atom economy of this method has to be considered. Compared to the Wittig reaction, in which a stoichiometric amount of phosphorus by-product is generated along with the desired alkene, in this procedure two phosphorus atoms are needed and consumed for each C=C double bond that is formed. Considering these issues, a new smaller protecting group could overcome the above-mentioned limitations, opening up the possibility of using electron- rich aldehydes and also ketones in the construction of more elaborated alkenes.

3.3 Reducing the size of the phosphorus protecting group In the field of low-coordinate organo-phosphorus compounds, the use of Mes* protecting group is a popular strategy to kinetically stabilize P=C and P=P double bonds (see section 2.2.2).[74, 88, 91a] Phosphaalkenes bearing smaller groups, such as Mes and Ph are known, but result in synthetic challenges. Therefore, only a few examples are reported in the literature.[91b, 113] Generally, trisubstituted phosphaalkenes that originate from ketones are more stable than the corresponding disubstituted ones.[105b, 105d] Thus, P-Mes-trisubstituted- phosphaalkenes are isolable,[91b, 94] whereas P-Mes-disubstituted-phosphaalkenes are not. Further decreasing the size to P-Ph-phosphaalkenes results in non-isolable compounds that rapidly dimerizes or oligomerize, even when ketones are used for the synthesis of trisubstituted P-Ph-phosphaalkenes (more detailed in section 6.2).[113] Figure 3.1 shows an up-to-date overview of phosphaalkenes' stability protected by Mes*, Mes and Ph groups.

Figure 3.1 List of tri- and disubstituted phosphaalkenes. The stability of these compounds decreases from left to right (left is the most stable, right is the least one).

37 In order to develop a new reductive cross-coupling of deactivated aldehydes and even ketones, P-Mes-phosphaalkenes seemed to be appropriate starting materials. Trisubstituted P-Mes-phosphaalkenes are stable,[83] and if reactive towards bases, new λ5σ4 olefinating reagents could be synthesized. The reduced size of the protecting group could enhance the reactivity of this olefinating reagent, hoping for the generation of more elaborated olefinic structures. Disubstituted P-Mes-phosphaalkenes have only been detected by 31P NMR,[113] but are not stable enough to be isolated. For the purpose of our work, it might be sufficient to synthesize disubstituted P-Mes-phosphaalkenes in high yield which could be quickly reacted with a nucleophilic base, to perform a similar reaction to that of Scheme 3.2. Thus, the chemistry of Mes protected compounds has been investigated, to understand the plausibility of these hypotheses.

3.4 A new potential coupling reagent - preparation and reactivity This section deals with the planning, synthesis and reactivity study of a new possible olefinating reagent based on the Mes protecting group.

3.4.1 Planning a new broader cross-coupling reaction As specified in section 3.3, Mes based λ3σ2 and λ5σ4 compounds could be ideal for new cross-coupling reactions of deactivated aldehydes and, to some extent, ketones. A preliminary plan of a new possible reaction pathway is shown in Scheme 3.3.

Scheme 3.3 General plan for a new broader one-pot cross-coupling of deactivated aldehydes and/or ketones to form unsymmetrical alkenes.

In the designed reaction scheme, “Mes-P” is a generic Mes containing phosphorus compound which could generate stable P-Mes-trisubstituted phosphaalkenes upon reaction with several ketones. These species are known to be attacked by several nucleophiles,[111, 135] therefore a hydroxide could add into the P=C double bond creating the corresponding phosphinite. Tautomerization

38 to the λ5σ4 phosphine oxide could generate the new coupling reagent which could finally form the desired alkene in reaction with a second carbonyl compound. Considering all potential synthetic difficulties that may emerge when phosphaalkenes with decreased kinetic stabilization are prepared, we were inter- ested in exploring the reactivity of λ5σ4 Mes phosphine oxides and to compare it with previously reported Mes* analogue, as shown in Figure 3.2.

tBu O Ph O Ph P tBu P H Ph H tBu Mes vs Mes* Figure 3.2 Mes and Mes* phosphine oxides, objects of the comparison.

As described above, phosphine oxides are crucial intermediates in the planned coupling sequence, therefore their downstream chemistry is imperative to ex- amine. Thus, it was decided to study the deprotonation behavior of pure isolated λ5σ4 Mes phosphine oxides for understanding their reactivity in basic environment.

3.4.2 Synthesis of benzhydryl(mesityl)phosphine oxide To probe the reactivity in the olefination step of the new potential Mes-based coupling reagent, its synthesis was performed avoiding challenging phosphaalkene chemistry, as reported in Scheme 3.4.

Scheme 3.4 Preparation of target compound 2-H.

Compound 1 was synthesized according to literature procedures.[136] 2-H was then prepared treating 1 with a concentrated aqueous HBr solution, until full conversion of starting material 1 was achieved. The reaction was carefully monitored with 31P NMR spectroscopy in which the signal of reagent 1 (δ = +67 ppm) disappears and a new peak of compound 2-H arises (δ = +24 ppm) in less than one hour. With compound 2-H in hand, a series of experiments were performed in order to deeply understand its behavior in basic conditions, with the ultimate goal of reacting 2-H with benzaldehyde to form the target olefin.

39 3.4.3 Reactivity study of benzhydryl(mesityl)phosphine oxide The deprotonation behavior of 2-H was investigated to understand its most acidic position. Therefore, 2-H was treated with 1 or 2 equivalents of n-BuLi and subsequently reacted with a methylating reagent, CH3I, to understand whether the P-H or C-H bond is methylated, as shown in Scheme 3.5.

1) 1 eq. BuLi O Ph 2) CH3I Mes P

CH3 Ph O Ph 3 Mes P H Ph 2-H O Ph Mes P CH 1) 2 eq. BuLi 3 CH3 Ph 2) CH3I 4 Scheme 3.5 Reactivity of 2-H with 1 or 2 equivalents of n-BuLi with consequent methylation reaction forming products 3 and 4.

A crucial analysis for understanding the reactivity of 2-H in the reaction conditions of Scheme 3.5 was performed via 1H NMR spectroscopy. Diagnostic signals of compounds 2-H, 3 and 4 are reported in Table 3.1.

Table 3.1 Set of diagnostic 1H NMR signals of compounds 2-H, 3 and 4.

P-H αC-H P-CH3 αC-CH3 8.10 ppm, dd: 4.53 ppm, dd: 2-H 1 3 2 3 none none JHP 477, JHH 3 Hz JHP 14, JHH 3 Hz 4.47 ppm, d: 1.90 ppm, d: 3 none 2 2 none JHP 9 Hz JHP 12 Hz 1.90 ppm, d: 1.62 ppm, d: 4 none none 2 3 JHP 16 Hz JHP 7 Hz

The 1H NMR spectrum of 2-H shows a dd at 8.10 ppm, which is assigned to 1 the proton directly attached to the phosphorus atom ( JHP = 477 Hz) and a second dd at 4.53 ppm, which is assigned to the proton at the α-benzyl carbon. When 2-H is treated with 1 eq of n-BuLi and excess of CH3I (upper row in Scheme 3.5), the large dd at 8.10 ppm is no longer observed in the 1H NMR spectrum of the reaction mixture. This is a clear indication of the absence of the proton directly attached to the phosphorus atom, suggesting reaction at the phosphorus center with subsequent formation of product 3. The formation of 3 is confirmed by analyzing the aliphatic region of the reaction mixture´s 1H NMR spectrum: a new doublet at 1.90 ppm integrating 3 protons appears, confirming the methylation at the phosphorus atom with formation of product 3. Treatment of 2-H with 2 equivalents of n-Buli and excess of CH3I (lower row in Scheme 3.5) results in the formation of product 4. The 1H NMR spectrum of the new reaction mixture does not show any signals in the region from 4 to 6 ppm, which could be assigned to protons at the α-benzyl carbon. The 1H

40 NMR spectrum features instead two doublets at 1.90 and 1.62 ppm, which could be assigned to the methyl groups at the phosphorus and carbon atoms, respectively. The analysis reported above shows some important features regarding the reactivity of 2-H in basic environments: with only 1 equivalent of base, the most acidic moiety of the molecule is the P-H bond, with no reactivity detected on the α-benzyl carbon atom. With 2 equivalents of base, the α-benzyl carbon atom is also deprotonated and the molecule now can react on both the carbon and phosphorus centers, as depicted in Figure 3.3.

Figure 3.3 Reactivity pattern of mono and doubly-lithiated 2-H derivatives, obtained treating 2-H with 1 or 2 equivalents of base, respectively.

Considering the formation of alkenes in the reaction of 2-H with carbonyl compounds as the primary goal of the project, reactivity at the α-benzyl carbon is a fundamental requirement. In order to test the ability of 2-H to form olefins when reacted with carbonyls, its reactivity was tested with 2 equivalents of base and benzaldehyde, as shown in Scheme 3.6.

Scheme 3.6 Reaction of 2-H with benzaldehyde forming the desired olefin 5 together with the side-product diphenyl methane 6.

As expected, the reaction of 2-H with benzaldehyde forms the trisubstituted alkene 5, albeit in only 11% isolated yield. As shown in Scheme 3.6, the synthesis of olefin 5 is accompanied by concomitant formation of the undesired side-product 6. Diphenyl methane 6 is unfortunately the main product of the reaction, indicating low stability of doubly lithiated 2-H. Cleavage of the P-C bond in similar compounds has already been shown,[135] which could explain the formation of the good leaving group 6. Because of the low isolated yield of desired alkene 5, and the low stability of the above described system, no further experiments were performed, whilst a new synthetic scheme was designed (described in chapter 4).

41 3.4.4 Conclusions and outlook Considering λ5σ4 P-Mes-phosphine oxide 2-H as a potential intermediate in a new reductive cross-coupling of deactivated aldehydes and ketones (see Scheme 3.3), its reactivity in basic environment was tested (see Scheme 3.5). Unfortunately, when 2-H is reacted with 1 equivalent of base it reacts only from the phosphorus center, and no reactivity at the α-benzyl carbon atom is observed. Only upon treatment of 2-H with 2 equivalents of base, reactivity from α-benzyl carbon is encountered (see Figure 3.3). Ultimately, doubly lithiated 2-H can react with benzaldehyde forming the desired alkene product 5, but in very low yield. Cleavage of the P-C bond seems to be the main reaction pathway instead, and the stable side-product 6 is formed as the major product of the reaction (see Scheme 3.6). As a result of the reactivity study performed and described above, λ5σ4 P- Mes-phosphine oxide 2-H doesn’t seem to be an appropriate intermediate for a possible new cross-coupling reaction. In the first published protocol,[133] the phosphine oxide intermediate was protected by the sterically bulky Mes* group. Supposedly because of the steric hindrance offered by this protecting group, reactivity of λ5σ4 Mes*-phosphine oxides takes place at the desired α- benzyl carbon and not the P-H moiety instead, leading to the formation of olefins in the reaction with different aldehydes (see Scheme 3.2). Considering these findings, a new reaction pathway which avoids the formation of 2-H was planned, the results of which are described in the next chapter.

42 4. Development of a new reductive cross- coupling of deactivated aldehydes (Paper II)

This chapter deals with the development of a new and broader cross-coupling olefination reaction of electron-rich aldehydes. A more detailed look at the phospha-Peterson reaction to form phosphaalkenes and their chemical behavior as electrophiles is given here. Optimization of P-Mes-phosphaalkene synthesis led to the creation of a new procedure in which even electron-rich aldehydes can be employed for the preparation of elaborated alkenes. In this context, important biologically active stilbenes have been synthesized.

4.1 Modification of the previous method As described in chapter 3, P-Mes-based phosphine oxides cannot be employed in new olefination reactions. Therefore, a different strategy needed to be adopted. In regard to this, it is clear that blocking the reactive P-H bond in 2- H is crucial to have further reactivity with carbonyl groups. To address this, activation of phosphaalkenes intermediates with alcohols (ROH), to avoid the formation of unreactive P-Mes-phosphine oxides, was proposed (Scheme 4.1).

Previous conditions : activation with OH

1 1 O 1 R "OH" HO R tautomerization R P P Mes P 1 1 Mes R Mes R1 H R

Unreactive towards carbonyls

New planned conditions : activation with ROH

1 1 1 R ROH RO R [O] O R P P H Mes P H 1 1 1 Mes R Mes R OR R

taut. Reactive towards carbonyls ?

O R1 Mes P 1 H R Scheme 4.1 Comparison of previous (top) and planned conditions (bottom).

43 Activation of phosphaalkene intermediates with nucleophilic bases (previously employed conditions, upper section in Scheme 4.1) could lead to phosphinite compounds that are able to tautomerize[134] to their corresponding unreactive P-Mes-phosphine oxides. Phosphaalkenes are known to behave as electrophiles towards many protic reagents.[135] Among these, alcohols would be interesting for the purpose of this work. As shown in the lower section of Scheme 4.1, activation of P-Mes-phosphaalkenes with ROH would lead to phosphinite intermediates that are unable to tautomerize to undesired unreactive P-Mes-phosphine oxides. Oxidation of the latter could form P-Mes-phosphinates that bear resemblance with typical HWE olefinating reagents and might thus be reactive towards carbonyl groups. In this scenario, a new one-pot methodology was planned based on a different activation of λ2σ3 P-Mes-phosphaalkenes to λ5σ4 P-Mes-phosphinate intermediates (Scheme 4.2).

Scheme 4.2 Proposed procedure based on addition of ROH to P-Mes-phosphaalkene intermediates. The reaction can be divided mechanistically in three steps: phosphaalkene formation, phosphaalkene activation and alkene formation.

The above proposed reaction scheme represents a valid synthetic pathway to be investigated in detail. The three separate steps have all been known for several decades, but were never combined into a one-pot olefination reaction. Scheme 4.3 represents the three known steps the new one-pot olefination reaction is based upon.

Scheme 4.3 Representation of the three isolated and known steps. “i.y.” indicates the isolated yield of the products of each reaction.

Firstly, the base-catalyzed phospha-Peterson reaction was used by Gates et al. to synthesize MesP=CPh2 with an isolated yield of 72% (reaction 1 in Scheme 4.3).[91b] The compound has been known for many years and is stable as a Mes- trisubstituted phosphaalkene (see section 3.3).[83] Its reaction with EtOH has also been described by the work of Van der Knaap and forms, in 85% isolated yield, its corresponding phosphinite which can be easily oxidized to its P(V) congener in almost quantitative yield (reactions 2 in Scheme 4.3).[111] Lastly, P-Ph-phosphinates have been reported by Horner to react in a basic environment with several aldehydes affording alkenes in 26-85% isolated yield (reaction 3 in Scheme 4.3).[137] The next three sections are dedicated to the presentation of the above described reactions. The combination of these three steps for shaping a new one- pot olefination is then presented in the last sections of this chapter.

4.2 The phospha-Peterson reaction As already described in section 2.2.3, the phospha-Peterson reaction is one of the most widely used methods in the synthesis of phosphaalkenes. The first report in 1976 by Becker was a landmark,[1] and many variations have been reported to-date. In all the possible modifications, the crucial step is a 1,3-silyl shift after an initial nucleophilic attack at a carbonyl group (Scheme 4.4).[3a, 91a, 91d]

45 O Me Me Me Me 1 Si Me R1 H Me Si OX 1,3-silyl shift R PR P R1 P X Nucleophilic attack R H R H

Me3Si-O-X

phospha-Peterson reagent: R = protecting group (Mes*, Mes, Ph, Ad, etc.) X = SiMe3 or Li Scheme 4.4 Mechanism of the phospha-Peterson reaction.[138]

The active phospha-Peterson reagents are di-silylated phosphines (RPTMS2, X = TMS in Scheme 4.4) or mono-silylated phosphines (RPLiTMS, X = Li in Scheme 4.4). These nucleophilic species attack carbonyl groups with formation of the σ-P-C bond. Because of the oxophilicity of silicon, a 1,3-silyl shift rapidly takes place with cleavage of the O-C bond and formation of the π-P=C double bond. In this step, the desired phosphaalkene product is formed along with the stable by-product of the reaction (TMS-O-TMS if X = TMS; TMS-O-Li when X = Li in Scheme 4.4). Concerning the reaction with lithiated phosphine species (RPLiTMS), two approaches have been developed in the last decades (Scheme 4.5).

From the primary phosphine

1) 1eq. n-BuLi O tBu Me tBu Me H 2) 1eq. tBuMe SiCl 1eq. n-BuLi 1 2 Si Si R1 H R PR Me Me PR PR P H (- tBuMe SiOLi) (-nBuH and LiCl) H (-nBuH) Li 2 R H

From the di-silylated phosphine

O Me Me TMS 1eq. MeLi 1 R1 H R PR Si Me PR P TMS (-MeTMS) (- TMSOLi) Li R H

Quenching of the by-products

1eq TMSCl R3Si-OLi R3Si-O-TMS LiCl Scheme 4.5 Different approaches for the synthesis of phosphaalkenes from MesPLiTMS by means of the phospha-Peterson reaction (R = protecting group, Mes*, Mes, Ph, Ad, etc.). Top: starting from the primary phosphine. Middle: starting from the di-silylated phosphine. Bottom: quenching of the lithiated byproduct of the reactions.

In the first method (upper section in Scheme 4.5), a long synthetic procedure begins from primary phosphines (RPH2). A first lithiation followed by t t- quenching with BuMe2SiCl generates a mono-silylated phosphine (RPHSi BuMe2) which, upon treatment with a second equivalent of n-BuLi, finally t creates the desired lithiated phospha-Peterson reagent (RPLiSi BuMe2). This

46 active species then reacts with a carbonyl group to form phosphaalkene and t [89c, 92a, 108, 139] BuMe2SiOLi products. In an easier and more recent method (middle section in Scheme 4.5), di- silylated phosphines (RPTMS2) are refluxed with 1 equivalent of MeLi in THF to directly afford the lithiated phospha-Peterson reagent that can then generate the desired products.[94, 113] In both methods, 1 equivalent of oxo-silylated compound (R3SiOLi) is formed at the end of the reaction. In most cases, quenching of this species is performed by adding 1 equivalent of TMSCl to the reaction mixture (bottom section in Scheme 4.5).[94a, 113] Formation of a second strong Si-O bond creates a volatile di-silylated ether (R3Si-O-TMS) that is easily removed from the reaction mixture. Regarding the phospha-Peterson reaction performed from di-silylated phosphines (RPTMS2), the group of Gates recently reported two different approaches which are presented in Scheme 4.6.[91b, 93]

Base-catalyzed phospha-Peterson reaction

TMS 1 O KOHcat R Ar P P 1 1 TMS R R (- TMSOTMS) Ar R1

Lewis acid-mediated phospha-Peterson reaction

TMS 1 O AlCl3 R PAlkyl P 1 1 TMS R R /H (- TMSCl) Alkyl R1/H -(Cl2AlOTMS) Scheme 4.6 Base catalyzed (top) and Lewis acid-mediated (bottom) phospha-Peter- son reactions.

A first protocol is base-catalyzed (upper part in Scheme 4.6) and has been useful in the synthesis of P-Mes-trisubstituted stable phosphaalkenes.[91b] In this approach, the only byproduct (hexamethyldisiloxane, TMS-O-TMS) is volatile and easy to remove. In a second method (lower part in Scheme 4.6), alkyl-protected di-silylated [93] phosphines are reacted with carbonyl groups in the presence of AlCl3. The reaction forms in one-pot the phosphaalkene product along with the corresponding silylated by-products. This protocol has been successful in the synthesis of P-alkyl-substituted-phosphaalkenes (e.g. Alkyl = tBu or Ad).[93] In conclusion, the phospha-Peterson reaction represents a powerful tool for the synthesis of phosphaalkenes. It has been shown to work not only with aldehydes[139a] and ketones,[91b] but also with several different carbonyl groups [1, 89c] [138] [140] [95a] like acyl chlorides, amides, isocyanide dichlorides, CO2, iso- cyanate [77b] and ketenes.[77b, 90]

47 4.3 Phosphaalkenes as electrophiles As described in sections 2.2 and 2.2.4, due to the difference in electronegativity between phosphorus and carbon, phosphaalkenes possess a polarized P=C double bond (Pδ+Cδ-).[3a, 3b, 110] The polarity of the P=C unit can be tuned according to the substituents on the carbon atom. In “classically” substituted phosphaalkenes, the phosphorus atom has electrophilic character and can be attacked by several nucleophiles (see Scheme 2.20 and Scheme 4.7 ).[2] In C- amino substituted phosphaalkenes, the polarity of the P=C double bond is reversed giving rise to the so called “inversely” polarized phosphaalkenes (Pδ- Cδ+, Figure 4.1).[91c]

Figure 4.1 Resonance structures of inversely polarized phosphaalkenes.

Both systems can react with several protic reagents (e.g. hydrogen halides, alcohols, amines or thiols)[91a, 135] but with opposite regioselectivity (see Scheme 2.20 and Scheme 4.7). In the case of “classical” phosphaalkenes, H+X- adds into the P=C double bond giving a product in which the nucleophilic part (X) is attached to the phosphorus atom and the electrophilic portion (H) to the carbon atom (upper part in Scheme 4.7). Conversely, in “inversely” polarized phosphaalkenes, the electrophilic portion (H) will end up on the phosphorus atom and the nucleophilic part (X) on the carbon atom of the P=C double bond (bottom part in Scheme 4.7).

+ - Scheme 4.7 1,2 addition of a general protic reagent (H X : HCl, ROH, HNR2, RSH) to classically (top row) and inversely polarized (bottom part) phosphaalkenes.

An example of trapping phosphaalkenes with nucleophiles is given in Scheme 4.8 in which benzoic acid adds onto a P=C double bond.[105a]

48 O Ph (OC) W Ph O Ph 5 PhCO2H P P H (OC) W Ph H 5 Ph H Scheme 4.8 Trapping of phosphaalkenes with benzoic acid.[105a]

The peculiar polarity of classic phosphaalkenes can also be exploited in anionic polymerization reactions. The group of Gates developed procedures in which small alkyl-lithium bases (e.g. MeLi and n-BuLi) can add onto P=C double bonds and act as initiators for anionic polymerizations. After nucleophilic addition, the resulting alkylated phosphaalkene can attack another P=C unit and propagate the polymerization reaction (Scheme 4.9).[4a, 4b, 112]

Scheme 4.9 Examples of anionic polymerizations of a P-Mes-phosphaalkene.

4.4 HWE-type olefinating reagents As specified in section 2.1.2, in the late 50´s and early 60´s, Horner, Wadsworth, and Emmons published their first works in which Wittig-type reagents afforded olefins upon reaction with carbonyl groups. More specifically, in 1958, Horner reported on the reaction of phosphine oxides and carbonyl compounds.[41] This type of reaction is now referred as the Horner-Wittig olefination.[16, 43] Three years later, Wadsworth and Emmons published the formation of olefins from the reaction of phosphonate carbanions and carbonyls.[19a] Lastly, during the development of these reactions, Horner discovered that phosphinates can also act as olefinating reagents towards ketones and aldehydes.[137] A presentation of the three P-based olefinating reagents is given in Figure 4.2.

Figure 4.2 Phosphorus-based olefinating reagents.

49 Among the three different phosphorus-based olefinating reagents, phosphonates are the most widely used (HWE reaction). Phosphine oxides are normally less employed than corresponding phosphonates. However, the Horner- Wittig reaction has a unique feature. Phosphine oxides carbanions add into carbonyls generating “betaine-type” intermediates that can be isolated and purified. This provides a possibility for controlling the stereochemistry of the reaction, since each pure diastereomeric intermediate selectively generates ge- ometrically pure alkenes.[43] Lastly, phosphinates are the least used coupling reagents. However, for the purpose of this work, reactivity of phosphinates is of interest because these compounds could be intermediates in the newly proposed coupling reaction (see section 4.1). Horner investigated the best reaction conditions for couplings of phosphinates and carbonyls. It was found that tBuOK promotes olefin synthesis upon reflux in toluene.[137] Thus, phosphinates can be interesting intermediates in the planned procedure shown in Scheme 4.2. The next three sections are dedicated to the optimization of each individual part of Scheme 4.2.

4.5 Optimization of the phospha-Peterson reaction The first step of the planned procedure shown in Scheme 4.2 concerns the synthesis of P-Mes-substituted phosphaalkenes by means of the phospha-Pe- terson reaction. This section deals with the optimization of such a reaction in order to produce P-Mes-phosphaalkenes in high yield. It is important to mention that P-Mes-phosphaalkenes were not isolated in this part of the work, but the quality of their synthesis was scrutinized by 31P NMR spectroscopy of the various reaction mixtures. Several approaches according to the general reaction Scheme 4.10 were explored.

Scheme 4.10 General phospha-Peterson formation of phosphaalkene 7. X = H, TMS.

The obtained results are presented in Table 4.1.

Table 4.1. Set of different approaches and corresponding results. Entry Starting material Approach Result

1 MesPH2 nBuLi, TMSCl Various side-products 2 MesPTMS2 KOH Various side-products 3 MesPTMS2 AlCl3 Various side-products 4 MesPTMS2 MeLi Unwanted polymerization of P=C 5 MesPTMS2 LiOEt Clean

At first, reactions from MesPH2 were tested (entry 1, Table 4.1). According to the sequence proposed in Scheme 4.5, lithiation of MesPH2 with one equivalent of n-BuLi and quenching with 1 equivalent of TMSCl generated in-situ MesPHTMS. Upon a second equivalent of n-BuLi, MesPLiTMS was obtained and it was reacted with benzophenone. MesP=CPh2 7 was generated, along with too many side-products. The tested sequence is not optimal for the purpose of the one-pot coupling, because four steps were needed for the formation of desired phosphaalkene 7. Subsequently, shorter protocols starting from MesPTMS2 were investigated. In entry 2, the base-catalyzed phospha-Peter- son reaction was tested. According to Scheme 4.6, MesPTMS2 was mixed with KOH and benzophenone. Target phosphaalkene 7 was formed, but with poor selectivity. The same results were obtained with the reaction of MesPTMS2 and AlCl3 (Table 4.1, entry 3). In entry 4, it was decided to generate MesPLiTMS intermediate directly from MesPTMS2, in order to have a shorter reaction sequence compare to the one used in Table 4.1 entry 1. In-situ formed MesPLiTMS, obtained by refluxing MesPTMS2 with MeLi in THF, was reacted with benzophenone, and phosphaalkene 7 was generated as the only detected species. However, broadening of the baseline in the 31P spectrum was noticed, most probably due to the use of MeLi. Indeed, small alkyl lithium- bases are known to catalyze polymerization reactions of P-Mes-phosphaalkenes (see Scheme 4.9).[4b, 112, 141] It is thus possible that unreacted MeLi might have attacked 7 giving rise to phosphorus-containing side-products that were not detectable via 31P NMR spectroscopy. Looking for a reagent capable of generating MesPLiTMS in high yield that is also unreactive towards P=C double bonds, alkoxy bases (e.g. tBuOK, LiOEt, etc.) seemed to be an ideal choice. It is known that such species can abstract a TMS moiety from dis- ilylated phosphines and generate the corresponding metallated phosphines in quantitative yield.[142] In this reaction, the driving force is the formation of the strong silicon-oxygen bond (upper part in Scheme 4.11). When using LiOEt, the byproduct of the reaction (TMS-O-Et) is a stable gas, which can be easily removed from the reaction mixture, leaving pure MesPLiTMS in solution.

Scheme 4.11 Top: formation of MesPLiTMS with LiOEt with generation of TMS-O- Et by-product. Bottom: synthesis of 7 from previously formed MesPLiTMS.

Generation of MesPLiTMS was monitored by 31P NMR spectroscopy. After four hours of stirring at room temperature, MesPLiTMS was the only detected compound. Removal of the volatiles (THF and TMS-O-Et) afforded a yellow solid. Initial attempts for the synthesis of 7 were performed from THF solutions of MesPLiTMS. However, it was found that ethereal solutions of the Li salt afforded cleaner P-Mes-phosphaalkenes 7. Therefore, solid MesPLiTMS was solubilized in Et2O. Addition of benzophenone resulted in pure formation of 7, as shown by its corresponding 31P NMR spectrum (Figure 4.3).

Figure 4.3 31P NMR spectrum of the reaction presented in Scheme 4.11.

Having developed a reliable synthesis of phosphaalkene 7, its quenching with alcohols and subsequent oxidation (second part of Scheme 4.2) to form desired P-Mes-olefinating reagents was studied. The results of this study are presented in the following section.

52 4.6 Bridging low- and high-valent phosphorus chemistry As presented in section 4.3, phosphaalkenes can be attacked by several protic reagents. Important to our work is the activation of P=C double bonds with alcohols (see Scheme 4.1 and Scheme 4.2). Having developed an optimized synthesis of P-Mes-phosphaalkenes, it was possible to study their reactivity with alcohols (ROH). Reactions were carefully monitored by 31P NMR spectroscopy and the corresponding products were not isolated. A general reaction scheme for the quenching of 7 with different ROH is shown in Scheme 4.12.

Ph ROH RO Ph P P Mes Ph Et2O, Mes Ph r.t., few min 7 Scheme 4.12 Reaction of phosphaalkene 7 with a generic alcohol (ROH) leading to the formation of the corresponding phosphinite specie.

Several reactions were tested, and the obtained results are summarized in Ta- ble 4.2.

Table 4.2 Set of different tested reagents and corresponding results. Conditions as in Scheme 4.12. Reagents Results NaOEt/EtOH exc. Various side-products NaOMe/MeOH exc. Various side-products LiOEt 1eq. No reaction MeOH 1eq. Clean

Initially, a commercially available solution of NaOEt in EtOH was used in the reaction with 7. Conversion of 7 to the corresponding phosphinite is achieved within a few minutes at room temperature. Unfortunately, formation of the product was not clean and many side-products were detected by 31P NMR analysis of the reaction mixture. A similar result was obtained using a Na- OMe/MeOH solution. Considering that MesPLiTMS was generated with Li- OEt (see Scheme 4.11), a second equivalent of LiOEt was added to the previously formed phosphaalkene 7. In this case, no reaction was detected and 7 was left untouched in the reaction mixture. Finally, pure MeOH was tested. Fortunately, upon addition of 1eq. of MeOH, phosphaalkene 7 is fully consumed within a few minutes, and a clean formation of the corresponding phosphinite is achieved, as visible from the 31P NMR spectrum of the reaction mixture (Figure 4.4).

Figure 4.4 31P NMR spectrum of the reaction between 7 and MeOH. Conditions as in Scheme 4.12, R = Me.

The next step that is needed to bridge low-coordinate with high-valent phosphorus chemistry is the oxidation of phosphinite intermediates to the corresponding phosphinates (see section 4.1). The general reaction pathway for the oxidation of P(III) intermediates to P(V) species is reported in Scheme 4.13.

Scheme 4.13 General oxidation of phosphinite 8 to phosphinate 9. [O] is a generic oxidating agent.

Several oxidizing agents have been tested, and the results are collected in Ta- ble 4.3.

Table 4.3 Set of different tested oxidizing reagents and corresponding results. Con- ditions as in Scheme 4.13. Reagents Results

aq. H2O2 exc. Various side-products DABCO-2H2O2 1 eq. No reaction Urea-H2O2 1 eq. Poor conversion Air Various side-products dry (Me)3NO 1 eq. No reaction dry tBuOOH 1 eq. Clean reaction

Firstly, aqueous hydrogen peroxide was added to an ethereal solution of 8. 31P NMR spectroscopy analysis of the reaction mixture revealed a complex reaction, since many undefined side-products were also formed. Apparently, aqueous H2O2 is a too strong oxidizing agent. Instead, a DABCO-2H2O2 complex proved to be unreactive towards intermediate 8. A new complex (urea-H2O2)

54 was also tested. Some reactivity was observed, although with a poor conversion to desired product 9. Only upon reaction with several equivalents of oxidizing agent, oxidation of 8 is achieved in a reasonable time. Bubbling air in the reaction mixture resulted in a fast oxidation of 8, but many side-products were also observed. An opposite result was obtained by treating 8 with 1eq. of dry (Me)3NO, since no reaction occurred with this oxidizing reagent. Ulti- mately, a dry solution of tBuOOH in benzene was tested. Oxidation of 8 to the desired product 9 was finally achieved within few minutes, as visible from its corresponding 31P NMR spectrum (Figure 4.5).

O Ph Mes P OMePh 9

Figure 4.5 31P NMR spectrum of the reaction between 8 and 1eq. of tBuOOH. Con- ditions as in Scheme 4.13.

As evidenced by these findings, phosphaalkene 7 can be converted to its phosphinate intermediate 9 with two fast and clean reactions (Scheme 4.12 and Scheme 4.13). Thus, the creation of the new one-pot methodology proposed in Scheme 4.2 is almost completed, with only the last step (alkene formation) left to be studied and optimized (presented in the next section).

4.7 Study of the olefination step As described in section 4.4, phosphinates are able to react with carbonyl compounds to form alkene products via Horner-Wittig type reactions. In the procedure developed by Horner in the early 60s,[137] P-Ph-phosphinates were used for the construction of C=C double bonds. In this work, P-Mes-phosphinates are used as new coupling reagents. Thus, their chemistry needed to be investigated. In order to screen different variables, P-Mes-phosphinate 11 was prepared in large scale according to Scheme 4.14.

Scheme 4.14 Synthesis of Mes-phosphinate 11.

With phosphinate 11 in hand, its reactivity towards aldehyde 12 was studied, as detailed in Scheme 4.15.

Scheme 4.15 General reaction scheme for the preparation of trisubstituted olefins from pure isolated 11.

Different combinations of reaction conditions (base, solvent, temperature and reaction time) were tested, and the obtained results are presented in Table 4.4.

Table 4.4 Reaction conditions tested for the olefination step reported in Scheme 4.15. Rec = recovered; r.t. = room temperature. Temp. Entry Base (1.5eq.) Solvent Time Result °C 1 KHMDS THF r.t. overnight 11 rec., 25% conversion 2 tBuOK THF r.t. overnight 11 rec., 35% conversion 3 tBuOK THF +40 °C 6 h 11 rec. and decomposition 4 tBuOK THF Reflux 6 h mainly decomposition 5 tBuOK DMF r.t. overnight 11 rec., 20% conversion 6 tBuOK DMF Reflux 2 h mainly decomposition 7 tBuOK Toluene Reflux overnight mainly decomposition

In the first entry, KHMDS was used as the base in THF at room temperature. Analysis of the reaction mixture after one night of stirring revealed poor conversion to the desired alkene 13. The main recovered material is unreacted 11, owing to a poor reactivity of this phosphinate in the tested reaction conditions. Conversion to product 13 was only 25%. A slightly better result was achieved in similar reaction conditions with tBuOK as the base (entry 2). In this case, 11 is still recovered but alkene 13 is formed with 35% conversion. In order to achieve an improved reactivity of 11, the same conditions were tested at a higher temperature (entry 3). Surprisingly, product 13 was not formed at all. The reaction contained instead a mixture of unreacted aldehyde 12, some non- reacted phosphinate 11, and another compound containing only aromatic pro-

56 tons. A more detailed analysis of this fraction revealed the formation of benzophenone. Having a closer look at the chemistry of phosphinates and phosphonates, it was found that in these compounds the P-C bond can be cleaved with formation of the corresponding carbonyl.[111, 143] Therefore, benzophenone can be formed by the degradation of reagent 11 which does not take part in the coupling with 12. In harsher reaction conditions (entry 4), it is not sur- prising that 11 is not recovered at all, with benzophenone being the main product of the reaction. Switching to a new solvent (DMF, entry 5) did not increase the formation of 13. After stirring at room temperature for one night, 11 was recovered and olefin 13 was synthesized with a poor conversion. In entry 6, the same conditions were tested, but the reaction was refluxed for two hours. With DMF having a considerably high boiling point, the main reaction was decomposition of 11 to benzophenone. The same result was obtained by refluxing 11 in toluene. In the two entries, no alkene 13 was formed. As it emerges from the analysis reported in Table 4.4, phosphinate 11 doesn’t seem an ideal coupling reagent for the construction of trisubstituted alkenes. Indeed, a fine tuning of its reactivity is needed. When reactions are performed at room temperature (entries 1, 2 and 5), some product is formed, albeit in a very low conversion. In these cases, poor reactivity is revealed, and 11 is recovered in the reaction mixture. At higher temperatures (entries 3, 4, 6 and 7), the main pathway is decomposition of 11 with no formation of the desired alkene. These results are a clear demonstration that trisubstituted olefins may not be accessible with this coupling reagent. Apparently, while normal HWE phosphonate reagents can be exploited for the construction of trisubstituted alkenes, P-Mes-phosphinates are in this regard unreactive. The reason may be due to the presence of the sterically demanding Mes group, which doesn’t allow a proper ketone-aldehyde cross-coupling. Thus, the reactivity of a less sterically hindered P-Mes-phosphinate was tested. In the first reported procedure from our group,[133] only electron-poor aldehydes bearing electron-withdrawing groups (EWG) could be employed for the synthesis of 1,2-disustituted alkenes (see section 3.2). Neutral and electron-rich aldehydes could not be used in this case. Decreasing the size of the protecting group from Mes* to Mes could allow the use of deactivated aldehydes for the construction of more elaborated C=C double bond containing products. Thus, a new coupling reagent was synthesized in a similar manner as for 11 (Scheme 4.16).

Scheme 4.16 Synthesis of 14 by means of the Michaelis-Arbuzov reaction.

57 Having pure 14, its reactivity towards para-bromobenzaldehyde 12 was tested, as shown in Scheme 4.17, and the obtained results are shown in Table 4.5.

Scheme 4.17 General reaction scheme for the synthesis of the 1,2-disubstituted olefin 15.

Table 4.5 Reaction conditions tested for the olefination step reported in Scheme 4.17. Rec = recovered; r.t. = room temperature. Entry Base (1.5eq.) Solvent Temp. °C Time Result 1 tBuOK THF r.t. Overnight 70% conversion 2 tBuOK DMF r.t. 3 h 14 rec., 50% conversion 3 nBuLi THF r.t. 3 h 14 rec., 60% conversion

In the first entry, tBuOK was used as the base and the reaction mixture was stirred in THF at room temperature for 16 hours. Finally, high yielding conversion to product 15 was achieved. In the reaction mixture, 14 was not recovered. Changing solvent to DMF (entry 2 in Table 4.5) revealed a lower reactivity of 14. Some starting material was recovered, and conversion to the desired alkene was lower than in entry 1. Lastly, n-BuLi was used as the base (entry 3 in Table 4.5). A lower reactivity than in entry 1 was observed, with some 14 being recovered from the reaction mixture. No more tests were performed. In conclusion, it was shown that phosphinate 14 is an ideal olefination reagent for the synthesis of 1,2-disubstituted olefins. Conditions tested in Table 4.5 entry 1 were selected to be employed for the new one-pot coupling of deactivated aldehydes. The three separate steps were all individually optimized, and their combination gave rise to the development of a new one-pot cross- coupling of deactivated aldehydes to alkenes, which is described in the next section.

4.8 Development of the new one-pot procedure Having developed each individual reaction detailed in Scheme 4.2, only the combination of the three isolated steps remained to be tested. The new one- pot cross-coupling of deactivated aldehydes to alkenes is reported in Scheme 4.18.[144]

Scheme 4.18 Detailed reaction scheme for the new one-pot cross-coupling of two different aldehydes to unsymmetrical alkenes.[144]

MesPLiTMS I reacts with several aldehydes at room temperature in a few minutes to generate its corresponding P-Mes-phosphaalkene II. Addition of MeOH instantly converts phosphaalkene II into its corresponding phosphinite intermediate III. Oxidation of III to phosphinate IV is achieved in a few minutes by using tBuOOH. Lastly, addition of tBuOK and a second aldehyde leads to the desired unsymmetrical alkene and the phosphonate V by-product. Each single step of the new one-pot procedure can be conveniently monitored by 31P NMR spectroscopy (see Figure 4.6). Addition of one equivalent of R1CHO to an ethereal solution of MesPLiTMS I (δ: -187 ppm) yields two new peaks with a chemical shift around +250 ppm. These are the characteristic deshielded signals of E- and Z-P-Mes-phosphaalkenes. As visible from the 31P NMR spectrum shown in Figure 4.6 b, the optimized phospha-Peterson reaction allows the synthesis of disubstituted P-Mes-phosphaalkenes in high selectivity. Upon MeOH addition, both E- and Z-phosphaalkenes II are converted into III (δ: + 129 ppm) which is oxidized to phosphinate IV (δ: +46 ppm) by using tBuOOH. Addition of tBuOK and a second aldehyde (R2CHO) yields the desired alkene product (not visible by 31P NMR spectroscopy) and the phosphonate by-product V (δ: +16 ppm).

Figure 4.6 31P NMR spectroscopy investigation of the steps reported in Scheme [144] 4.18. Each reaction was checked with an internal C6D6 capillary. (a → b) phospha-Peterson reaction between I and R1CHO to from E- and Z- phosphaalkenes II. (b → c) addition of MeOH to form phosphinite III. (c → d) oxidation of III to the corresponding phosphinate IV. (d → e) olefination step between IV and R2CHO with formation of phosphonate V by-product. Conditions as in Scheme 4.18.

With the optimized one-pot protocol demonstrated, the scope of the new method was investigated. The obtained results are detailed in the next section.

4.9 Substrate scope of the new method Having established the mechanism of the new cross-coupling procedure, the effect of the reduced steric demand of the phosphorus protecting group towards the substrate scope of the newly developed protocol was investigated. It is important to remember that, as discussed in section 3.2, the previous procedure did not allow the use of any electron-rich aldehydes or ketones.[133] A series of differently substituted benzaldehydes can be used in the new coupling procedure to form various unsymmetrical stilbenes (Table 4.6).

Table 4.6 Symmetrical and unsymmetrical alkenes from the homo- and cross-coupling of two aldehydes, respectively. Conditions as in Scheme 4.18. a Yields are calculated relative to the second aldehyde which was used as a limiting reagent (0.5 eq. to the first aldehyde). Yield %, Entry 1st aldehyde 2nd aldehyde Product E/Z ratio O 1 58, Br 100/0

60 O 45, 2 100/0 Br

O MeO MeO OMe 47, 3 100/0 OMe MeO O MeO 51, 4 100/0 OMe

O O MeO OMe 33, 5 100/0 OMe OMe O MeO 35, 6 98/2 OMe

52, 7 75/25

45, 8 85/15

O O 56, 9 57/43 N Br O

40%, 10 50/50

O a 79, 11 98/2 Br

O a 75, 12 100/0 Br

a 80, 13 37/63

Initially, electron-poor substrates like para-bromobenzaldehyde and 3,5-dimethoxybenzaldehyde were used as the first coupling partners for the genera-

61 tion of the corresponding P-Mes-phosphaalkenes. Conversion of these intermediates and subsequent coupling with a second carbonyl proceeds with complete control of stereoselectivity with the E-alkenes being the only isolated products (Table 4.6, entries 1 to 6). Thus, para-bromobenzaldehyde can be coupled to itself, generating the corresponding symmetrical stilbene in 58% isolated yield (Table 4.6, entry 1), or to unsubstituted benzaldehyde affording the cross-coupling product in 45% yield (Table 4.6, entry 2). Similar satisfactory results are obtained using 3,5-dimethoxybenzaldehyde as the first coupling partner. Strongly electron-rich aldehydes, that were not compatible with the previous protocol,[133] can be used as the second carbonyl, leading to pre- cursors of important biologically active stilbenes (Table 4.6, entries 3 and 4).[8, 145] Resveratrol[10] is derived from the olefin synthesized in entry 3,[146] whereas acidic cleavage of the OTHP group in alkene from entry 4 affords pterostilbene (see also Figure 1.1).[11, 147] The substrate scope of the new protocol is not limited to electron-deficient aldehydes in the first step. Indeed, the new method allows for the use of neutral (Table 4.6, entries 7 and 8) and electron-rich benzaldehydes (Table 4.6, entries 9 and 10) as the first coupling partners. Such substrates could not be used in the first reported procedure.[133] An interesting trend in the stereoselectivity of the new protocol emerges when non electron-poor aldehydes are used in the first step of the coupling, owing to a correlation between the E-Z selectivity of the method and the electronic nature of the first carbonyl. By comparing entries 1-10 in Table 4.6, it is clear that the reaction becomes increasingly less selective for the E-isomer. With unsubstituted benzaldehyde as the first reagent, a roughly 3:1 ratio of E/Z stilbene products is observed (Table 4.6, entries 7 and 8). When strong electron-rich substrates are used in the first step, the Z-isomer is obtained in almost equal amounts as the E-isomer (Table 4.6, entries 9 and 10). No such correlation is noticed for the second carbonyl, the nature of which seems to be less important in determining the stereoselectivity of the new procedure. The yields reported in entries 1-10 in Table 4.6 are relative to the aldehydes, which were used in equimolar amounts. An economically valuable substrate can be used as the limiting reagent in the second step of the procedure. For instance, using a second carbonyl in a 1:2 relative ratio to the first aldehyde, increases dramatically the isolated yields up to 80% (Table 4.6, entries 11-13). It is important to mention that in all the reactions presented in Table 4.6, no homo-coupled side-product was ever observed. Having achieved the coupling of two aldehydes, the synthesis of trisubstituted alkenes was also tested. A great advantage of this new method, compared to the McMurry reaction, is that its mechanism is based on a sequence of different steps. For instance, trisubstituted olefins could be synthesized with two different combinations of carbonyls: aldehyde + ketone or vice-versa. In section 4.7, it was shown that P-Mes-phosphinates deriving from ketones are almost unreactive in the last olefination step. Thus, the combination ketone (1st

62 carbonyl) + aldehyde (2nd carbonyl) cannot be performed with this method. However, the high reactivity of P-Mes-phosphinates synthesized from aldehydes could be exploited in the coupling with ketones as the second carbonyls. Thus, the reaction of para-bromobenzaldehyde and some aliphatic ketones was tested, and the results are shown in Table 4.7.

Table 4.7 Intermolecular cross-coupling of aldehydes and aliphatic ketones yielding trisubstituted olefins. Yield %, Entry 1st aldehyde 2nd carbonyl Product E/Z ratio

20, 1 -

24, 2 -

29, 3 75/25

Applying the room temperature protocol depicted in Scheme 4.18 and used for the reactions of Table 4.6 did not afford any product in the coupling with ketones. However, if the reaction between phosphinates and ketones is performed in refluxing THF, trisubstituted alkenes are indeed formed (Table 4.7). While the isolated yields are modest, it is important to realize that ketones and aldehydes have been reductively coupled to olefins under an ionic mechanism for the first time.

4.10 Advantages and limitations Compared to the McMurry reaction, the newly developed method has the following advantages: The new procedure omits the use of transition-metals in any step. In the McMurry reaction, tedious in-situ preparation of low-valent Ti0 active species is required combining TiCl3 or TiCl4 with strong reducing agents like Zn or Zn-Cu. Due to the difficulty of preparing the starting materials, the McMurry reaction is often described as “tricky” and has some problems in affording [16] reproducible results. Instead, MesPTMS2 can be synthesized on a multi- gram scale. Active MesPLiTMS can be stored under inert conditions for at least one month without visible degradation. Moreover, all transformations in our coupling are performed under mild reaction conditions, while the

63 McMurry coupling requires refluxing high-boiling solvents for extended pe- riods of time. Due to the harsh reaction conditions typically employed in the McMurry reaction, along with the use of strong reducing agents, some functional groups are not tolerated in the McMurry coupling. For instance, the NO2 group adopted in Table 4.6 entry 6 is not compatible with typical McMurry reaction conditions. Lastly, but most importantly, due to the ionic mechanism of our procedure, the synthesis of unsymmetrical olefins from the cross-coupling of two different carbonyls can be easily achieved. This is a vast improve- ment compared to the McMurry reaction. Indeed, due to its radical mechanism, the McMurry coupling leads to mixtures of symmetrical and unsymmetrical olefins when two different carbonyls are reacted (see Scheme 2.3 and section 3.1). Compared to phosphorus based- and free-olefination reactions described respectively in sections 2.1.2 and 2.1.3, this new methodology represents a one-pot procedure that does not require any purification of the corresponding olefinating reagents. Typically, two or three steps are required for the synthesis of phosphonium ylides and phosphonate coupling reagents. This interesting feature of the newly developed procedure saves time and materials, and olefins can be synthesized in a one-pot manner from feedstock carbonyls. Ad- ditionally, in a retrosynthetical analysis, all of the previously described olefinations do not represent the coupling of two carbonyls. This difference with our methodology offers an interesting and distinct opportunity for the construction of C=C double bonds. In comparison to the first method developed in our group (see section 3.2),[133] this new procedure offers a considerably broader substrate scope. In- deed, in the protocol described in section 3.2, only electron-poor aldehydes could be utilized for the synthesis of alkenes. Decreasing the size of the Mes* protecting group to the smaller Mes opened up the possibility of using not only EWG substituted aldehydes, but also carbonyls bearing electron-donating groups (EDG). With this new broader reaction scope, important biologically active stilbenes have been synthesized. Additionally, in the first reported procedure, two phosphorus atoms were consumed for each olefin that was synthesized. Instead, in this new protocol, only one phosphorus atom is needed to synthesize one olefin, leading to a better atom economy of the entire method. However, like every reaction in organic chemistry, our methodology also has some drawbacks that should be mentioned. Regarding the coupling of two aldehydes, the new procedure is limited to aromatic compounds. Aliphatic aldehydes do not take part in the reaction with MesPLiTMS, thus crucial P- Mes-phosphaalkene intermediates cannot be synthesized. When employed in the second step, considerably low formation of the desired olefin was detected. Concerning the coupling with ketones, very low yields were observed. Ke- tones cannot be used in the first step since the corresponding P-Mes-phosphinate shows poor reactivity in the last olefination step (see section 4.7). If

64 ketones are utilized in the second part, considerably low yields of the final trisubstituted olefins are observed (see results in Table 4.7).

4.11 Conclusions The present one-pot procedure provides a new entry into carbonyl-carbonyl cross-coupling reactions to form dissimilarly substituted alkenes. The newly developed method offers vast improvements in comparison to the McMurry reaction and to the previously published procedure from our group. Mild reaction conditions, wide substrate scope, and an ionic mechanism are all nota- ble features of this method. It is important to remember that the work described in this chapter does not only represent advantageous findings regarding new C=C bond formation reactions, but also shows great achievements in the field of low-coordinate phosphorus chemistry. Prior to this work, disubstituted phosphaalkenes bearing smaller protecting groups like Mes were never synthesized. Considering the impact of the phospha-Peterson reaction for the synthesis of P=C double bond containing molecules (see section 2.2.3), our optimized procedure represents new possibilities in the field of low-coordinate organophosphorus chemistry.

65 5. E-Z selectivity of the new cross-coupling procedure (Paper III)

This chapter gives more insights into the E-Z selectivity of the previously described reductive cross-coupling of two aldehydes into 1,2-disubstituted alkenes. It is shown that, depending only on the order of addition of the two coupling partners, the same stilbene can be afforded in a E- or Z-enriched form.

5.1 Introduction The stereochemistry of carbon-carbon double bonds has a crucial impact on the properties of alkenes.[148] Thus, establishing protocols to selectively access either the E- or the Z-isomer is still at the heart of organic chemistry. In classic carbonyl olefinations described in sections 2.1.2 and 2.1.3, the factors that dictate the stereoselectivity are well understood and depend upon different factors, such as the nature of the olefinating reagent, the base, the solvent, etc.[48, 50c, 149] Among these, it should be pointed out that there is only one example in which the nature of the carbonyl affects the stereochemical outcome of the reaction. This has been observed in Wittig olefinations of benzaldehydes bearing heteroatoms in the ortho-position. This phenomenon is described as the “ortho-effect”,[31a, 150] and it has been used to synthesize Z-enriched stilbenes.[151] As shown in Scheme 5.1, Z-alkenes are generated in the reaction with ortho-substituted benzaldehydes (Scheme 5.1a). In contrast, when the ortho-substitution is on the ylide, E-alkenes are obtained with unsubstituted benzaldehyde (Scheme 5.1b).

Scheme 5.1 Z- and E-stereoselective Wittig reactions with ortho-substituted benzaldehydes and phosphonium ylides. X = OMe, Br.[150c]

66 From a mechanistic point of view, this effect arises from a secondary bonding interaction between the phosphorus and the ortho-substituent in the transition state (TS), as shown in Scheme 5.2.[150c]

Scheme 5.2 Z-stereoselective Wittig reaction with an ortho-substituted benzaldehyde.[150c] Upper part: cis-OPA TS, leading to the Z-olefin, the major product. Lower part: trans-OPA TS with destabilizing 2,3-interactions leading to the minor E-stilbene. X = Cl.

As shown in Scheme 5.2, the cis-OPA TS is less crowded than the corresponding trans-OPA TS, which translates to a lower energy for the cis-OPA TS compare to the trans-OPA. Indeed, the puckering of the TS generates unfa- vorable 2,3-interaction in the trans-OPA, which are not present in the cis- OPA. Thus, the Z-alkene is the major product formed in the reaction. Having developed a new cross-coupling olefination of two benzaldehydes into 1,2-disubstituted alkenes,[144] curiosity arose to see whether such “ortho- effect” is present in the methodology described in chapter 4 as well. A systematic study that correlates the substitutions of benzaldehydes and the stereochemical outcome of the reaction is presented in the next section.

5.2 Synthesis of E- and Z- alkenes; results and discussion While developing the synthetic protocol described in chapter 4, the existence of two separate effects that influence the E-Z selectivity was noticed. A first effect takes places when the aldehydic coupling partners are substituted in the

67 para-position; a second arises instead with ortho-substituted benzaldehydes. The two effects will be presented separately. At first, a series of reactions with different para-substituted benzaldehydes were investigated, and the collected results are shown in Table 5.1.

Table 5.1 E- and Z-alkenes from the coupling of two para-substituted benzaldehydes. X = substitution on the first aldehyde A; Y = substitution on the second aldehyde B.

Entry E/Z ratio Aldehyde A Aldehyde B Yield %

1 H Br 46 70/30 2 H Me 52 75/25 3 H OMe 75 87/13 4 Br H 45 100/0 5 Me H 39 52/48 6 OMe H 57 38/62 7 Br Br 58 100/0 8 H H 52 75/25 9 Me Me 44 69/31 10 OMe OMe 40 50/50 11 Br Me 32 100/0 12 Me Br 74 49/51 13 Br OMe 35 100/0 14 OMe Br 48 30/70

The results presented in Table 5.1 are in accordance with the findings already described in section 4.9. The substitution on the second aldehyde (aldehyde B) has little influence on the stereochemical outcome of the reaction, and E- stilbenes are the main products regardless the Y group on aldehyde B (Table 5.1, entries 1-3). On the contrary, the electronic nature of the substituent on the first coupling partner (aldehyde A) has severe impacts on the outcome of the reaction (Table 5.1, entries 4-6). With EWG like Br (entry 4), the reaction is 100% E-selective. When a strong EDG such as OMe (entry 6) is used, the Z-olefin is obtained as major product. The same trend is confirmed in entries 7-10, which represent a series of homo-coupling reactions. The stronger the electron-donating X group on aldehyde A, the more Z-olefin is formed. Lastly, and more interestingly, entries 11-12 and 13-14 show how the same olefin can be prepared in either E- or Z-manner under identical reaction conditions. In entry 11, the olefin is synthesized in a 100% E-fashion due to the presence of Br substituent on aldehyde A, with the Me group employed on the second carbonyl having no influences on the stereochemical outcome of the process. By reversing the order of addition of the two aldehydes, as shown in entry 12, a 1:1 mixture of E-Z olefins is obtained. With the stronger EDG OMe used in

68 entries 13-14 a similar trend is observed, but with a greater selectivity for the Z- product when aldehyde A is electron-rich (entry 14). These last examples clearly show the strength of the procedure. Keeping the same reaction conditions, the same alkene can be synthesized in a E- or Z- enriched fashion only by changing the order of addition of the two coupling partners. Subsequently, and inspired by the “ortho-effect” present in Wittig reactions, a series of experiments with different ortho-substituted benzaldehydes were performed. The obtained results are presented in Table 5.2. Table 5.2 E- and Z- alkenes from the coupling of two ortho-substituted benzaldehydes. X = substitution on the first aldehyde A; Y = substitution on the second aldehyde B.

Entry E/Z ratio Aldehyde A Aldehyde B Yield % 1 H Br 51 33/67 2 H Me 38 34/66 3 H OMe 67 41/59 4 Br H 43 100/0 5 Me H 76 71/29 6 OMe H 57 75/25 7 Br Br 36 62/38 8 Me Me 33 37/63 9 OMe OMe 57 36/64

In entries 1-3, ortho-substituted benzaldehydes were used as second coupling partner. The results are in line with what is expected from the “ortho-effect” present in Wittig reactions: the Z-olefins are major products, with no influence noticed from the three different Y group tested (Y = Br, Me, OMe). Subse- quently, the order of addition of the two carbonyls was reversed. In entries 4- 6, E-olefins are predominately synthesized, as expected from the results presented in Scheme 5.1 for Wittig reactions. Interestingly, in entry 4 (X = Br), the reaction is 100% E-selective, while with EDG substituents (X = Me and OMe, entries 5-6) a certain amount of Z-isomer is present. Lastly, couplings with two ortho-substituted benzaldehydes were performed (entries 7-9). In entry 7 (X = Y = Br), the reaction is E-selective due to the presence of Br atom on aldehyde A. Analyzing entries 1 and 4 in Table 5.2, ortho-Br is Z-directing for aldehyde B and E-directing for aldehyde A. The two effects are working in opposite directions, explaining thus the reaction outcome. In entries 8-9, the presence of ortho-substituent on aldehyde B, as well as the presence of EDG groups on aldehyde A, are leading to the formation of Z-products. As already

69 noticed from the results presented in Table 5.1, the same product can be produced in a higher E- or Z- form only by changing the order of addition of the two aldehydes. The reader can compare entries 1-4, 2-5 and 3-6 in Table 5.2. To conclude this analysis, depending on the substitution of the two coupling partners, it is possible to control and predict the stereochemistry of the newly formed C=C double bond by adding the two aldehydes with a correct order of addition. An example is shown in Scheme 5.3, where the same olefins are synthesized in either E- or Z-fashion, just by choosing an appropriate order of addition of aldehydes A and B.

O O Br Br

a) + MeO MeO AB 67 % Z-selective

O O OMe

b) + Br OMe Br AB 100 % E-selective

Br O O Br

c) + N O AB N

O 67 % Z-selective

O O O N d) + Br N ABO Br 95 % E-selective Scheme 5.3 A proper combination of electronic-effect on aldehyde A and ortho-effect on aldehyde B can be used to predict product stereochemistry.

5.3 Conclusions The stereochemical outcome of the one-pot procedure presented in chapter 4 has been studied in more depth. The existence of two separate effects that influence the stereochemistry of the newly formed C=C double bonds has been noticed. It emerges that, by choosing an appropriate combination of substitutions on the two coupling partners, the same olefin can be formed in a E- or Z- enriched fashion, under identical reaction conditions. The stereochemical outcome depends on the order of addition of the two partners. A graphical representation of this effect is shown in Figure 5.1.

Figure 5.1. Visual representation of the connection between stereochemical outcome of the process and order of addition of the two coupling partners.

71 6. Reducing the steric bulk on phosphaalkenes – studies on the stability of P-Ph- phosphaalkenes (Paper IV)

This chapter describes synthesis and stability studies performed on phosphaalkenes which lack steric protection by bulky substituents. More specifically, P- Ph-trisubstituted phosphaalkenes are studied. The general applicability of the optimized phospha-Peterson procedure developed in the previous chapter is here confirmed, and exploited for the formation of PhP=CR2 phosphaalkenes in high yield. Additionally, a detailed investigation on the chemical equilibria of their corresponding monomeric, dimeric and oligomeric species is presented.

6.1 Introduction As described in chapters 3 and 4, reducing the size of the phosphorus protecting group is a fundamental strategy to enlarge the reaction scope of the reductive cross-coupling of carbonyls to alkenes. Indeed, when the sterically demanding Mes* group was employed, the scope of the coupling procedure was limited to activated electron-poor aldehydes (see section 3.2). Moving to the smaller P-Mes-phosphaalkenes allowed the use of deactivated electron-rich aldehyde substrates for the construction of elaborated olefinic structures (see chapter 4). However, the synthesis of trisubstituted alkenes with the newly developed method is limited (see results presented in Table 4.7). Thus, we sought the possibility of employing P-Ph-trisubstituted phosphaalkenes as intermediates for the cross-coupling of ketones and aldehydes. However, synthetic challenges arise in synthesizing phosphaalkenes bearing such a small protecting group. To better estimate the stability and reactivity of P-Ph-phosphaalkenes, their synthesis and chemical behavior was investigated separately, with the long- term intention to use these systems as intermediates in the cross-coupling of ketones and aldehydes.

72 6.2 Kinetic stabilization of phosphaalkenes As described in section 2.2.2, kinetic stabilization is a common strategy to protect the reactive P=C unit.[83] In order to stabilize acyclic transition-metal free phosphaalkenes, sterically demanding groups can be installed as phosphorus protecting groups (see Figure 2.7).[88] Among these, the super mesityl (Mes*) being the most widely used (sections 2.2.2 and 3.3). Disubstituted P- Mes*-phosphaalkenes can be isolated and purified by silica gel column chromatography.[89a] Reducing the size of the protecting group to Mes has severe impacts on the reactivity of the corresponding phosphaalkenes (see Figure 3.1), which however can be prepared and purified by distillation and/or crys- tallization.[91b] On the other hand, disubstituted P-Mes-phosphaalkenes can only be observed in solution by 31P NMR spectroscopy.[113] To date, only one example of an isolated disubstituted P-Mes-phosphaalkene has been reported,[93] obtained by the reaction of MesPLiTMS with a sterically bulky aldehyde (tBuCHO). However, the system degrades within a few days under inert atmosphere to the corresponding diphosphine, as shown in Scheme 6.1.[93]

Scheme 6.1 Degradation pathway of pure isolated MesP=CtBuH phosphaalkene.[93]

With a smaller protecting group like Ph, the stability of the corresponding phosphaalkenes dramatically decreases (Figure 3.1). Disubstituted P-Ph- phosphaalkenes have not been observed to date, while trisubstituted systems have been reported, but not isolated. In a recent work by Gates et al, in which P-Ph-phosphaalkenes are described in detail, the phospha-Peterson reaction between PhPLiTMS and benzophenone is used for the synthesis of [113] 31 PhP=CPh2. P NMR spectroscopy analysis of the reaction mixture showed the formation of three different species (Scheme 6.2).

Scheme 6.2 Reaction between PhPTMS2 and benzophenone leading to the formation of desired phosphaalkene 16a together with the side-products 17a and 18a.[113]

Attempts to isolate phosphaalkene 16a were unsuccessful. The system is unstable towards dimerization, forming the head-to-head dimer 17a. Diphos- phirane 18a can also be detected in the reaction mixture, however its formation remained unexplained.[113] Additionally, when 17a was isolated and

73 redissolved in THF, 16a was detected, suggesting the existence of a chemical equilibrium between the two species. From these examples, it is evident that decreasing the size of the protecting group of phosphaalkenes results in synthetic difficulties. In order to use P-Ph- phosphaalkenes as intermediates to broaden the scope of the carbonyl cross- coupling reaction, a high-yielding and reliable synthesis of these systems is needed. Additionally, a better understanding of the properties of P-Ph-phosphaalkenes is required, since these species would be crucial intermediate in the new cross-coupling procedure. Thus, a series of triphenylphosphaalkenes with different electronic properties were prepared as described in the following section.

6.3 Development of a new reliable synthesis of P-Ph- phosphaalkenes With a reliable and optimized procedure for the synthesis of P-Mes-phosphaalkenes in hand (section 4.5), its applicability to P-Ph-phosphaalkenes was investigated. Reaction of PhPTMS2 with one equivalent of LiOEt in THF for 4 hours at room temperature resulted in complete conversion to the desired PhPLiTMS as confirmed by 31P NMR spectroscopy. Subsequently, ethereal solutions of PhPLiTMS were reacted with benzophenone at room temperature to form the trisubstituted phosphaalkene 16a (Scheme 6.3).

Scheme 6.3 Reaction of PhPLiTMS with one equivalent of benzophenone leading to the formation of products 16a, 17a and 18a. In this example, an ethereal solution of benzophenone is added to a solution of PhPLiTMS in water-free Et2O.

Similarly to the findings of Gates et al.,[113] three species were observed in the 31P NMR spectrum of the reaction mixture (Figure 6.1): the desired phosphaalkene 16a (δ: +233 ppm), its head-to-head dimer 17a (δ: +5 ppm) and the diphosphirane 18a (δ: -118 ppm).

Figure 6.1 31P NMR spectrum of the reaction depicted in Scheme 6.3.

While monitoring the equilibrium between 16a and 17a, it was noticed that the concentration of 18a remained constant, suggesting that 18a was not involved in the above-mentioned equilibrium. It was postulated that 18a may instead be formed from a nucleophilic attack of PhPLiTMS into 16a. In order to test this hypothesis, the reaction stoichiometry was adjusted using one equivalent of benzophenone and two equivalents of PhPLiTMS, as reported in Scheme 6.4.

Scheme 6.4 The reaction between benzophenone and one equivalent of PhPLiTMS gives rise selectively to phosphaalkene 16a. Addition of a second equivalent of PhPLiTMS generates the diphosphirane 18a. In this example, an ethereal solution of PhPLiTMS is added to a solution of benzophenone in water-free Et2O.

With the stoichiometry reported in Scheme 6.4, formation of 18a is immediate and quantitative, as determined by 31P NMR spectroscopy (Figure 6.2 a), confirming our hypotheses.

75 18a

16a

18a c

Figure 6.2 31P NMR spectra of the reactions reported in Scheme 6.4. a) Selective formation of 18a by reacting one equivalent of benzophenone and two equivalents of PhPLiTMS. b) Selective synthesis of 16a achieved by adding one equivalent of a solution of PhPLiTMS to an ethereal solution of benzophenone. c) Transformation of 16a to the diphosphirane 18a by addition of a second equivalent of PhPLiTMS to the mixture shown in figure b.

Having understood how 18a is formed, the reaction was adjusted to optimize the synthesis of desired phosphaalkene 16a. Selective formation of 16a was achieved by reversing the addition of the reagents: a PhPLiTMS solution was added to a benzophenone solution in a 1:1 ratio (first part in Scheme 6.4), with no traces of 18a detected by 31P NMR spectroscopy (Figure 6.2 b). The mechanism for the formation of 18a was confirmed by addition of PhPLiTMS to the reaction mixture (second part in Scheme 6.4), resulting in the clear formation of 18a, as evidenced by 31P NMR spectroscopy (Figure 6.2 c).

6.4 Investigations on the chemical equilibria of P-Ph- phosphaalkenes Having developed a selective and high-yielding synthesis of tri- phenylphosphaalkene 16a, its applicability for the formation of P-Ph-phosphaalkenes with different substitutions was tested (Scheme 6.5).

Scheme 6.5 Synthesis of phosphaalkenes 16a-c from PhPLiTMS and different para- substituted ketones a-c, followed by the dimerization equilibrium to form head-to- head dimers 17a-c.

The reactivity of electron-rich or poor ketones (b and c, respectively) with PhPLiTMS might differ from that of benzophenone a. Furthermore, changing the electronic properties of the resulting P-Ph-phosphaalkenes may have an impact on their reactivity. This was investigated by monitoring the chemical equilibria between phosphaalkenes 16a-c and their dimers 17a-c by 31P NMR spectroscopy. A P-Mes*-phosphaalkene, more specifically (E)-(4-methox- ybenzylidene) (2,4,6-tri-tert-butylphenyl)phosphane), was used as internal standard in order to obtain quantitative information from these experiments. A typical example of 31P NMR spectroscopy monitoring starting from ketone a is presented in Figure 6.3.

Figure 6.3 31P NMR investigation of the equilibrium between 16a (δ: +233 ppm) and 17a (δ: +5 ppm), conditions as in Scheme 6.5; *Internal standard (δ: +244 ppm). a) First measurement after 14 minutes, b) after 2 hours and c) after 48 hours.

Spectroscopic analysis suggests that the concentration of phosphaalkenes 16a- c slowly decreases with concomitant formation of dimers 17a-c (Figure 6.3). The time taken for the reactions to reach steady-state equilibrium concentra- tions ranges from hours to days. For R = H, equilibrium between monomer 16a and dimer 17a is reached within 48 hours with a 1:2.4 ratio. In case of R = O-octyl, the equilibrium is reached in four days and is shifted towards the

77 monomer 16b. Conversely for R = F, equilibrium is reached in 13 hours, and favors the formation of dimer 17c. The use of an internal standard revealed an interesting feature about the reactivity of these systems. During the equilibria processes, the total amount of phosphorus species decreased significantly as determined by peak integra- tion. Once the equilibrium was reached, phosphorus containing species were no longer being depleted, suggesting additional equilibria from monomers 16a-c, as shown in Scheme 6.6. The consumption of phosphorus containing species from the reaction mixture was assigned to the formation of higher oligomers of 16a-c, which are not observable by 31P NMR spectroscopy (Scheme 6.6).

Scheme 6.6 Proposed equilibria between monomers 16a-c, dimers 17a-c and higher oligomers and consequent quenching reaction with MeOH.

If an equilibrium exists between monomers 16a-c and higher oligomers, according to Le Chatelier´s principle, all phosphorus containing species should be recovered in an irreversible reaction that depletes monomers from the reaction mixture. Methanol is known to irreversibly add across P=C double bonds.[111] Thus, MeOH was added to solutions containing 16a-c, 17a-c and oligomers in equilibrium (Scheme 6.6) and reactions were carefully monitored by 31P NMR spectroscopy (Figure 6.4).

78 17a a * 19a

b 19a * 17a

c 19a *

250 230 210 190 170 150 130 110 102030405060708090 0 -10 Figure 6.4 31P NMR investigation of the equilibria reported in Scheme 6.6. Addition of MeOH leads to the formation of phosphinite 19a (δ: + 125 ppm). a) First measurement after several minutes, b) after 30 hours and c) after 7 days. *Internal standard (δ: + 244 ppm).

As expected, MeOH immediately reacts with phosphaalkenes 16a-c to form the corresponding phosphinites 19a-c, while dimers 17a-c are not immediately affected by addition of MeOH (Figure 6.4, a). However, on a longer timescale (hours to days), the dimers 17a-c undergo conversion to phosphinites 19a-c through their equilibrium with monomers 16a-c (Figure 6.4, b). After several days, all dimers are converted into phosphinites 19a-c (Fig- ure 6.4, c). On the assumption that MeOH addition across P=C double bonds is quantitative, the final concentration of 19a-c represents the yield of the initial phospha-Peterson reaction (Table 6.1), revealing a remarkably high yield for the synthesis of phosphaalkenes 16a and 16b (87 and 73%, respectively). A lower yield (54%) was obtained for 16c, which can be attributed to the higher reactivity of the electron-deficient ketone c with PhPLiTMS. Careful 31P NMR spectroscopic analysis allowed the obtained yields to be compared to the yields for initial phosphaalkenes 16a-c formation. As seen in Table 6.1, the data are in accordance with minor deviations. This study has demonstrated that oligomerization and dimerization processes are slow when compared to phosphaalkene formation. This finding is of crucial importance and sets the basis for a potential application of triphenylphosphaalkenes as intermediates in new cross-coupling reactions of carbonyls to alkenes. Additionally, few comments on the tendency of monomers 16a-c in taking part in chemical equilibria can be made from the results shown in Table 6.1.

79 Table 6.1 Results from 31P NMR spectroscopy investigation of equilibria between 16a-c, 17a-c and higher oligomers. aYields determined using P-Mes*-phosphaalkene internal standard.

NMR yield NMR yield 16a-c at 17a-c at oligomers at of 16a-ca, of 19a-ca, equilibrium, equilibrium, equilibrium, % % % % %

a: R = H 97 87 18 45 37 b: R = O-octyl 65 73 64 17 19 c: R = F 40 54 14 31 55

The results from the quantitative analysis of the species present in equilibrium (e.g. phosphaalkenes 16a-c, their dimers 17a-c and oligomers) are summarized in Table 6.1. It was found that electron-deficient phosphaalkenes are more prone to oligomerization, since oligomers of 16c represents 55% of the corresponding reaction mixture. In comparison for the electron-rich ketone b, the equilibrium lies on the side of monomer 16b, with only 19% of the reaction mixture present in the form of oligomers. With unsubstituted benzophenone, the situation is in between the two extreme cases. These data are in accordance with the initial assumption that the electronic nature of the three systems will vary and it will influence their chemical behavior.

6.5 Conclusions Firstly, the modified phospha-Peterson reaction developed for the synthesis of P-Mes-phosphaalkenes is also compatible with the Ph substituent. The reaction of PhPLiTMS with benzophenone a allows the selective synthesis of phosphaalkene 16a, with the previously reported side-product diphosphirane 18a being not detected in the optimized synthetic procedure. Secondly, these new reaction conditions enable the formation of triphenylphosphaalkenes 16a-c with different substitution patterns from PhPLiTMS and ketones a-c. The formed phosphaalkenes 16a-c are involved in chemical equilibria with the head-to-head dimers 17a-c and oligomeric species not detectable by 31P NMR spectroscopy. The electronic nature of starting ketones a-c influences the yield of the phospha-Peterson reaction. The electron-deficient ketone c generates phosphaalkene 16c in a notably reduced yields compared to the electron-rich ketone b or unsubstituted benzophenone a. Moreover, also the equilibria are affected by the electronic nature of the starting ketones a-c. In general, electronic-rich phosphaalkenes favor the monomeric side, while electron-poor systems are present mainly in the oligomeric form. Additionally, the mixtures of monomers 16a-c, dimers 17a-c and oligomers can be all converted into their corresponding phosphinites 19a-c via irreversible quenching with

80 MeOH. Despite the poor kinetic stabilization offered by the Ph protecting group, P-Ph-phosphaalkenes can be used as intermediates in subsequent chemistry when converted to their corresponding phosphinite species. As shown in chapter 4, phosphinites can be oxidized to their phosphinate congeners which then take part in new olefination reactions. The study performed in this chapter shows that P-Ph-triphenylphosphaalkenes could be great can- didates for the development of a new cross-coupling reaction of ketones and aldehydes to form trisubstituted alkenes.

81 7. Trisubstituted alkenes from the cross- coupling of ketones and aldehydes (Paper V)

This chapter describes the development of a new cross-coupling olefination reaction of ketones and aldehydes to afford trisubstituted alkenes. The new one-pot procedure arises from a combination of findings described in chapters 4 and 6. Specifically, the new methodology presented here builds on the optimized synthesis of P-Ph-triphenylphosphaalkenes described in chapter 6.

7.1 Reducing the size of the phosphorus protecting group for increased reactivity As described in section 3.2 and chapter 4, the size of the phosphorus protecting group plays a key role in determining the substrate scope of the previously reported methodologies. A general representation of the importance of the phosphorus protecting group in relation to the outcome of the new olefination procedures is presented in Figure 7.1.

Figure 7.1 General representation of the connection between the size of the phosphorus protecting group (represented by black circles) and the outcome of the olefination procedures.

In the first report by our group, the sterically demanding Mes* group was used to kinetically stabilize the phosphaalkenes,[133] resulting in a narrow substrate scope limited to electron-poor aldehydes (R1 = para-CN-benzaldehyde, para-

82 bromo-benzaldehyde, etc. Figure 7.1). The electron-withdrawing nature of the substrate was required for the formation of the olefin products, as evidenced by the lack of reactivity of P-Mes*-diphenylphosphaalkene in the second part of the procedure.[133] While the substrate scope for the second coupling partner on the other hand is broader, olefin products could not be obtained from highly electron-rich aldehydes (R2 = para-methoxy-benzaldehyde, para-morpholine- benzaldehyde, etc. Figure 7.1).[133] Employing a smaller protecting group (Mes), allowed a much broader scope in the synthetic protocol.[144] As described in chapter 4, the new procedure has almost no limitations in the choice of the aldehydic starting materials, with excellent reactivity observed for benzaldehyde, electron-poor and electron-rich aromatic aldehydes (Figure 7.1) as first coupling partners (see Table 4.6). The scope for the second carbonyl coupling partner is also broad, with good conversions for both electron-poor and electron-rich aromatic aldehydes. While some reactivity was observed with ketones as second coupling reagents, the synthesis of trisubstituted alkenes using the P-Mes-based procedure remains unsatisfactory as poor olefins yields were obtained (see Table 4.7).[144] In order to develop a new high yielding synthetic procedure for coupling ketones and aldehydes, a smaller phosphorus protecting group seems to be needed. Hence, P-Ph-triphenylphosphaalkenes could be used as intermediates in the coupling of ketones and aldehydes to trisubstituted olefins (Figure 7.1, bottom). Building on the knowledge gained in chapters 4 and 6, the development of new olefination reactions was investigated, the result of which are presented in the following section.

7.2 Developing a new cross-coupling reaction The bases of the new olefination procedure are described in chapters 4 and 6. The optimized phospha-Peterson reaction used for the high yielding formation of P-Ph-phosphaalkenes and phosphinites represents the first part of a general and plausible new cross-coupling procedure (Scheme 7.1). After a successful oxidation of P-Ph-phosphinites, the corresponding P-Ph-phosphinates may form alkenes upon reaction with a carbonyl compound (Scheme 7.1, right), analogously to the results discussed in chapter 4.[144]

Scheme 7.1 Plausible new cross-coupling reaction of ketones and aldehydes to trisubstituted alkenes based on the work described in chapters 4 and 6.

83 Employing the reaction conditions developed for P-Mes-based phosphaalkenes resulted in the reductive coupling of a ketone and an aldehyde, according to Scheme 7.2.[152]

Scheme 7.2 Detailed reaction scheme for the new one-pot coupling of ketones and aldehydes to form unsymmetrical alkenes.[152]

As discussed in chapter 6, P-Ph-phosphinites VIII can be synthesized in a straightforward manner and in excellent yields by the reaction of PhPLiTMS VI and ketones, followed by MeOH addition. Oxidation of VIII with tBuOOH provides phosphinates IX within a few minutes. In the last step, addition of tBuOK and an aldehyde leads to the formation of desired trisubstituted alkenes. As observed for the Mes case, the procedure depicted in Scheme 7.2 can be conveniently followed by 31P NMR spectroscopy (Figure 7.2).[152]

TMS a Ph P VI Li

R2 b P VII Ph R1

MeO R2 c P VIII Ph R1

O R2 d Ph P IX 1 MeO R

Figure 7.2 31P NMR spectroscopy investigation of the elementary steps reported in [152] Scheme 7.2. Each reaction was checked with the use of a C6D6 internal capillary. (a → b) phospha-Peterson reaction between VI and R1COR2 to form P-Ph-phosphaalkenes VII. (b → c) Addition of MeOH to form phosphinite VIII. (c → d) Oxi- dation of VIII to the corresponding phosphinate IX.

Initially, addition of one equivalent of an ethereal solution of PhPLiTMS VI (δ: -143 ppm) to a solution of ketone R1COR2 results in the formation of a new

84 peak with a chemical shift around +230 ppm (Figure 7.2, b). This signal can be assigned to P-Ph-phosphaalkene VII, as noticeable from its diagnostic and typical deshielded resonance.[113] Upon MeOH addition, P-Ph-phosphaalkenes VII are rapidly converted to phosphinites VIII (δ: +126 ppm, Figure 7.2 c). Oxidation to the corresponding phosphinate IX (δ: +41 ppm) is completed in a few minutes (Figure 7.2, d). Addition of tBuOK and an aldehyde generates the desired olefin, which is not visible by 31P NMR spectroscopy. Unlike the case of bulkier phosphorus substituents, the phosphorus containing byproducts of the olefin elimination cannot be identified by 31P NMR spectroscopy. The reaction endpoint was determined by monitoring the reaction by TLC until the aldehyde reagent was fully consumed. The reaction conditions (time and temperature) required optimization based on the electronic nature of both coupling partners and this process is described in the following section.

7.3 Substrate scope With this new synthetic protocol in hand, the influence of the small P-Ph- protecting group on the substrate scope was investigated. As summarized in Table 7.1, various ketones and aldehydes can be coupled in good to excellent yields overall.

Table 7.1 Intermolecular cross-coupling of ketones and aldehydes yielding trisubstituted alkenes Entry 1st carbonyl 2nd carbonyl Product Yield %

1 63

O 2 67 Br

3 79

4 58

5 34

85 6 24

7 56

O 8 49

9 48

10 33

OctylO O

11 Br 46 Br OOctyl

12 23

Initially, reactions with unsubstituted benzophenone were tested (Table 7.1, entries 1-6). Coupling with benzaldehyde yielded the corresponding tri- phenylethylene product with 63% isolated yield (Table 7.1, entry 1). Electron- deficient aldehydes (Table 7.1, entries 2 and 3) resulted in an increased yield of the final olefin products (67 and 79%, respectively) owing to the higher reactivity of electron-poor aldehydes in the olefination step. Using the electron-rich para-methoxy-benzaldehyde resulted in a lower isolated yield of 58% (Table 7.1, entry 4), confirming the connection between the electronic nature of the aldehyde and product’s yield. Conjugated and aliphatic aldehydes (Table 7.1, entries 5 and 6 in, respectively) give rise to considerably lower yields, most likely because of competing aldol condensation under the basic reaction conditions. Subsequently, electron-deficient bis(4-fluorophenyl)methanone was tested (Table 7.1, entries 7-10). In line with the results described in section 6.4, lower phosphinite yields are obtained when electron-deficient ketones are employed (results presented in Table 6.1). Assuming the oxidation of P(III) P-Ph-phosphinites as quantitative, the formation of P(V) P-Ph-phosphinates with bis(4- fluorophenyl)methanone is estimated to be roughly 54% (result deriving from

86 the findings presented in Table 6.1). The moderate yields associated with this part of the reaction sequence impose a limit on the overall yield for the formation of the olefin. Coupling of benzaldehyde, electron-poor and rich aldehydes to bis(4-fluorophenyl)methanone was achieved in isolated yields ranging 33 to 56% (Table 7.1, entries 7-10). The results collected in entries 7-10 provide two precious pieces of information. Firstly, the correlation between the electronic properties of the aldehyde and product yield hinted in entries 1- 4 is further confirmed. The lowest yield of 33% was obtained with electron- rich para-methoxy-benzaldehyde (Table 7.1, entry 10), while higher yields were obtained for neutral or electron-poor benzaldehyde (Table 7.1, entries 7- 9). Secondly, the electronic nature of P-Ph-phosphinates intermediates have an impact on the olefin yields. As previously discussed, the yield of the P(V) species is approximately 54% for electron-poor ketones and provide olefins in comparable yields ranging 48 to 56% after reaction with benzaldehydes (Ta- ble 7.1, entries 7-9). This suggests that the reaction of electron-poor P-Ph- phosphinates and aldehydes is near quantitative. In contrast, tri- phenylphosphinates can be formed in high yields but subsequent reaction with aldehydes furnish olefins in isolated yields between 58 and 79% (Table 7.1, entries 1-4). These data suggest a lower reactivity of neutral tri- phenylphosphinates compared to electron-poor P-Ph-phosphinates. A further confirmation of these trends is found when the electron-rich ketone bis(4-octyloxyphenyl)methanone is used (Table 7.1, entries 11 and 12). While P-Ph-phosphinates are synthesized in excellent yield with electron-rich ketones, low alkene yields are obtained after reaction with aldehydes. For instance, reaction with electron-poor para-bromobenzaldehyde forms the alkene product in 46% isolated yield (Table 7.1, entry 11), owing to the poor reactivity of electron-rich P-Ph-phosphinates in the olefination step. Addition- ally, coupling with electron-rich para-methoxybenzaldehyde forms, as expected, the corresponding product in remarkably poor yield of 23% (Table 7.1, entry 12). In addition to the isolated yields of the final products, the reaction times and conditions are also affected by the electronic nature of the two coupling partners. The general trend is that coupling with electron-poor aldehydes is fastest, with full conversion achieved between 5 and 30 minutes at room temperature while electron-rich aldehydes require longer reaction times (hours) or elevated temperatures. The reaction conditions also depend on the reactivity of the P(V) intermediates. As described earlier, electron-poor P(V) species are comparably more reactive, and result in short reaction times with aldehydes. Reactions of electron-poor P(V) intermediates and electron-deficient aldehydes result in the formation of alkenes in a few minutes, while neutral tri- phenylphosphinates require usually 30 to 60 minutes. Lastly, the coupling of electron-rich P(V) intermediates are the most challenging. The most extreme

87 case is presented in entry 12 Table 7.1 where both reagents exhibit poor reactivity, and reflux in THF for 24 hours was needed for olefin’s formation. These trends are visually summarized in Table 7.2.

Table 7.2 Visual representation of the connection between electronic nature of the coupling partners and their reactivity.

R = EWG: - Poor yield of P(V) formation + High yield of P(V) formation good reactivity

R = EDG: + Good reactivity - Poor reactivity poor reactivity

With the reaction scope determined for symmetrical ketones, the suitability of the procedure to preferentially synthesize either E- or Z-isomers of the trisubstituted alkenes from unsymmetrical ketones was investigated (Table 7.3).

Table 7.3 Synthesis of E/Z-trisubstituted olefins by the coupling of unsymmetrical ketones and aldehydes. Yield %, Entry 1st carbonyl 2nd carbonyl Product E/Z ratio

Br O 39, 1 55/45 Br Br

Br O 36, 2 59/41 Br OMe

O 61, 3 61/39

44, 4 57/43

88 Similar reactivity trends were observed for unsymmetrical ketones with re- spect to the aldehyde coupling partner. Electron-poor para-bromobenzaldehyde resulted in higher final product yields than electron-rich para-methoxybenzaldehyde (Table 7.3, entries 1-2 and 3-4, respectively). Electron-rich ketones showed better results than electron-poor substrates (Table 7.3, entries 1- 3 and 2-4) which may be due to the higher yield of formation of electron-rich P-Ph-phosphaalkenes compared to electron-poor congeners. As in the case of symmetrical ketones, the isolated yields depend on the electronic nature of the two coupling reagents while the stereochemical outcome seems unaffected. As shown in Table 7.3, no specific trend in the E/Z product ratio was found, with E-isomers being only slightly favored than Z ones.

7.4 Advantages and limitations In analogy to the discussion in chapter 4, the method that is subject of this chapter has several advantageous features compared to the classic McMurry coupling. The newly developed protocol does not require any transition metals, eliminating tedious in-situ preparations of Ti0. Moreover, this new reaction proceeds smoothly at room temperature in most cases, and trisubstituted alkenes can be formed within a few minutes from readily available carbonyls feedstocks. Undesired homo-coupled products, often found in typical McMurry reactions, where not detected in this procedure. In comparison with classical olefination reactions, the presented method has the main advantage of forming alkenes in a one-pot procedure. Previously inaccessible trisubstituted olefins were synthesized in high yields with short reaction times using the P-Ph reagents, circumventing the need to isolate olefinating reagents with lengthy syntheses such as phosphonium ylides and phosphonates. There is contrasting reactivity between the P-Ph- and P-Mes-phosphinate based systems. For instance, using P-Mes-phosphinates obtained from ketones results in unsatisfactory conversions. Decreasing the protecting group to Ph, proved crucial to enhance the reactivity of the phosphinate species, confirming the hypotheses outlined in section 7.1: the steric demand of the protecting group has a significant influence in dictating the reactivity of the coupling protocols developed in our group. However, this new method has some limitations: only aromatic carbonyl reagents could be used in the procedure. Aliphatic ketones proved to be challenging substrates for the initial phospha-Peterson reaction. The corresponding P-Ph-phosphaalkenes lack kinetic stabilization, resulting in decomposition. Similarly, the scope of aldehyde substrates could not be extended to aliphatic aldehydes, the olefins of which could only be isolated in low yields, owing to competing aldol side reactions. Lastly, attempts towards the coupling of two ketones to synthesize tetrasubstituted olefins were unsuccessful, even

89 with forcing conditions, and only decomposition products could be observed. This suggests that a new approach has to be considered to target tetrasubstituted olefins.

7.5 Conclusions The presented one-pot procedure allows ketones and aldehydes to be reductively coupled under an ionic mechanism with synthetically useful yields for the first time. Trisubstituted alkenes can be prepared in short reaction times under mild conditions from feedstock materials, without the need of purifying intermediates. The procedure is applicable for the coupling of various aromatic carbonyls with different substituents. The reactivity trends explored thus far provide a basis for suitable substrate combinations for future syntheses of trisubstituted olefins.

90 8. Outlook

The studies performed in this thesis are centered around phosphorus-mediated one-pot syntheses of alkenes from the cross-coupling of two different carbonyl compounds. The size of the phosphorus protecting group plays a crucial role in determining the applicability of our new procedures to challenging substrates. Moving from the large Mes* to the small Ph group had dramatic effects in the substrate scope of the procedures, allowing the cross-coupling of ketones and aldehydes with great results. One can imagine that also the E/Z selectivity could be tuned modifying the nature of the phosphorus protecting group. In- troducing fluorinated protecting groups (e.g. 2,4,6-trifluorophenyl or perfluor- ophenyl) could impact the electronic properties of the intermediates and, in turn, the stereochemical outcome of the reaction. This is compatible with the Still-Gennari modification of the classic HWE olefination which takes advantage of CF3CH2O-phoshponates for the stereoselective synthesis of Z-alkenes (see Figure 2.5).[48] Similar reactivity can be expected from new fluorinated protecting groups. The newly developed procedure can also be modified changing the quenching reagents of phosphaalkenes intermediates. P=C double bonds are known to react with different nucleophiles (see section 4.3). Among these, secondary amines (R2NH) represent interesting trapping reagents to be tested in our procedures. Attack of amines onto P=C double bonds followed by oxidation, may generate Corey-Kwiatkowski type olefinating reagents (see Figure 2.5), which are known to promote stereoselective Z-olefination of carbonyls.[50a, 50b] Regarding the substrate scope of the procedures described in this thesis, di- and trisubstituted alkenes have successfully been synthesized. Terminal and tetrasubstituted olefins represent interesting targets, which have eluded our procedures thus far. Terminal olefins are fascinating and synthetically useful functional groups in modern organic chemistry. For instance, Heck cross-coupling[153] and olefin metathesis[154] are two fundamental reactions performed on terminal olefins. Terminal alkenes could be envisaged from the reductive cross-coupling of ketones or aldehydes with commercially available formaldehyde. Tetrasubstituted alkenes also represent an important class of molecules in organic chemistry.[155] Their synthesis could be derived from ketone- ketone cross-coupling reactions. Thus far, our efforts towards these compounds using P-Ph-phosphinates have been unsuccessful. Optimization of conditions may be required to obtain these highly substituted alkenes.

91 Another aspect that could be improved in the procedures developed in this thesis is their atom economy. The protocols described in chapters 4 and 7 generate one equivalent of phosphorus by-product per olefin equivalent. This waste could be recovered for the preparation of the initial coupling reagents, as shown in Scheme 8.1. The reduction of phosphine oxides, phosphinates and phosphonate to phosphines has been reported.[156] Once generated, phosphines can form desired RPLiTMS starting material in two high yielding steps.

Scheme 8.1 Proposed cycle for the re-use of phosphonate byproducts. R = Mes or Ph.

Lastly, a final and obvious goal of this project is its application in the synthesis of more complex alkenes or olefins resembling natural products (see for instance molecules shown in Figure 1.1).

92 9. Conclusions and summary

This thesis contributes to the fields of low-coordinate and high-valent organophosphorus chemistry. More specifically, this thesis deals with the chemistry of phosphaalkenes (low-coordinate phosphorus species) and the formation of olefins from high-valent phosphorus based olefinating reagents. In the context of compounds containing P=C double bonds, the results obtained in this work are remarkable for the following reasons: the optimized phospha-Peterson reaction conditions developed in this thesis allow the synthesis of P-Mes- and P-Ph-phosphaalkenes that had previously been reported unstable. A crucial discovery was made in the use of LiOEt for the preparation of the RPLiTMS starting reagents rather than the harsh conditions (e.g. refluxing with MeLi) previously used in the literature. The other advantage comes in the subsequent reaction of RPLiTMS with carbonyls to yield the corresponding phosphaalkenes, where small alkyl lithium bases (e.g. nBuLi or MeLi) can catalyze anionic polymerization reactions of sterically unprotected phosphaalkenes. In the work described in this thesis, LiOEt selectively cleaves one P-Si bond of starting RPTMS2, leading to the quantitative formation of desired RPLiTMS, together with the volatile and highly stable silyl-ether by- product (TMS-O-Et). Traces of LiOEt are unreactive with the thereby formed phosphaalkenes, which can be present in solution up to several days, although protected by small Mes and Ph groups! The optimized synthesis of such species is crucial for the development of new coupling reactions. P-Mes- and P- Ph-phosphaalkenes described herein have a perfect balance between reactivity and stability. Indeed, these systems are stable enough when in ethereal solution, but react immediately and quantitatively with MeOH. Oxidation of the latter concludes the “bridging” of low-coordinate to high-valent phosphorus chemistry. High-valent phosphorus olefinating reagents have been known for several decades. Phosphonium ylides (Wittig), phosphine oxides (Horner-Wittig), phosphonates (HWE) and phosphinates (Horner) have all been employed in couplings with ketones and aldehydes to form C=C double bonds. Usually, between two and three synthetic steps are required for their synthesis. Thus, time and materials are consumed for the generation of phosphorus-based olefinating reagents. Moreover, these species are frequently formed from aryl and alkyl halides. Hence, in a retrosynthetical view, olefins are synthesized from carbonyls and aryl/alkyl halides. Instead, in the present thesis, phosphinate olefinating reagents can be synthesized in a few minutes upon activation of

93 phosphaalkenes, which are generated from carbonyl groups. Without any purification, P-Mes- and P-Ph-phosphinates react with a second carbonyl generating desired alkenes from two different carbonyls. As shown by the results reported in this thesis, the size of the phosphorus protecting groups plays a crucial role for determining the substrate scope of the newly developed procedures. Figure 9.1 summarizes the connection between size of the protecting group and obtained alkenes.

O O O O R1 O R1 O R1 2 2 2 2 *Mes P R Mes P R R Ph P R 1 H OMe OMeR

EWG O

O R1 R2 EWG O R1 Aliph Aliph R2 1 O R

Aliph R1 R2 O

R2 1 R 2 EWG EWG R Reactivity of the carbonyl compound carbonyl the of Reactivity O

EDG EDG Figure 9.1 Connection between the size of the phosphorus protecting group and the formation of alkene products. On the left side, the reactivity of carbonyls in the olefination step increases from bottom to top (top: most reactive, bottom: less reactive).

As illustrated in the first column of Figure 9.1, P-Mes* phosphine oxides offer a quite poor substrate scope. Firstly, only disubstituted olefins arising from aldehyde-aldehyde couplings can be synthesized. As first coupling partner (red rectangle in Figure 9.1), only electron-poor aldehydes can be used. The choice of the second substrate (blue rectangle in Figure 9.1) is somewhat broader, although it is still restricted to electron-neutral aldehydes. Electron- rich carbonyls, unreactive with the Mes* procedure, show instead reactivity with the Mes-based protocol (second column in Figure 9.1). With the reduced size of the phosphorus protecting group, all types of aromatic aldehydes (electron-poor, neutral and rich) can be employed for the synthesis of elaborate olefinic structures. To some extent, also ketones can be utilized as second coupling reagents, but the yields of the resulting trisubstituted olefins are low. In order to achieve satisfactory cross-couplings of ketones to aldehydes, the size

94 of the protecting groups has to be decreased to the smaller Ph moiety (third column in Figure 9.1). Indeed, aromatic ketones can be used as first coupling partners (not aldehyde since the corresponding P-Ph-phosphaalkene would be too unstable), to be then reacted with all types of aldehydes. As it emerges from Figure 9.1 and from the analysis reported in chapter 8, the formation of terminal (coupling with formaldehyde) and tetrasubstituted olefins (coupling of two ketones) has still to be accomplished. Lastly, a comparison with the only known method used to reductively couple two carbonyls, namely the McMurry reaction, can be done. The McMurry coupling requires tedious in-situ preparations of low valent Ti0 species from mixture of TiCl3 or TiCl4 with strong reducing agents like Zn or Zn-Cu. In the procedures developed in this thesis, transition metals are omitted at any stage. Alkenes can be formed within short reaction times and mild conditions; in opposition to harsh reaction conditions applied in typical McMurry couplings (refluxing in high boiling solvent for several hours). But the most advantageous feature of the newly developed protocols is found in their ability of generating selectively unsymmetrical alkenes from the coupling of two different carbonyls. This selectivity arises from a sequential ionic mechanism, in which the two carbonyl reagents are added at different stages of the one-pot reaction. On the contrary, in the McMurry coupling the two carbonyls are present at the same time and, due to its radical mechanism, the reaction lacks selectivity and yields mixtures of symmetrical and unsymmetrical alkenes when two different carbonyls are reacted.

95 Svensk sammanfattning

Denna avhandling bidrar till fältet för kemi med lågt koordinerade och högva- lenta organofosfor föreningar. Mer specifikt behandlar avhandlingen, kemi med fosforalkener (lågt koordinerade fosforföreningar) och formationen av olefiner från högvalenta fosforbaserade olifineringsreagens. I kontexten för föreningar som innehåller P=C dubbelbindningar är resultaten anmärkningsvärda av följande anledningar: de optimerade reaktionsför- hållandena i fosfor-Peterson reaktionen som utvecklats i denna avhandling, tillåter den tidigare okända syntesen av P-Mes- och P-Ph-fosforalkener. An- vändandet av LiOEt för syntesen av RPLiTMS reagenset visade sig vara en avgörande upptäckt då man kan undvika de stränga förutsättningarna (d.v.s återloppskokning med MeLi) som tidigare använts i litteraturen. En fördel till kommer i följande reaktion då RPLiTMS reagerar med car- bonyler för att bilda korresponderande fosforalken. Små litierade alkylbaser (så som nBuLi eller MeLi) kan katalysera anjoniska polymerisationsreakt- ioner för steriskt oskyddade forsforalkener. I arbetet som beskrivs i denna avhandling så klyver LiOEt selektivt en P-Si bindning i RPTMS2 reaktanten. Reaktionen leder till en kvantitativ produktion av den önskade RPLiTMS med den flyktiga och mycket stabila silyeter (TMS-O-Et) som biprodukt. Spår av LiOEt reagerar inte vidare med fosfoalkenprodukten, som är stabil i lösning flera dagar, även om den endast är skyddad av små Mes och Ph grupper! Den optimerade syntesen av en sådan specie är avgörande för utvecklingen av nya kopplade reaktioner. P-Mes- och P-Ph-fosforalkenerna som beskrivs här visar en perfekt balans mellan reaktivitet och stabilitet. Som tidigare beskrivet är dessa system tillräckligt stabila i lösning men reagerar direkt och kvantitativt med MeOH. Oxidation av den efterföljande färdigställer länknings processen för kemi med lågt koordinerade och högvalenta organofosfor föreningar. Högvalenta olifinerings reagens baserade på fosfor har varit kända i flera decennier. Fosfoniumylider (Wittig), fosfinoxider (Horner-Wittig), fosfonater (HWE) och fosfinater (Horner) har alla använts i kopplings reaktioner med ketoner och aldehyder för att forma C=C dubbelbindningar. I de flesta fall krävs det mellan två och tre syntetiska steg för denna syntes, därför krävs mycket tid och material för att generera fosforbaserade olifinerings reagens. Utöver detta så kommer dessa specier ofta från aryl-och alkylhalider. Därför är, från ett retrosyntetiskt perspektiv, olefiner syntetiserade från karbonyler och aryl- eller alkylhalider. I denna avhandling presenteras alternativet att an-

96 vända fosfinat som olifierings reagens dessa kan syntetiseras på några få mi- nuter efter aktivering av fosforalkener, som is sin tur är genererade från kar- bonylgrupper. Utan att rena reagerar P-Mes- och P-Ph-fosfinater med en andra karbonyl, vilket leder till att den eftersträvade karbonylen genererats från två karbonyler. Resultaten i denna avhandling visar att storleken på fosforets skyddande grupp spelar en avgörande roll för att bedöma vilka substrat som går att an- vända för de nyutvecklade procedurerna. Figur 9.1 sammanfattar kopplingen mellan storleken på den skyddade gruppen och alkenens formation.

O O O O R1 O R1 O R1 2 2 2 2 *Mes P R Mes P R R Ph P R 1 H OMe OMeR

EWG O

O R1 R2 EWG O R1 Aliph Aliph R2 1 O R

Aliph R1 R2 O

R2 1 R 2 EWG EWG R O

EDG EDG Figur 9.1. Sambandet mellan storleken på fosforets skyddande grupp och formationen av alkeneprodukter. På vänster sida, reaktiviteten hos karbonylerna i olifinerings steget ökar nedifrån och upp (högst upp: mest reaktiv, längst ned: minst reaktiv).

Så som det är illustrerat i den första kolumnen i Figur 9.1, så är reaktions möjligheterna för P-Mes* fosfinoxider begränsade. För det första kan endast dubbelsubstituerade olifiner, från aldehyd-aldehydkopplingar syntetiseras med denna gigantiska skyddande grupp. Som första koppling partner (röda rektangeln i Figur 9.1) kan endast elektronfattiga aldehyder användas. Valet av andra reagens (blå rektangeln i Figur 9.1) är något mer öppet men fortfa- rande begränsat till elektronfattiga aldehyder. Elektronrika karbonyler som inte reagerar med Mes*-proceduren, visar reaktivitet med det Mes- baserade protokollet (andra kolumnen i Figur 9.1). En mindre skyddande grupp gör det

97 möjligt att använda alla typer av aromatiska aldehyder (elektronfattiga, ne- utrala och elektronrika) kan användas för syntes av komplicerade olifinstruk- turer. Till viss utsträckning kan även alifatiska ketoner användas som andra kopplingsreagens, men utbytet av de korresponderande trippelsubstituerade olefinerna är lågt. För att nå en tillfredställande kopplingsreaktion mellan ketoner och aldehyder måste storleken på den skyddande gruppen minskas yt- terligare till den mindre Ph gruppen (tredje kolumnen i Figur 9.1). Aromatiska ketoner kan användas som första kopplings partner (inte aldehyder då den korresponderande P-Ph-fosforalkenen är för ostabil), dessa kan sedan reagera med alla typer av aldehyder. Så som det visas i Figur 9.1 och från analysen som beskrivs i kapitel 8, så kan formationen av terminala (kopplingar med formaldehyder) och tetrasubstituerade olefiner (koppling mellan två ketoner) inte än syntetiseras med denna metod. Avslutande, så kan en jämförelse göras med den enda tidigare kända metoden för att reduktivt koppla två karbonyler göras, nämligen McMurry reaktionen. McMurry koppling kräver ansträng- 0 ande in-situ förberedelser av lågvalenta Ti specier från en blandning av TiCl3 eller TiCl4 bed starka reducerande medel så som Zn eller Zn-Cu. I proceduren som utvecklats i denna avhandling undvikes övergångmetaller i alla reaktions- stegen. Alkener kan syntetiseras under korta reaktionstider och milda omstän- digheter; till skillnad från de stänga reaktions förutsättningarna som används i de flesta McMurry kopplingarna (återloppkoking i flera timmar). Den största fördelen med den ny procedur är dock egenskapen att skapa selektivt osymmetriska alkener genom att koppla ihop två olika karbonyler. Denna selektivitet kommer från den sekvensbaserade joniska mekanismen, där de två karbonyl reagensen tillsätts vid olika faser av enkelkärlsreaktionen. Jämförelse- vis så saknar McMurry reaktionen selektivitet på grund av radikalmekanismen då båda aldehyder tillsätts samtidigt, detta leder till blandningar av symmet- riska och osymmetriska alkener när två olika karbonyler reagerar.

Översatt av Martin Axelsson

98 Acknowledgements

I have been extremely lucky to pursue my PhD studies in the unique student town of Uppsala and in such an amazing and international environment offered by Uppsala University. In the last four incredible years, I have met many people from all around the world. This has helped me not only in growing my network, but most importantly in developing an open-minded vision of life and to learn from different cultures and countries. Thus, I would like to grate- fully thank the following people:

My mentor and PhD advisor Prof. Sascha Ott. I am deeply grateful for the opportunity you gave me and the trust you always had for me. It has been of a great pleasure working for you during my PhD studies. You are friendly, open-minded, always ready to help me when is needed, and of a great support both under a personal and working manner.

My co-advisor Dr. Anna Arkhypchuk, what should I say... You have perfectly guided me during this journey into phosphorus chemistry. It was great fun working together, creating what was thought to be impossible only a few years ago. We are the cowboys of the synthesis! I wish you all the best for your future career, you deserve it!

My second co-advisor Dr. Andreas Orthaber, I am happy I came as an Eras- mus exchange student many years ago in your group. Thanks to the opportunity you gave me, I had the chance to be enrolled as a PhD student at Upp- sala University. Apart from that, it has been a great pleasure discussing science and, most importantly, climbing with you.

Dr. Sébastien Bontemps, thank you for giving me the chance to visit and dis- cover your lab, and such an amazing country like France! You have literally helped me in everything, I really appreciate it. Merci!

Then I would like to thank my climbing family. In the last four years, I became a bit addicted to this great sport. But on top of everything, having this amazing group of friends, a family, has made my staying in (cold and dark) Uppsala unique. Thanks to: Dr. Mario, everything started with meeting you in a normal (boring) gym. You brought me to Allis, and you have been like a big brother during this

99 journey. “Mparozzo” Marco, well, how many things did we do together? How many times did we freeze on the way back home? There would be a thousand things to say, but a thanks is probably the best! Juliana, I enjoyed every single warm hug between us; as well as great time spent outdoor in lovely Swedish nature. Mikel, I liked doing so much sport together. Thanks for showing me great spot for MTB around Uppsala! And for drinking great craft beers together. Venu, thanks for great long climbing sessions together, and, of course, for amazing drinking nights at Palermo. Luca P., I enjoyed not only climbing together but also cooking delicious Italian food with you! Prune, thanks for keeping up the party mood in Uppsala! We need this (and you). Tom, thanks for offering us delicious beer tastings and for showing us some of the best camping-climbing spots in Sweden. Erik C., it was amazing to spend so many sunny days together this last year; and to complain about (horrible) Swedish weather with a Swede. Daniel, I really enjoyed our few sessions of climbing outdoor and intense BBQ as well. Oyvind, it was always so much fun hanging out with an always smiling person like you. Elena and Julian, I am really happy we met, climbing wise but also for sharing some nice underground parties together. Katarzyna, I had fun following some PhD courses together and especially hugging you whenever we met. Achmed, thanks for giving great and wise advice about my future. Moritz and Jana, you just joined this amazing group. Even though we had a short time together, I really enjoyed meeting you and having fun together in my last months in Sweden. Disa, I am sorry we did not have enough time to develop our friendship, we miss you. And in general, I would like to thanks all the amazing people I met at the gym for all the climbing sessions and beers we drank together: Luca M., Ilari, Jon, Oskar, Tobias, Erik D., Miriam, and Mauricio.

Then it is time for my German family: Dr. Juri, it was more than super spending time together! In the lab, in the gym and in our (sh)amazing (couple) lovely trips! We enjoyed our time at the best we could! I am happy for it! Dr. Lukas, I feel so lucky to have lived together during these years! It was always fun watching football and Bojack together; I especially liked meeting your great family and friends that visited you! (Dr.) Johanna, I am extremely happy that I met you. We started a crazy journey together, one that not so many would have taken. You are a nice person, and I am glad we shared our lives for quite some time. (Mr) Martin, we had a great connection although the time together was rather short. I enjoyed every long techno night in Sweden, Berlin and Russia we spent together! Looking forward for more to come! (Mr Bambino) Sebastian, we also had great time in Uppsala. Talking about food and often closing Palermo together was super fun. (Ms) Dulcie, I am counting you in the German family now! We had so much fun in underground techno-nights in Uppsala!

100 (Mr) Michael, my journey in Uppsala started many years ago with you. I am happy we shared that time together!

Uppsala has offered me the chance to meet so many amazing colleagues and friends: Dr. Keyhan, we had so much fun in the lab, at the gym and in all the crazy Katushka parties! Dr. Shameem, it was shamazing sharing time with you in the lab and in the gym. Dr. Daniel Salazar, I needed my latino-hermano in these years. I am really happy for the crazy time we had together! And also for the nice time I spent with your lovely family. Dr. Michele, thanks for being my older brother during these years. I really appreciated it. Dr. Emilie, I am glad I met you. We did so many amazing things together; I really have good memories of it. Merci! The “Harem”. Dr. Starla, thank you for your amazing energy, for your hugs, for belaying me, for dancing together and for teaching us English; I meeaan! Daniel, I am super happy I met you! I loved doing so many things together. Cooking, climbing, listening to techno and travelling together! I wish you all the best for your PhD! Luuudo, it was short but intense. It was super nice to meet you and to form this amazing group! Crazy stories! Ashleigh, it was really nice to have coffee breaks together, as well as speak- ing about tennis and all the rest. And, of course, thanks for proof reading my Italian-English thesis! Dr. Jordann, thank you for showing me great tricks in the lab and sharing the responsibility of the stills. We are all lucky you came to our labs! Bambinissimo Alessandro, it was cool having you as a student. I appreciated meeting you and “guiding” you during your master thesis project. To the great group of skiers and climbers: Dr. Anna, Leon, Dr. Bryan, Dr. Kelly, Manuel and Kate; we had a lot of fun in this last year. I really enjoyed our few group trips, our dinners and just simply spending time with you in the lab and outside. Martin A., thanks for the translating my thesis’ summary to Swedish, as well as for some nice lunch break discussions we had over the last two years. And to all the past and present members of SMC for creating such a nice and relaxing working atmosphere: Dr. Eszter B., Dr. Henrik O., Dr. Anders T., Dr. Hemlata A., Dr. Ben J., Timofey L., Dr. Noémie L., Dr. Josh G., Dr. Avind G., Dr Mariia P., Dr. Jacinto S., Dr. Luca D., Belinda P., Astrid N., Dr. Claudia D., Dr. Daniel K., Holly R., Eirini S., Salauat K., Maxime L, Nina S., Wanja G. Yunchen, you are literally the first person I met in Uppsala a long time ago. I am extremely happy that we have kept meeting regularly even after such a

101 long time. It has been always super nice! Good luck for the ending of your PhD and for your future. Luce, as you already know, I am really happy we met. We have had a great time together during the last year. Playing tennis, cooking and simply spending time with you was sensational. The pool guys, Jorgos, Ruzhdi and Bahati, studying Swedish together and, most importantly, having our nice tradition of playing pool together every week was so much fun! Thank you for it!

Additionally, to all the nice people I met during my short (and weird) time in amazing sunny Toulouse. You have made me feel really good during this challenging experience. Nicolas and Charlenne, it was fun to meet you again after spending some time together in Uppsala. Aurèle, my stay in Toulouse would have never been the same without you. Thank you for all the beautiful nights we shared during this new and weird situation we were all in. I am grateful I got to meet you! Thank you also Dan, Sarah, Ramaraj, Carlos and Angelica.

To my Italian friends. Luciano, grazie per averci regalato un po di notti italiane con ottimo cibo, vino e la tua immensa simpatia. Arthur, grazie infinito per esserme stato vicino in questi anni lontani. E grazie per avermi visitato ben due volte, creando altri ricordi indelebili. Luca Ska e Aldo, che viaggio, dalle superiori a diventare grandi insieme! Sono felice di avervi. Vale, anche se molto lontano, ho sempre amato le nostre video telefonate e il nostro mega viaggio nel Sud-est Asiatico. Raffone, grazie per scrivermi e per aver sempre passato delle belle serate quando tornavo a casa per Natale. Quelli del Lasmo, Bianca, Masche, Dani e Divo. Sono contento che siamo riusciti a mantenere i contatti nonostante la distanza e gli anni passati! Spero di poter passare altre belle serate insieme. L’onnipresente Branco (Br6). Cosi lontani e sparsi per il mondo, ma sempre presenti quando ne avevo bisogno! E grazie a tutti i miei cari amici che hanno speso un pó di tempo per tenersi in contatto. In particolare, Manuelone, Ale Anza, Gremp, Andre Orru, Alfa, Ely e Luigi D.

E infine, un grazie infinito alla mia splendida e unica famiglia. Grazie per avermi supportato in tutti questi anni lontani da voi. Vi voglio bene.

102 References

[1] V. G. Becker, Z. anorg. allg. Chem. 1976, 423, 242-254. [2] M. Regitz, O. Scherer, Multiple Bonds and Low Coordination in Phosphorus Chemistry, Thieme Verlag, Stuttgart, Germany, 1990. [3] a) M. François, Angew. Chem., Int. Ed. 2003, 42, 1578-1604; b) K. B. Dillon, F. Mathey, J. F. Nixon, Phosphorus: The Carbon Copy; From Organophos- phorus to Phospha-organic Chemistry, John Wiley & Sons, New York, 1998; c) P. L. Floch, Coord. Chem. Rev. 2006, 250, 627-681; d) F. Mathey, Acc. Chem. Res. 1992, 25, 90-96; e) F. Mathey, F. Mercier, C. Charrier, J. Fischer, A. Mitschler, J. Am. Chem. Soc. 1981, 103, 4595-4597. [4] a) J. I. Bates, J. Dugal-Tessier, D. P. Gates, Dalton Trans. 2010, 39, 3151- 3159; b) K. J. T. Noonan, D. P. Gates, Angew. Chem., Int. Ed. 2006, 45, 7271- 7274; c) V. A. Wright, D. P. Gates, Angew. Chem., Int. Ed. 2002, 41, 2389- 2392. [5] a) L. Weber, Angew. Chem., Int. Ed. 1996, 35, 271-288; b) J. Dugal-Tessier, E. D. Conrad, G. R. Dake, D. P. Gates, in Phosphorus (III) Ligands in Homo- geneus Catalysis, 1st ed. (Eds.: C. J. Kamer, P. W. N. M. Leeuwen), John Wiley & Sons, Ltd., Chichester, UK, 2012, pp. 321 - 341. [6] a) M. A. Shameem, A. Orthaber, Chem. Eur. J. 2016, 22, 10718-10735; b) R. C. Smith, J. D. Protasiewicz, J. Am. Chem. Soc. 2004, 126, 2268-2269; c) M. P. Duffy, P. Bouit, M. Hissler, in Main Group Strategies towards Functional Hybrid Materials: Applications of Phosphorus-Based Materials in Optoelec- tronics (Eds.: T. Baumgartner, F. Jäkle), John Wiley & Sons, Ltd., New Jersey, USA, 2018; d) R. J. Gilliard, J. M. Kieser, J. D. Protasiewicz, in Main Group Strategies towards Functional Hybrid Materials: Synthons for the Develop- ment of New Organophosphorus Functional Materials (Eds.: T. Baumgartner, F. Jäkle), John Wiley & Sons, Ltd., New Jersey, USA, 2018. [7] a) S. E. Kelly, in Comprehensive Organic Synthesis (Eds.: B. M. Trost, I. Flem- ing), Pergamon, Oxford, 1991, pp. 729-817; b) J. M. J. Williams, Preparation of Alkenes: A Practical Approach, Oxford University Press, UK, 1996; c) W. S. Knowles, Angew. Chem., Int. Ed. 2002, 41, 1998-2007; d) R. Noyori, An- gew. Chem., Int. Ed. 2002, 41, 2008-2022; e) K. B. Sharpless, Angew. Chem., Int. Ed. 2002, 41, 2024-2032; f) Y. Chauvin, Angew. Chem., Int. Ed. 2006, 45, 3740-3747; g) R. R. Schrock, Angew. Chem., Int. Ed. 2006, 45, 3748-3759; h) R. H. Grubbs, Angew. Chem., Int. Ed. 2006, 45, 3760-3765. [8] a) C. Rivière, A. D. Pawlus, J.-M. Mérillon, Nat. Prod. Rep. 2012, 29, 1317- 1333; b) B. De Filippis, A. Ammazzalorso, M. Fantacuzzi, L. Giampietro, C. Maccallini, R. Amoroso, Chem. Med. Chem 2017, 12, 558-570. [9] Y. Shang, Nat. Rev. Cancer 2006, 6, 360. [10] M. Jang, L. Cai, G. O. Udeani, K. V. Slowing, C. F. Thomas, C. W. W. Bee- cher, H. H. S. Fong, N. R. Farnsworth, A. D. Kinghorn, R. G. Mehta, R. C. Moon, J. M. Pezzuto, Science 1997, 275, 218-220. [11] J. McNulty, D. McLeod, Tetrahedron Lett. 2013, 54, 6303-6306.

103 [12] F. McTaggart, P. Jones, J. Cardiovasc. Drugs Ther. 2008, 22, 321-338. [13] a) Y.-J. Cheng, S.-H. Yang, C.-S. Hsu, Chem. Rev. 2009, 109, 5868-5923; b) H. E. Katz, Z. Bao, S. L. Gilat, Acc. Chem. Res. 2001, 34, 359-369. [14] J. E. McMurry, M. P. Fleming, J. Am. Chem. Soc. 1974, 96, 4708-4709. [15] a) I. Beadham, J. Micklefield, Curr. Org. Synth. 2005, 2, 231-259; b) T. Takeda, Modern Carbonyl Olefination, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2004. [16] L. Kurti, B. Czako, Strategic Applications of Named Reactions in Organic Syn- thesis, 2005. [17] J. Wang, Stereoselective Alkene Synthesis, Springer-Verlag, Berlin Heidel- berg, 2012. [18] a) G. Wittig, G. Geissler, Liebigs Ann. Chem. 1953, 580, 44-57; b) G. Wittig, U. Schollkopf, Chem. Ber. 1954, 87, 1318-1330. [19] a) W. S. Wadsworth, W. D. Emmons, J. Am. Chem. Soc. 1961, 83, 1733-1738; b) J. Boutagy, R. Thomas, Chem. Rev. 1974, 74, 87-99. [20] a) M. Julia, J.-M. Paris, Tetrahedron Lett. 1973, 14, 4833-4836; b) P. J. Ko- cienski, B. Lythgoe, S. Ruston, J. Chem. Soc., Perkin Trans. 1978, 1, 829-834. [21] a) D. J. Peterson, J. Org. Chem. 1968, 33, 780-784; b) L. F. v. Staden, D. Gravestock, D. J. Ager, Chem. Soc. Rev. 2002, 31, 195-200. [22] a) J. E. McMurry, Acc. Chem. Res. 1974, 7, 281-286; b) B. E. Kahn, R. D. Rieke, Chem. Rev. 1988, 88, 733-745. [23] F. N. Tebbe, G. W. Parshall, G. S. Reddy, J. Am. Chem. Soc. 1978, 100, 3611- 3613. [24] N. A. Petasis, E. I. Bzowej, J. Am. Chem. Soc. 1990, 112, 6392-6394. [25] a) L. A. Paquette, L.-Q. Sun, D. Friedrich, P. B. Savage, J. Am. Chem. Soc. 1997, 119, 8438-8450; b) I. Martı́nez, A. R. Howell, Tetrahedron Lett. 2000, 41, 5607-5611; c) R. R. Schrock, J. Am. Chem. Soc. 1976, 98, 5399-5400. [26] a) J. E. McMurry, Chem. Rev. 1989, 89, 1513-1524; b) M. Ephritikhine, Chem. Commun. 1998, 2549-2554. [27] G. Wittig, W. Haag, Chem. Ber. 1955, 88, 1654-1666. [28] a) P. A. Byrne, D. G. Gilheany, Chem. Soc. Rev. 2013, 42, 6670-6696; b) O. I. Kolodiazhny, Phosphorus Ylides: Chemistry and Application in Organic Syn- thesis, Wiley, New York, 1999. [29] H. Schmidbaur, W. Bucher, D. Schentzow, Chem. Ber. 1973, 106, 1251-1255. [30] a) A. Streitwieser, A. Rajca, R. S. McDowell, R. Glaser, J. Am. Chem. Soc. 1987, 109, 4184-4188; b) S. M. Bachrach, J. Org. Chem. 1992, 57, 4367-4373; c) D. G. Gilheany, Chem. Rev. 1994, 94, 1339-1374. [31] a) P. A. Byrne, D. G. Gilheany, J. Am. Chem. Soc. 2012, 134, 9225-9239; b) B. E. Maryanoff, A. B. Reitz, Chem. Rev. 1989, 89, 863-927. [32] J. Mai, Bridging McMurry and Wittig in One-Pot – Olefins from Stereoselec- tive, Reductive Couplings of Two Aldehydes via Phosphaalkenes, Acta Univer- sitatis Upsaliensis, Uppsala, 2019. [33] G. Wittig, A. Haag, Chem. Ber. 1963, 96, 1535-1543. [34] a) E. Vedejs, K. A. J. Snoble, J. Am. Chem. Soc. 1973, 95, 5778-5780; b) E. Vedejs, C. F. Marth, J. Am. Chem. Soc. 1990, 112, 3905-3909. [35] a) R. Hoeller, H. Lischka, J. Am. Chem. Soc. 1980, 102, 4632-4635; b) F. Vol- atron, O. Eisenstein, J. Am. Chem. Soc. 1984, 106, 6117-6119; c) H. Yamataka, S. Nagase, J. Am. Chem. Soc. 1998, 120, 7530-7536. [36] a) E. Vedejs, G. P. Meier, K. A. J. Snoble, J. Am. Chem. Soc. 1981, 103, 2823- 2831; b) E. Vedejs, C. F. Marth, J. Am. Chem. Soc. 1989, 111, 1519-1520. [37] M. Schlosser, H. B. Tuong, Angew. Chem., Int. Ed. 1979, 18, 633-634.

104 [38] a) M. Schlosser, K. F. Christmann, Liebigs Ann. Chem. 1967, 708, 1-35; b) M. Schlosser, K.-F. Christmann, A. Piskala, Chem. Ber. 1970, 103, 2814-2820. [39] G. L. Parker, L. K. Smith, I. R. Baxendale, Tetrahedron 2016, 72, 1645-1652. [40] P. A. Byrne, K. V. Rajendran, J. Muldoon, D. G. Gilheany, Org. Biomol. Chem. 2012, 10, 3531-3537. [41] a) L. Horner, H. Hoffmann, H. G. Wippel, Chem. Ber. 1958, 91, 61-63; b) L. Horner, H. Hoffmann, H. G. Wippel, G. Klahre, Chem. Ber. 1959, 92, 2499- 2505. [42] W. S. Wadsworth, Jr. Org. React. 1977, 25, 73. [43] J. Clayden, S. Warren, Angew. Chem., Int. Ed. 1996, 35, 241-270. [44] A. K. Bhattacharya, G. Thyagarajan, Chem. Rev. 1981, 81, 415-430. [45] F. A. Carey, R. J. Sundberg, Advanced Organic Chemistry part B: Reactions and Synthesis, 5th ed., Springer, New York, 2008. [46] S. K. Thompson, C. H. Heathcock, J. Org. Chem. 1990, 55, 3386-3388. [47] W. Y. Siau, Y. Zhang, Y. Zhao, Stereoselective Alkene Synthesis: Stereoselec- tive Synthesis of Z-Alkenes, Springer Berlin Heidelberg, Berlin, Heidelberg, 2012. [48] W. C. Still, C. Gennari, Tetrahedron Lett. 1983, 24, 4405-4408. [49] a) K. Ando, Tetrahedron Lett. 1995, 36, 4105-4108; b) K. Ando, J. Org. Chem. 1997, 62, 1934-1939. [50] a) E. J. Corey, G. T. Kwiatkowski, J. Am. Chem. Soc. 1966, 88, 5652-5653; b) E. J. Corey, G. T. Kwiatkowski, J. Am. Chem. Soc. 1968, 90, 6816-6821; c) E. J. Corey, G. T. Kwiatkowski, J. Am. Chem. Soc. 1966, 88, 5653-5654. [51] E. J. Corey, G. T. Kwiatkowski, J. Am. Chem. Soc. 1966, 88, 5654-5656. [52] a) P. J. Kocienski, B. Lythgoe, D. A. Roberts, J. Chem. Soc., Perkin Trans. 1978, 1, 834-837; b) P. J. Kocienski, B. Lythgoe, I. Waterhouse, J. Chem. Soc., Perkin Trans. 1980, 1, 1045-1050. [53] P. R. Blakemore, J. Chem. Soc., Perkin Trans. 2002, 1, 2563-2585. [54] a) A. S. Kende, J. S. Mendoza, Tetrahedron Lett. 1990, 31, 7105-7108; b) G. E. Keck, K. A. Savin, M. A. Weglarz, J. Org. Chem. 1995, 60, 3194-3204; c) I. E. Markó, F. Murphy, S. Dolan, Tetrahedron Lett. 1996, 37, 2089-2092; d) I. E. Markó, F. Murphy, L. Kumps, A. Ates, R. Touillaux, D. Craig, S. Carbal- lares, S. Dolan, Tetrahedron 2001, 57, 2609-2619. [55] J. B. Baudin, G. Hareau, S. A. Julia, O. Ruel, Tetrahedron Lett. 1991, 32, 1175- 1178. [56] P. R. Blakemore, W. J. Cole, P. J. Kocieński, A. Morley, Synlett 1998, 1998, 26-28. [57] C. Aïssa, J. Org. Chem. 2006, 71, 360-363. [58] a) P. J. Kocienski, A. Bell, P. R. Blakemore, Synlett 2000, 2000, 365-366; b) D. A. Alonso, M. Fuensanta, C. Nájera, M. Varea, J. Org. Chem. 2005, 70, 6404-6416. [59] a) G. Kim, M. Y. Chu-Moyer, S. J. Danishefsky, G. K. Schulte, J. Am. Chem. Soc. 1993, 115, 30-39; b) A. B. Smith, B. M. Brandt, Org. Lett. 2001, 3, 1685- 1688; c) P. Liu, E. N. Jacobsen, J. Am. Chem. Soc. 2001, 123, 10772-10773. [60] a) D. J. Ager, Synthesis 1984, 1984, 384-398; b) A. G. M. Barrett, J. M. Hill, E. M. Wallace, J. A. Flygare, Synlett 1991, 1991, 764-770. [61] P. F. Hudrlik, D. Peterson, J. Am. Chem. Soc. 1975, 97, 1464-1468. [62] P. F. Hudrlik, E. L. O. Agwaramgbo, A. M. Hudrlik, J. Org. Chem. 1989, 54, 5613-5618. [63] a) A. G. M. Barrett, J. A. Flygare, J. Org. Chem. 1991, 56, 638-642; b) L. Duhamel, J. Gralak, A. Bouyanzer, J. Chem. Soc., Chem. Comm. 1993, 1763- 1765.

105 [64] a) K. Utimoto, M. Obayashi, H. Nozaki, J. Org. Chem. 1976, 41, 2940-2941; b) P. F. Hudrlik, A. M. Hudrlik, R. N. Misra, D. Peterson, G. P. Withers, A. K. Kulkarni, J. Org. Chem. 1980, 45, 4444-4448; c) J.-N. Heo, E. B. Holson, W. R. Roush, Org. Lett. 2003, 5, 1697-1700. [65] a) K. Shimoji, H. Taguchi, K. Oshima, K. Yamamoto, H. Nozaki, J. Am. Chem. Soc. 1974, 96, 1620-1621; b) J.-M. Galano, G. Audran, H. Monti, Tetrahedron Lett. 2001, 42, 6125-6128. [66] a) S. E. Denmark, S.-M. Yang, J. Am. Chem. Soc. 2002, 124, 15196-15197; b) A. Toró, P. Nowak, P. Deslongchamps, J. Am. Chem. Soc. 2000, 122, 4526- 4527. [67] M. B. Gillies, J. E. Tønder, D. Tanner, P.-O. Norrby, J. Org. Chem. 2002, 67, 7378-7388. [68] Z. Wang, Comprehensive Organic Name Reactions and Reagents, John Wiley & Sons, Inc., 2010. [69] M. P. Washington, V. B. Gudimetla, F. L. Laughlin, N. Deligonul, S. He, J. L. Payton, M. C. Simpson, J. D. Protasiewicz, J. Am. Chem. Soc. 2010, 132, 4566- 4567. [70] J. E. Huheey, E. A. Keiter, R. L. Keiter, Inorganic Chemistry: Principles of Structure and Reactivity, 4th ed., HerperCollins, New York, 1993. [71] R. L. DeKock, H. B. Gray, Chemical Structure and Bonding, 2nd ed., Univer- sity Science Books, Sausalita, USA, 1989. [72] J. Waluk, H. P. Klein, A. J. Ashe, J. Michl, Organometallics 1989, 8, 2804- 2808. [73] a) K. S. Pitzer, J. Am. Chem. Soc. 1948, 70, 2140-2145; b) R. S. Mulliken, J. Am. Chem. Soc. 1950, 72, 4493-4503; c) R. S. Mulliken, J. Am. Chem. Soc. 1955, 77, 884-887. [74] M. Yoshifuji, I. Shima, N. Inamoto, K. Hirotsu, T. Higuchi, J. Am. Chem. Soc. 1981, 103, 4587-4589. [75] a) M. Regitz, Chem. Rev. 1990, 90, 191-213; b) G. Becker, G. Gresser, W. Uhl, Z. Naturforsch 1981, 36, 16-19. [76] C. N. Smit, F. M. Lock, F. Bickelhaupt, Tetrahedron Lett. 1984, 25, 3011- 3014. [77] a) A. Rolf, F. Peter, J. Bernhard, S. Mustafa, W. Volker, K. Falk, Angew. Chem., Int. Ed. 1984, 23, 619-620; b) M. Yoshifuji, K. Toyota, K. Shibayama, N. Inamoto, Tetrahedron Lett. 1984, 25, 1809-1812. [78] W. W. Schoeller, Chem. Commun. 1985, 334-335. [79] S. Lacombe, D. Gonbeau, J. L. Cabioch, B. Pellerin, J. M. Denis, G. Pfister- Guillouzo, J. Am. Chem. Soc. 1988, 110, 6964-6967. [80] a) M. W. Schmidt, P. N. Truong, M. S. Gordon, J. Am. Chem. Soc. 1987, 109, 5217-5227; b) P. v. R. Schleyer, D. Kost, J. Am. Chem. Soc. 1988, 110, 2105- 2109. [81] a) M. C. Simpson, J. D. Protasiewicz, Pure Appl. Chem. 2013, 85, 801-815; b) D. Nauroozi, A. Orthaber, Eur. J. Inorg. Chem. 2016, 2016, 709-717. [82] T. Baumgartner, R. Réau, Chem. Rev. 2006, 106, 4681-4727. [83] T. C. Klebach, R. Lourens, F. Bickelhaupt, J. Am. Chem. Soc. 1978, 100, 4886- 4888. [84] a) G. Märkl, Angew. Chem., Int. Ed. 1966, 5, 846-847; b) A. J. Ashe, J. Am. Chem. Soc. 1971, 93, 3293-3295. [85] a) O. J. Scherer, Angew. Chem., Int. Ed. 1985, 24, 924-943; b) J. F. Nixon, Chem. Rev. 1988, 88, 1327-1362. [86] T. C. Klebach, R. Lourens, F. Bickelhaupt, C. H. Stam, A. Van Herk, J. Organomet. Chem. 1981, 210, 211-221.

106 [87] I.-P. Lorenz, W. Pohl, H. Nöth, M. Schmidt, J. Organomet. Chem. 1994, 475, 211-221. [88] M. Yoshifuji, J. Organomet. Chem. 2000, 611, 210-216. [89] a) K. Esfandiarfard, A. I. Arkhypchuk, A. Orthaber, S. Ott, Dalton Trans. 2016, 45, 2201-2207; b) A. H. Cowley, J. E. Kilduff, T. H. Newman, M. Pakul- ski, J. Am. Chem. Soc. 1982, 104, 5820-5821; c) Y. Masaaki, T. Kozo, S. Kat- suhiro, I. Naoki, Chem. Lett. 1983, 12, 1653-1656; d) R. Appel, F. Knoch, H. Kunze, Angew. Chem., Int. Ed. 1984, 23, 157-158; e) R. Knorr, C. Pires, C. Behringer, T. Menke, J. Freudenreich, E. C. Rossmann, P. Böhrer, J. Am. Chem. Soc. 2006, 128, 14845-14853. [90] M. Yoshifuji, J. Chem. Soc., Dalton Trans. 1998, 3343-3350. [91] a) R. Appel, F. Knoll, I. Ruppert, Angew. Chem., Int. Ed. 1981, 20, 731-744; b) M. Yam, J. H. Chong, C.-W. Tsang, B. O. Patrick, A. E. Lam, D. P. Gates, Inorg. Chem. 2006, 45, 5225-5234; c) L. Weber, Eur. J. Inorg. Chem. 2000, 2000, 2425-2441; d) L. N. Markovski, V. D. Romanenko, Tetrahedron 1989, 45, 6019-6090; e) A. C. Gaumont, J. M. Denis, Chem. Rev. 1994, 94, 1413- 1439. [92] a) A. Jouaiti, M. Geoffroy, G. Bernardinelli, Tetrahedron Lett. 1992, 33, 5071- 5074; b) A. Jouaiti, M. Geoffroy, G. Bernardinelli, Tetrahedron Lett. 1993, 34, 3413-3416. [93] J. I. Bates, B. O. Patrick, D. P. Gates, New J. Chem. 2010, 34, 1660-1666. [94] a) S. C. Serin, F. S. Pick, G. R. Dake, D. P. Gates, Inorg. Chem. 2016, 55, 6670-6678; b) B. W. Rawe, C. M. Brown, M. R. MacKinnon, B. O. Patrick, G. J. Bodwell, D. P. Gates, Organometallics 2017, 36, 2520-2526. [95] a) R. Appel, B. Laubach, M. Siray, Tetrahedron Lett. 1984, 25, 4447-4448; b) K. Takeuchi, D. W. Stephan, Chem. Commun. 2012, 48, 11304-11306. [96] a) S. M. Cornet, K. B. Dillon, A. E. Goeta, J. A. K. Howard, M. D. Roden, A. L. Thompson, J. Organomet. Chem. 2005, 690, 3630-3637; b) B. Schäfer, E. Öberg, M. Kritikos, S. Ott, Angew. Chem., Int. Ed. 2008, 47, 8228-8231; c) H. Miyake, T. Sasamori, N. Tokitoh, Angew. Chem., Int. Ed. 2012, 51, 3458- 3461. [97] a) K. Toyota, S. Kawasaki, M. Yoshifuji, J. Org. Chem. 2004, 69, 5065-5070; b) M. van der Sluis, J. B. M. Wit, F. Bickelhaupt, Organometallics 1996, 15, 174-180. [98] a) R. Appel, J. Peters, A. Westerhaus, Tetrahedron Lett. 1981, 22, 4957-4960; b) P. Pellon, J. Hamelin, Tetrahedron Lett. 1986, 27, 5611-5614; c) A. Rolf, W. Axel, Angew. Chem., Int. Ed. 1980, 19, 556-557. [99] R. Appel, M. Immenkeppel, Z. anorg. allg. Chem. 1987, 553, 7 - 14. [100] V. D. Romanenko, A. V. Ruban, M. I. Povolotskii, L. K. Polyachenko, L. N. Markovskii, J. Gen. Chem. USSR 1986, 56, 1044. [101] a) M. Yam, C.-W. Tsang, D. P. Gates, Inorg. Chem. 2004, 43, 3719-3723; b) Y. A. Veits, N. B. Karlstedt, I. P. Beletskaya, Tetrahedron Lett. 1995, 36, 4121-4124. [102] a) D. E. Hibbs, C. Jones, A. F. Richards, J. Chem. Soc., Dalton Trans. 1999, 3531-3532; b) J. Renner, U. Bergsträßer, P. Binger, M. Regitz, Eur. J. Inorg. Chem. 2000, 2000, 2337-2340. [103] a) H. Aktas, J. C. Slootweg, M. Schakel, A. W. Ehlers, M. Lutz, A. L. Spek, K. Lammertsma, J. Am. Chem. Soc. 2009, 131, 6666-6667; b) T. L. Breen, D. W. Stephan, J. Am. Chem. Soc. 1995, 117, 11914-11921. [104] S. Shah, J. D. Protasiewicz, Coord. Chem. Rev. 2000, 210, 181-201.

107 [105] a) P. Le Floch, A. Marinetti, L. Ricard, F. Mathey, J. Am. Chem. Soc. 1990, 112, 2407-2410; b) A. Marinetti, S. Bauer, L. Ricard, F. Mathey, Organome- tallics 1990, 9, 793-798; c) A. Marinetti, L. Ricard, F. Mathey, Organometal- lics 1990, 9, 788-793; d) A. Marinetti, F. Mathey, Angew. Chem., Int. Ed. 1988, 27, 1382-1384. [106] a) S. Shah, J. D. Protasiewicz, Chem. Commun. 1998, 1585-1586; b) S. Shah, T. Concolino, A. L. Rheingold, J. D. Protasiewicz, Inorg. Chem. 2000, 39, 3860-3867. [107] a) A. I. Arkhypchuk, Y. V. Svyaschenko, A. Orthaber, S. Ott, Angew. Chem., Int. Ed. 2013, 52, 6484-6487; b) A. I. Arkhypchuk, M.-P. Santoni, S. Ott, Or- ganometallics 2012, 31, 1118-1126. [108] M. Yoshifuji, K. Toyota, N. Inamoto, Tetrahedron Lett. 1985, 26, 1727-1730. [109] a) M. Abbari, P. Cosquer, F. Tonnard, Y. Y. C. Y. L. Ko, R. Carrié, Tetrahed- ron 1991, 47, 71-82; b) T. A. Van Der Knaap, T. C. Klebach, F. Visser, R. Lourens, F. Bickelhaupt, Tetrahedron 1984, 40, 991-997. [110] T. V. D. Does, F. Bickelhaupt, Phosphorus, Sulfur Silicon Relat. Elem. 1990, 49-50, 281-284. [111] T. A. Van der Knaap, T. C. Klebach, R. Lourens, M. Vos, F. Bickelhaupt, J. Am. Chem. Soc. 1983, 105, 4026-4032. [112] B. H. Gillon, K. J. T. Noonan, B. Feldscher, J. M. Wissenz, Z. M. Kam, T. Hsieh, J. J. Kingsley, J. I. Bates, D. P. Gates, Can. J. Chem. 2007, 85, 1045- 1052. [113] S. Wang, K. Samedov, S. C. Serin, D. P. Gates, Eur. J. Inorg. Chem. 2016, 2016, 4144-4151. [114] M. Regitz, Angew. Chem. 1975, 87, 259 - 268. [115] M. Regitz, G. Maas, in Short-lived phosphorus(V) compounds having coordination number 3, Springer Berlin Heidelberg, Berlin, Heidelberg, 1981, pp. 71-120. [116] a) M. Regitz, H. Scherer, W. Illger, H. Eckes, Angew. Chem., Int. Ed. 1973, 12, 1010-1011; b) H. Eckes, M. Regitz, Tetrahedron Lett. 1975, 16, 447-450; c) M. Regitz, W. Illger, G. Maas, Chem. Ber. 1978, 111, 705-725; d) M. Regitz, H. Eckes, Chem. Ber. 1980, 113, 3303-3312; e) M. Regitz, H. Eckes, Tetrahe- dron 1981, 37, 1039-1044. [117] a) R. C. Smith, X. Chen, J. D. Protasiewicz, Inorg. Chem. 2003, 42, 5468- 5470; b) S. P. W., S. S. C., K. Ivo, H. T. W., G. D. P., Angew. Chem., Int. Ed. 2013, 52, 6967-6970; c) K. Dück, B. W. Rawe, M. R. Scott, D. P. Gates, An- gew. Chem., Int. Ed. 2017, 56, 9507-9511. [118] A. Orthaber, H. Löfås, E. Öberg, A. Grigoriev, A. Wallner, S. H. M. Jafri, M.- P. Santoni, R. Ahuja, K. Leifer, H. Ottosson, S. Ott, Angew. Chem., Int. Ed. 2015, 54, 10634-10638. [119] X.-L. Geng, Q. Hu, B. Schäfer, S. Ott, Org. Lett. 2010, 12, 692-695. [120] a) T. Mukaiyama, T. Sato, J. Hanna, Chem. Lett. 1973, 2, 1041-1044; b) S. Tyrlik, I. Wołochowicz, Bull. Soc. Chim. Fr. 1973, 2147. [121] J. E. McMurry, M. P. Fleming, K. L. Kees, L. R. Krepski, J. Org. Chem. 1978, 43, 3255-3266. [122] A. Fürstner, B. Bogdanović, Angew. Chem., Int. Ed. 1996, 35, 2442-2469. [123] S. Gauthier, J. Mailhot, F. Labrie, J. Org. Chem. 1996, 61, 3890-3893. [124] a) T. Eguchi, K. Arakawa, T. Terachi, K. Kakinuma, J. Org. Chem. 1997, 62, 1924-1933; b) T. Eguchi, K. Ibaragi, K. Kakinuma, J. Org. Chem. 1998, 63, 2689-2698.

108 [125] a) E. J. Corey, R. L. Danheiser, S. Chandrasekaran, J. Org. Chem. 1976, 41, 260-265; b) N. Balu, S. K. Nayak, A. Banerji, J. Am. Chem. Soc. 1996, 118, 5932-5937. [126] J. E. McMurry, M. Silvestri, J. Org. Chem. 1975, 40, 2687-2688. [127] J. E. McMurry, M. P. Fleming, J. Org. Chem. 1975, 40, 2555-2556. [128] G. H. Timms, E. Wildsmith, Tetrahedron Lett. 1971, 12, 195-198. [129] T. L. Ho, C. M. Wong, Synth. Comm. 1973, 3, 37-38. [130] J. E. McMurry, J. Melton, J. Org. Chem. 1973, 38, 4367-4373. [131] J. E. McMurry, L. R. Krepski, J. Org. Chem. 1976, 41, 3929-3930. [132] T. Takeda, A. Tsubouchi, Organic reactions, Vol. 82, (Denmark, S. E.) ed., John Wiley & Sons, Inc., 2013. [133] K. Esfandiarfard, J. Mai, S. Ott, J. Am. Chem. Soc. 2017, 139, 2940-2943. [134] a) A. Christiansen, C. Li, M. Garland, D. Selent, R. Ludwig, A. Spannenberg, W. Baumann, R. Franke, A. Börner, Eur. J. Org. Chem. 2010, 2010, 2733- 2741; b) B. Kurscheid, W. Wiebe, B. Neumann, H.-G. Stammler, B. Hoge, Eur. J. Inorg. Chem. 2011, 2011, 5523-5529. [135] T. A. Van Der Knaap, T. C. Klebach, F. Visser, R. Lourens, F. Bickelhaupt, in Synthesis and Properties of Phosphaalkenes, Vol. 171, American Chemical Society, Washington, USA., 1981, pp. 401-404. [136] T. A. Van Der Knaap, T. C. Klebach, F. Visser, F. Bickelhaupt, P. Ros, E. J. Baerends, C. H. Stam, M. Konijn, Tetrahedron 1984, 40, 765-776. [137] L. Horner, H. Hoffmann, W. Klink, H. Ertel, V. G. Toscano, Chem. Ber. 1962, 95, 581-601. [138] G. Becker, O. Mundt, Z. anorg. allg. Chem. 1980, 462, 130-142. [139] a) A. Termaten, M. van der Sluis, F. Bickelhaupt, Eur. J. Org. Chem. 2003, 2003, 2049-2055; b) M. Yoshifuji, K. Toyota, I. Matsuda, T. Niitsu, N. Ina- moto, K. Hirotsu, T. Higuchi, Tetrahedron 1988, 44, 1363-1367. [140] R. Appel, B. Laubach, Tetrahedron Lett. 1980, 21, 2497-2500. [141] B. H. Gillon, D. P. Gates, Chem. Commun. 2004, 1868-1869. [142] a) F. Uhlig, R. Hummeltenberg, J. Organomet. Chem. 1993, 452, C9-C10; b) U. Englich, K. Hassler, K. Ruhlandt-Senge, F. Uhlig, Inorg. Chem. 1998, 37, 3532-3537. [143] T. Huang, T. Chen, L.-B. Han, J. Org. Chem. 2018, 83, 2959-2965. [144] A. I. Arkhypchuk, N. D’Imperio, S. Ott, Org. Lett. 2018, 20, 5086-5089. [145] R. Csuk, S. Albert, B. Siewert, S. Schwarz, Eur. J. Med. Chem. 2012, 54, 669- 678. [146] Q.-Q. Li, Z. Shah, J.-P. Qu, Y.-B. Kang, J. Org. Chem. 2018, 83, 296-302. [147] D. McCormack, D. McFadden, J. Surg. Res. 2012, 173, e53-e61. [148] a) I. N. Ioffe, A. A. Granovsky, J. Chem. Theory Comput. 2013, 9, 4973-4990; b) J. Jiang, C. Zheng, K. Zhu, J. Liu, N. Sun, C. Wang, H. Jiang, J. Zhu, C. Luo, Y. Zhou, J. Med. Chem. 2015, 58, 2538-2546; c) M. Cushman, D. Na- garathnam, D. Gopal, A. K. Chakraborti, C. M. Lin, E. Hamel, J. Med. Chem. 1991, 34, 2579-2588; d) M. Cushman, D. Nagarathnam, D. Gopal, H. M. He, C. M. Lin, E. Hamel, J. Med. Chem. 1992, 35, 2293-2306. [149] A. B. Juan, R. O. Liliana, Curr. Org. Chem. 2015, 19, 744-775. [150] a) H. Yamataka, K. Nagareda, K. Ando, T. Hanafusa, J. Org. Chem. 1992, 57, 2865-2869; b) G. L. Keldsen, W. E. McEwen, J. Am. Chem. Soc. 1978, 100, 7312-7317; c) E. C. Dunne, É. J. Coyne, P. B. Crowley, D. G. Gilheany, Tet- rahedron Lett. 2002, 43, 2449-2453; d) P. A. Byrne, L. J. Higham, P. McGov- ern, D. G. Gilheany, Tetrahedron Lett. 2012, 53, 6701-6704.

109 [151] a) X.-p. Zhang, M. Schlosser, Tetrahedron Lett. 1993, 34, 1925-1928; b) H. D. C., G. I. L., N. Lana, Angew. Chem., Int. Ed. 2006, 45, 2242-2245; c) D. C. Harrowven, I. L. Guy, M. Howell, G. Packham, Synlett 2006, 2006, 2977- 2980. [152] A. I. Arkhypchuk, N. D'Imperio, S. Ott, Chem. Commun. 2019, 55, 6030-6033. [153] R. F. Heck, J. Am. Chem. Soc. 1968, 90, 5518-5526. [154] O. M. Ogba, N. C. Warner, D. J. O’Leary, R. H. Grubbs, Chem. Soc. Rev. 2018, 47, 4510-4544. [155] A. B. Flynn, W. W. Ogilvie, Chem. Rev. 2007, 107, 4698-4745. [156] a) Y. Al Jasem, R. El-Esawi, T. Thiemann, J. Chem. Res. 2014, 38, 453-463; b) D. Hérault, D. H. Nguyen, D. Nuel, G. Buono, Chem. Soc. Rev. 2015, 44, 2508-2528.

110

Acta Universitatis Upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1962 Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science and Technology, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology. (Prior to January, 2005, the series was published under the title “Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology”.)

ACTA UNIVERSITATIS UPSALIENSIS Distribution: publications.uu.se UPPSALA urn:nbn:se:uu:diva-419124 2020