!  "    #

  

 !"! ! # $%& $! "" %%! ' ( ) * ))(() $$&+&           8!"                

          

   %( DT7I

                                                                                                   †                                                                                                                            

       †‡r r‚pr‡r †                Gvthq†Pyrsv†      

 

         Bear in mind that the wonderful things you learn in your schools are the work of many generations. All this is put in your hands as your inheritance in order that you may receive it, honor it, add to it, and, one day, faithfully hand it on to your children. Albert Einstein

Till Petra och Zion Scientific advisor

Prof. Pher G. Andersson Department of Biochemistry and Organic Chemistry Uppsala University

Faculty opponent

Prof. Lise-Lotte Gundersen Department of Chemistry University of Oslo

Examination committee

Dr. Sverker Hansson AstraZeneca AB Södertälje

Prof. Mats Larhed Department of Medicinal Chemistry Organic Pharmaceutical Chemistry Uppsala University

Prof. David Tanner Department of Chemistry Organic Chemistry Technical University of Denmark

Dr. Jan Vågberg iNovacia AB Stockholm

Prof. Björn Åkermark Department of Organic Chemistry Stockholm University

List of Papers

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

I Mattias Engman; Jarle Diesen; Alexander Paptchikhine; Pher G. Andersson. Iridium-Catalyzed Asymmetric Hydrogenation of Fluori- nated Olefins Using N,P-Ligands: A Struggle with Hydrogenolysis and Selectivity. Journal of the American Chemical Society, 2007, 129, 4536–4537.

II Päivi Kaukoranta; Mattias Engman; Christian Hedberg; Jonas Ber- gquist; Pher G. Andersson. Iridium Catalysts with Chiral Imidazole- Phosphine Ligands for Asymmetric Hydrogenation of Vinyl Fluorides and other Olefins. Advanced Synthesis & Catalysis, 2008, 350, 1168–1176.

III Mattias Engman; Pradeep Cheruku; Päivi Tolstoy; Jonas Bergquist; Sebastian F. Völker; Pher G. Andersson. Highly Selective Iridium- Catalyzed Asymmetric Hydrogenation of Trifluoromethyl Olefins: A New Route to Trifluoromethyl-Bearing Stereocenters. Advanced Synthesis & Catalysis, 2009, 351, 375–378.

IV Päivi Tolstoy; Mattias Engman; Alexander Paptchikhine; Jonas Ber- gquist; Tamara L. Church; Abby W.-M. Leung; Pher G. Andersson. Irid- ium-Catalyzed Asymmetric Hydrogenation yielding Chiral Diarylmethi- nes with either Coordinating or Non-Coordinating Substituents (Manuscript)

Reprints were made with permission from the publishers.

Contribution Report

The author wishes to clarify his contribution to the papers I–IV in the thesis

I Performed a significant part of the experimental work and the develop- ment of new methods for chiral separation of the products; contributed significantly to interpreting the results and writing the paper.

II Contributed significantly to the experimental work, development of separation methods and interpretation of the results mainly concerning fluorine substrates; contributed in the preparation of the manuscript.

III Performed a significant part of the experimental work; developed a ma- jority of the methods needed for separation of products; contributed sig- nificantly to the interpretation of the results and partly to the writing of the paper.

IV Performed a significant part of the experimental work; contributed sig- nificantly to separating the products, interpreting the results and prepar- ing the manuscript.

Contents

1 Introduction ...... 11 1.1 Chirality...... 11 1.2 Asymmetric hydrogenation ...... 13 2 Asymmetric Hydrogenation using Iridium Complexes — Focus on Fluorine-Containing Substrates (Paper I and II)...... 15 2.1 Introduction ...... 15 2.2 Catalysts...... 18 2.3 Olefin synthesis...... 19 2.4 Hydrogenations ...... 22 2.4.1 Hydrogenolysis: Defluorination...... 22 2.4.2 Hydrogenation of other substrates ...... 25 2.4.3 Evaluation of catalysts with imidazole-based ligands for the hydrogenation of fluoroolefins...... 29 2.5 Hydrogenation of <-chlorocinnamic ester and the corresponding alcohol ...... 34 2.6 Further studies on the new catalyst (R)-II...... 35 3 Hydrogenation of trifluoromethyl-substituted olefins (Paper III) ...... 37 3.1 Introduction ...... 37 3.2 Hydrogenation...... 38 4 Formation of 1,1-diarylmethine stereocenters (Paper IV)...... 45 4.1 Introduction ...... 45 4.2 Hydrogenation...... 46 4.2.1 Unfunctionalized olefins...... 48 4.2.2 Esters, alcohols and acetates ...... 50 5 Enantiodiscrimination...... 52 5.1 Introduction ...... 52 5.2 Fluorine- and trifluoromethyl-substituted olefins...... 54 5.3 1,1-Diaryl olefins ...... 57 6 Conclusions and Outlook ...... 59 Acknowledgements ...... 61 Summary in Swedish...... 63 References...... 65

Abbreviations

* Chiral center Abs. Conf. Absolute configuration Ac Acetyl Ar Aryl - BArF Tetrakis[3,5-bis(trifluoromethyl)phenyl]borate B3LYP Becke’s 3 parameter hybrid functional using the Lee- Yang-Parr correlation functional B-DM Astec CHIRALDEX -cyclodextrin (dimethyl) column Caro’s acid Peroxysulfuric acid, H2SO5 Cat. Catalyst COD Cyclooctadiene Conv. Conversion Cy Cyclohexyl c-EWG Conjugated electron-withdrawing group de Diastereomeric excess DCM Dichloromethane DFT Density functional theory DIBAL Diisobutylaluminium hydride DNA Deoxyribonucleic acid ee Enantiomeric excess Et Ethyl EWG Electron-withdrawing group GC Gas chromatography h Hour(s) HPLC High performance liquid chromatography HWE Horner-Wadsworth-Emmons (reaction) i-Pr Isopropyl LACVP Los Alamos effective core valence potential LCD Liquid crystal display mCPBA meta-Chloroperoxybenzoic acid Me Methyl MS Mass spectrometer n-BuLi n-Butyllithium NCS N-Chlorosuccinimide NMR Nuclear magnetic resonance NOE Nuclear overhauser effect

(N,P)* Chiral N,P ligand o.n. Overnight Pd/C Palladium on carbon Ph Phenyl rac Racemic r.t. Room temperature RNA Ribonucleic acid THF Tetrahydrofuran Å Ångström

Catalyst numbering:

Ph Ph Ph Ph Ph BArF Ph BAr BArF P F R P Ir P N Ir O Ir N N N Ph Ph Ph Ph S S O II III I a: R = Me b: R = Bn

Ar Ar Ar Ar BArF Ar Ar BArF BArF P P P Ir Ir Ir N N N N R R Ph N O R' S VI IV V a: Ar = o-tol, R = i-Pr, R' = H a: Ar = Ph, R = H a: Ar = Ph b: Ar = Ph, R = Ph, R' = Ph b: Ar = 3,5-dimethylphenyl b: Ar = Ph, R = Ph c: Ar = Ph, R = i-Pr, R' = H c: Ar = o-tol c: Ar = o-tol, R = 3,5-dimethylphenyl d: Ar = Cy, R = Ph, R' = Ph

1 Introduction

1.1 Chirality

A molecule containing a carbon with four different substituents can occur in two forms that are mirror images of each other. These will have the same physical properties, such as boiling point, molecular weight, hardness, color etc. However, since the beginning of the 19th century scientists have ob- served that these so-called enantiomers also have different properties. For example, Louis Pasteur separated the enantiomers of sodium ammonium tartrate manually by a pair of tweezers and studied how they rotated plane- polarized light in different directions.1 In the 1960s, we learned more about how two enantiomers interact differently with our world, when the effects of thalidomide (named Neurosedyn on the Swedish market) were revealed.2,3 Children whose mothers had been taking thalidomide to alleviate morning sickness and to help them sleep during pregnancy were born with malformed extremities. Investigations revealed that one of the two mirror forms had the wanted therapeutic effect, and the other one caused the toxic effects (Figure 1) In the same way, the two forms of limonene taste and smell different. (R)- Limonene smells like orange and (S)-limonene like lemon. The idea that a similar molecule can have different effects in the human body might sound strange, but is not more complicated than the right hand fitting better in a right glove.

Figure 1. The therapeutic (R)-thalidomide (white) and its harmful mirror image, (S)- thalidomide (black).

11 It is clearly important to have control over, and perhaps even take advantage of, the different behavior of the mirror forms of a molecule. Therefore, we greatly need methods of getting hold of them in their pure forms. The case of Neurosedyn, is a poor argument for the importance of asymmetric synthesis because (R)-thalidomide racemizes in vivo, but it is a perfect example of the different properties of two enantiomers. In many cases, enentiopure com- pounds are required and these can be achieved in many ways. If we synthe- size a molecule using a reaction that is not stereoselective, we get a mixture of its two forms. These can be separate by the so-called resolution of race- mates. This can be done, for example, via the recrystallisation of a di- astereomeric salt or even with chiral preparative columns. Nature’s chiral pool can provide stereochemically pure chiral materials, but there are also ways of doing reactions that produce more of one enantiomer. There are four principal ways of doing this:

• By using a chiral substrate • By using a chiral auxiliary • By using a chiral reagent • By using a chiral catalyst

Treating a single enantiomer of a starting material (a chiral substrate) with an achiral reagent may lead to a stereoselective reaction. Chirality can also be introduced by adding a chiral directing group (auxiliary) to an achiral substrate to direct the outcome of the reaction. The directing group can then be removed after the reaction. A third way of directing a reaction, in this case without a chiral starting material, is to use a chiral reagent. In all three methods above, a stoichiometric amount of chiral material is needed to direct the reaction. The fourth way of performing a stereoselective reaction is to use a chiral catalyst. This could be considered the most satisfying method of the four, as the use of hard-to-get chiral material is minimized. However, the chiral ligands needed for the chiral catalysts may be very exotic and hard to synthesize.

Now that we know that chirality is required to introduce chirality, one might wonder how to obtain the chiral substrates, auxiliaries, reagents and catalysts in the first place. Again, these must come from either Nature’s chiral pool, by resolution (separation) of racemates, or by stereoselective synthesis. To hatch a chicken, you need an egg.

12 1.2 Asymmetric hydrogenation

In the area of stereoselective synthesis, there are many types of catalysts that catalyze a lot of different reactions. The work in this thesis has focused on one type of reaction and a few different classes of catalysts, all containing iridium. These catalysts, based on N,P-ligated iridium centres, were used to add hydrogen gas to olefins. The enantioselective hydrogenation of olefins is one of the most powerful transformations in asymmetric catalysis. The reac- tion has been very well studied using complexes of chiral P,P- or N,P- chelating ligands with metals like ruthenium and rhodium.4,5 However, ru- thenium and rhodium catalysts have been limited to functionalized substrates having coordinating groups like alcohols, acids, esters or amides. In the past decade, iridium complexes have proven to be efficient catalysts for the enan- tioselective hydrogenation of both functionalized and unfunctionalized ole- fins.5-9 The iridium catalyst developed by Crabtree in 197910 was modified by Pfaltz in 1998 to give an iridium-phosphanodihydrooxazole (PHOX) catalyst that could hydrogenate tri- and tetrasubstituted olefins with enanti- omeric excesses (ee’s) up to 98% (Figure 2).11

PF6 BArF

Cy3P Ir Ar2P Ir Many new N,P-ligated catalysts N N R

O Crabtree´s catalyst Pfaltz' catalyst Figure 2. Crabtree’s catalyst was modified by Pfaltz in 1998 and seeded the design of many new ligands during the past decade.

Another important improvement on the technique revealed that the anion used for the hydrogenation catalyst is important. After the weakly coordinat- - ing anion BArF (Figure 3) was introduced to this field, catalytic loadings as low as 0.02% could sometimes be used. Since then, a constantly growing pool of stereoselective iridium-based, N,P-ligand catalysts have been re- ported for the hydrogenation of a limited range of substrates.12-37 As the number of efficient N,P-ligated catalysts grow, the desire among scientists to evaluate these with new types of substrates grow even faster.22,27,38-47

- CF3 F3C

F3C CF3 B CF F3C 3

F C CF3 3 - Figure 3. The structure of the better-performing anion; BArF .

13 The aim of this thesis has been to broaden the scope of useful substrate classes for the asymmetric hydrogenation of olefins using N,P-ligated irid- ium catalysts. This has resulted in many new, efficient reactions with very high enantioselectivities and conversions that yield diverse chiral products; these can be used in applications reaching all the way from building blocks for medicinal chemistry to additives in LCDs.

14 2 Asymmetric Hydrogenation using Iridium Complexes — Focus on Fluorine- Containing Substrates (Paper I and II)

"The hectic state of activity in this field and the surprises that often emerge from what should be the simplest transformations of fluorine derivatives make it hard for me to resist coining a new term: flustrates ( fluorine-containing substrates)." Dieter Seebach48

2.1 Introduction

The size of a fluorine atom and the length of a C–F bond (1.47 Å and 1.35 Å respectively) are intermediate between the properties of hydrogen (1.2 Å and 1.09 Å) and oxygen (1.52 Å and 1.43 Å), but closer to oxygen.49,50 The C–F bond is the strongest in organic chemistry (105.4 kcal mol-1, cf. C–H 98.8 kcal mol-1 or C–O 84.0 kcal mol-1). As the fluorine atom possesses the high- est electronegativity, much higher than the electronegativity of carbon, the C–F bond is highly polarized with most of the electron density on fluorine. The particular strength of the bond is attributed to electrostatic interactions between F- and C+ rather than the electron sharing of a covalent bond.50 The fluorine atom prefers to bond to sp3- rather than sp2-hybridized carbons. This is revealed, for example, in calculations on the interconversion of meth- ane and fluoroethene, in which the production of fluoromethane and ethene is thermodynamically favoured.50,51 There is a lot of evidence that the C–F bound fluorine atom can coordinate cations. The interactions are strongest with hard cations such as Li+, Na+ and K+, but extend to softer cations (e.g. Ca2+).50,52 Grubbs and co-workers have reported an acceleration effect in a fluorine-bearing metathesis catalyst due to fluorine coordinating to the ru- thenium cation.53 Therefore, we could expect fluorine coordination in other situations.

Fluorine-containing molecules have become very popular in the chemical industry lately.54,55 Up to 20% of pharmaceuticals and 30–40% of agro- chemicals are estimated to contain fluorine56 and the incorporation of fluo- rine into a molecule is considered able to almost magically change the prop- erties of a molecule for the better and “smuggling fluorine into a lead struc- ture enhances the probability of landing a hit almost 10-fold.”57 In pharma-

15 ceuticals, fluorine is used as a bioisostere for hydrogen. The exchange of hydrogen for fluorine might:58,59

• Block a metabolically unstable site or work as an enzyme in- hibitor • Affect binding affinity to a receptor • Affect basicity of adjacent basic groups, e.g. amines • Facilitate new patents – It is a new compound

In one example, 5-fluorouracil can be used to inhibit the enzyme thymidylate synthase, which converts uracil to thymine in cells. The purpose of the en- zyme is to add a methyl group to the 5-position of uracil. Adding a substrate with fluorine stationed in that position inhibits the enzyme and interrupt the synthesis of the pyrimidine thymidine, which is a nucleotide required for DNA replication (Figure 4). This in combination with other mechanisms of actions makes 5-fluorouracil a useful anti-cancer drug.60

O O F O HN HN HN O N O N H O N H H Uracil Thymine 5-fluorouracil

Nucleosides Deoxynucleosides Nucleotides Deoxynucleotides A, G, C, U A, G, C, T

RNA DNA Figure 4. 5-Fluorouracil can inhibit the enzyme thymidylate synthase and therefore decrease the conversion of uracil to thymine in the cell.

Another example in which a fluorine atom can radically change the proper- ties of a molecule is when it is placed in the vicinity of, for example, an amine. The 5-HT1D agonist 1 (Figure 5) has a good affinity for the receptor, but because of its high pKa value, it also has very low bioavailability. If one fluorine atom is added to the molecule, the pKa drops to 8.7, resulting in higher bioavailability and without lowering the receptor affinity. Adding a second fluorine affects the basicity too much, resulting in a large loss of affinity to the serotonin receptor.54

16 IC50 = 0.3 nM

pKa = 9.7 N N Very low bioavailability N N 1

N IC50 = 0.9 nM

pKa = 8.7 Medium bioavailability N N N N F N 2 IC50 = 78 nM pKa = 6.7 N N N F N F N 3 Figure 5. Fluorine can be used to tune the basicity of an amine in a 5-HT1D agonist, affecting both its bioavailability and affinity for the receptor.

There are many examples in which adding fluorine to a structure takes a molecule from useless to priceless. However, the synthetic methods avail- able for introducing fluorine-containing stereocenters are still few; most involve the asymmetric fluorination of -ketoesters or - ketophosphonates.61-68 Methods for creating a CHF-bearing stereocenter by asymmetric hydrogenation are even rarer.55 In a recent patent, Nelson et al. reported the asymmetric hydrogenation of cyclic vinyl fluorides using Rh- Walphos.69 We therefore wanted to evaluate our iridium catalysts in the asymmetric hydrogenation of vinylic fluorine compounds to explore the ability of these catalysts to create stereocenters that bear fluorine atoms. Surprisingly, the hydrogenation of fluorine-containing olefins has only been reported a few times, possibly due to the ability of vinylic fluorine to be cleaved.70-72 This might sound bizarre, as C–F bonds are usually considered very stable,50 but as discussed before, the hybridization of the fluorine atom can be of importance. It is not uncommon for a fluorine atom situated on an olefin to leave when exposed to hydrogen gas and a transition-metal cata- lyst.70-72 This was one of the problems we had to overcome.

17 2.2 Catalysts

Expanding the substrate scope of the iridium-catalyzed asymmetric hydro- genation of olefins required the use of several classes of complexes. The syntheses of catalysts I,37 II,46 III,32 IV,34,73 V,31 VI35,36 were developed in the group (Figure 6).

Ph Ph Ph Ph Ph BArF Ph BAr BArF P F R P Ir P N Ir O Ir N N N Ph Ph Ph Ph S S O II III I a: R = Me b: R = Bn

Ar Ar Ar Ar BArF Ar Ar BArF BArF P P P Ir Ir Ir N N N N R R Ph N O R' S VI IV V a: Ar = o-tol, R = i-Pr, R' = H a: Ar = Ph, R = H a: Ar = Ph b: Ar = Ph, R = Ph, R' = Ph b: Ar = 3,5-dimethylphenyl b: Ar = Ph, R = Ph c: Ar = Ph, R = i-Pr, R' = H c: Ar = o-tol c: Ar = o-tol, R = 3,5-dimethylphenyl d: Ar = Cy, R = Ph, R' = Ph Figure 6. Catalysts used in this work.

The ligands used in these catalysts vary in several ways. The heterocyclic rings can bear different substituents, though phenyl groups are used most frequently. A bulkier group, such as 3,5-dimethylphenyl or a small atom, hydrogen, can be used to vary the bulk in this part of the ligand. In the case of catalyst VI, the substituent on the oxazoline is often the moderately bulky i-Pr group. The ligand backbone, which contains the asymmetric element, differs widely between the ligand classes. Catalysts I, II, IV and V are quite similar in structure. Catalyst III also resembles these, but differs in that it has an ‘open’ (i.e. acyclic) backbone. The backbone atom that connects to the phosphorus atom (the ‘linker’), can also be varied. So far, oxygen (cata- lyst I), carbon (catalysts III–V) and nitrogen (catalysts II and VI) have been used. The substituents at phosphorus can also differ, and small changes here, for example replacing phenyl with o-tolyl substituents can sometimes have a dramatic effect on a catalyst’s properties. Therefore, expanding the substrate scope of iridium-catalyzed asymmetric hydrogenation requires the testing of a variety of catalysts with different properties. Slight changes in the ligand structure, even if they produce only small changes in the optimized geometry of the active hydrogenation complex, can have huge effects on the outcome

18 of the reaction. An overlay of the calculated optimized geometries of com- plexes I, IV, V and VI shows how similar they appear (Figure 7).

Figure 7. Two views of the optimized geometries of the active catalytic complex (B3LYP/LACVP). The activated complex of I (oxazole): white, IVb (thiazole): dark grey, Va (imidazole): grey, VIa : black. Hydrogens, iridium atom and ethene re- moved for clarity.

2.3 Olefin synthesis

There are many ways of making fluorine-containing olefins. However, any chemist that has attempted to introduce fluorine into his/her chemistry has probably noticed that the outcomes of the reactions are not always as clean and predictable as desired. Some of the synthetic routes used in this work are sketched below.

Esters, alcohols and acetates were synthesized by refluxing ethyl bro- mofluoroacetate in triethyl phosphite overnight (Scheme 1). The resulting mixture was purified by distillation and then subjected to a Horner- Wadsworth-Emmons (HWE) reaction with either acetophenone or benzalde- hyde to produce a mixture of E and Z olefins.

O O O n-BuLi O P(OEt)3 O Acetophenone Br P O DIBAL OH O O O F Reflux o.n. THF Et O F F 2 4 5 F n-BuLi E/Z- 6 E/Z- 7 Benzaldehyde THF

O O O O DIBAL O O OH O Et2O Pyridine F F F E/Z - 8 E/Z- 9 E/Z- 10 Scheme 1. Overview of the synthetic routes to esters, alcohols and acetates from an -fluoroester.

19 To synthesize the Z isomer of 8 selectively, tandem reduction-olefination was used instead of direct HWE olefination. In this approach, a benzoyl group was added to 5 and the product was isolated. The ketone carbonyl was then reduced using NaBH4, leaving the ester intact and resulting in an elimi- nation of phosphorus oxide that was Z-selective (Scheme 2). The ester was converted to alcohol and acetate as above, and all three components were evaluated as substrates for asymmetric hydrogenation.

n-BuLi O O O O O Benzoyl chloride O O NaBH4 O P O P O O THF F O EtOH F F O

511Z- 8 Scheme 2. Synthesis of isomerically pure Z-8 via tandem reduction-olefination.

We were also interested in the cinnamate ester derivative with a fluorine atom in the benzylic position, for several reasons. First, this type of substrate would have electron-withdrawing groups on each olefin terminus, rather than two on the same carbon as in substrate 8. Second, this substrate would be structurally similar to 8, and its behavior upon hydrogenation could there- fore offer hints as to the properties that are needed for successful hydrogena- tion without defluorination. Therefore, E- and Z-15 were synthesized accord- 74 ing to a literature procedure, via the radical addition of CF2Br2 to ethyl vinyl ether, oxidation, conversion to organozinc reagent, Pd(0)/Cu(I)- cocatalyzed cross-coupling with phenyliodide, and purification via standard flash chromatography (Scheme 3).

O Na2S2O4/NaHCO3 F F Caro's acid O O F F Zn CF2Br2 Br O Br O EtOH DMF 12 13

I O F F O O F O O ZnBr O Pd(PPh ) /CuI 14 3 4 F DMF Z-15 E-15 Scheme 3. The synthesis of E- and Z-15 for evaluation in asymmetric hydrogena- tion.

We planned to synthesize 17 in order to compare to the standard substrate 18. The synthesis of 17 was attempted using the Shapiro reaction with Se- lectfluorTM as an electrofile (Scheme 4). This procedure resulted in only non- isolable traces of product. A similar substrate, 29 (Table 3, entry 6), was

20 later synthesized in the group and proved inert to hydrogenation under these conditions (vide infra), so the synthesis of similar substrates discontinued.

O 18 H O N N S F n-BuLi O

O O SelectfluorTM 16 THF 17 Scheme 4. The synthesis of fluorinated substrate 17 via Shapiro reaction.

The synthesis of 24 was attempted from 23, a fluorine-containing analogue of the Julia reagent75 that was obtained from a long synthesis (Scheme 5). The synthesis yielded an E/Z mixture of 24, which provided inert to hydro- genation. The E/Z isomers of 24 were inseparable by standard flash chroma- tography and were hard to handle because of their volatility.

O Cl P N O N N O NCS ZnBr SH S S K CO S 2 3 S DCM S 3HF NEt3 19Acetone 20 21 Cl Acetonitrile N N O mCPBA O F S S S DCM S O t-BuOK 22 F F THF 23 E/Z- 24 Scheme 5. Julia synthesis to reach substrates E- and Z-24.

Trans--methylstilbene is one of the standard substrates used when evaluat- ing new catalysts for asymmetric hydrogenation. Therefore, evaluating its fluoro counterpart was of interest. The synthesis of this olefin, 26, was per- formed using diarylzinc-promoted Wittig reaction76 and a subsequent Suzuki coupling with phenylboronic acid (Scheme 6). The diastereoisomers were separated through column chromatography.

O Et Zn 2 Suzuki PPh 3 F coupling F CBr3F THF E/Z- 25 Br E/Z- 26

Scheme 6. The synthesis of E/Z--fluorostilbene 26.

21 2.4 Hydrogenations 2.4.1 Hydrogenolysis: Defluorination

As previously mentioned, the fluorine-carbon bond is considered to be very stable,50 which is often true. During metal-catalyzed hydrogenation however, vinylic, allylic, benzylic and aromatic carbon-fluorine bonds are easily cleaved (‘defluorinated’).70-72 This reaction was observed during our initial experiments. The hydrogenation of a 10:1 mixture of E:Z-8 using different catalysts (Table 1) showed that fluorine loss was highly ligand-dependent.

Table 1. Initial study of defluorination in the hydrogenations of 8.a

O O O O F O H2 (100 bar) O O F Ir catalyst (0.5 mol %) F + O +

r.t. CH2Cl2 X Y Bacicity of N in Entry Catalyst Conv. (%) X:Y ligand (pKa) 1 I 0.8 44 60:40 2 IVb 2.5 31 68:32 IVb 3 2.5 28 79:21 + acetic acidb IVb 4 2.5 25 78:22 + triethylamineb 5 Va 7 16 79:21 6 VIa 5 99 95:5 7 VIb 5 49 90:10 8 VIc 5 35 71:29 9 VId 5 46 90:10

a) Reaction conditions: 0.2–0.3 M substrate in CH2Cl2, 0.5 mol % catalyst 100 bar H2, 72 h. Conversions refer to the % of olefin converted to X or Y and were determined by 1H NMR spectroscopy; ee values were determined by chiral HPLC. b) 5 mol % of the additive (acetic acid or triethylamine).

Catalyst I displaces the fluorine quite efficiently, particularly in ,,- trifluorotoluene solution, where 62% defluorination was observed (not shown in Table 1). There are two possible routes for the fluorine displace- ment. In one, the carbon-carbon bond of the olefin is hydrogenated first, resulting in 8a, which undergoes fluorine displacement to yield 27a. In this case, long reaction times would result in more defluorination, as the product 8a would be unstable under the reaction. The other possibility is that fluorine is displaced before the hydrogenation. This would result in 27, which is eas-

22 ily reduced by the system to yield 27a. To further investigate the fluorine displacement in this particular reaction, we synthesized the racemate of 8a (using Wilkinson’s catalyst, RhCl(PPh3)3) and subjected it to hydrogenation under the conditions that produced the most defluorination: complex I in <,<,<-trifluorotoluene (Scheme 7). Only starting material could be detected after this reaction, indicating that fluorine loss does not occur via the first of the proposed paths.

O O

F Complex I H ,100 bar, 40 oC, 52 h, 8a 2 ,,-trifluorotoluene

O O O O F E/Z 8 27a

O O

E/Z 27

Scheme 7. Two possible defluorination routes and the attempted defluorination of rac-8a using catalyst I in ,,-trifluorotoluene.

Although no fluorine-free unsaturated substrate was detected following the hydrogenations of 8, this does not eliminate the defluorination- hydrogenation route; it simply requires that the reactivity of the fluorine-free olefin is much higher than that of its fluorinated counterpart. However, a concerted mechanism, like the ones proposed earlier for similar systems,70 might be operative. Either way, long reaction times clearly do not affect the defluorination ratio by allowing fluorine to be displaced from the product. Although it is hard to draw conclusions about what factors are involved in increasing the loss of fluorine, it is clear that ligands based on the 2-aza- norbornane-oxazoline backbone (Table 1, entries 6–9) give the overall high- est conversions and lowest degrees of defluorination.

A comparison of the fluorine-displacement abilities of complexes I, IVb and Va, which have very similar ligands (Tables 1 and 4), demonstrates that a higher pKa of the ligands heteroaromatic nitrogen lead to lower defluorina- tion ratio. The overall least defluorinating complexes VI have pKa-values somewhere between those of complexes I, IVb and Va which does not fol- low this trend. The ee values were also correlated to ligand pKa; this will be discussed further in Chapter 2.4.3.

23 Cleaving a C–F bond via a hydrogenation reaction would reasonably pro- duce hydrogen fluoride. Because this is an acid, we wanted to evaluate whether adding base (triethylamine) or acid (acetic acid) affects the de- fluorination in any way. In fact, adding either base or acid slightly reduced the loss of fluorine, which cannot be explained.

The different catalysts also have different atoms linking the ligand backbone and the phosphorus atom. Catalysts VI all have a nitrogen atom in this posi- tion, whereas IVb, Va and I have methylene groups and an oxygen atom, respectively. To distinguish the effects of the aza-norbornyl-backbone of complexes VI from the effects of a nitrogen linker, an oxazole or thiazole scaffold containing an amine linker could be evaluated as a ligand in this reaction. Because thiazole-based catalysts have generally performed better than the oxazole-based catalysts for olefin hydrogenation,34,37 thiazole was used as the backbone for the new amine-linked ligand (Figure 8).

Ar Ar P Ir BArF N Ph Ph Ph S P BArF IV N Ir N Ph Ar S Ar IIa P BArF Ir N

N R O R' VI Figure 8. Combining the thiazole ligand with an N-P linker resulted in a new cata- lyst for the hydrogenation of fluoroolefins.

The new catalyst IIa was tested in hydrogenation of a 10:1 mixture of E:Z-8, under the same conditions used previously. This catalyst unexpectedly pro- duced 48% defluorination and only 23% conversion, worse than all previous results in CH2Cl2. Clearly, the incorporation of an amine in the linker did not result in the desired decrease in defluorination. However, this catalyst per- formed very well in the hydrogenation of other fluorinated substrates.

24 2.4.2 Hydrogenation of other substrates

Overall, the use of iridium complexes based on the aza-norbornyl scaffold (VI) minimized defluorination. Complex VIa gave the best results in terms of the level of defluorination and conversion. It was chosen for further evalua- tion with some trisubstituted, fluorinated substrates (Table 2). The ester (en- tries 1 and 2) required high pressures (100 bar) and elevated temperatures (40 °C) to undergo hydrogenation, whereas the acetate (entries 3 and 4) and alcohol (entries 5 and 6) reacted more readily. The highest reaction rates and conversions were observed for the alcohol (entries 5 and 6). Full conversion was reached with 20 bar H2 and 0.5 mol % catalyst after 24 h at room tem- perature. The ee values varied from poor (7%, entry 2) to good (87% entry 3 and 80% entry 5). Despite its initial poor performance in the defluorination study, complex IIa, which was designed especially for the hydrogenation of fluorine-containing olefins, was extremely stereoselective when screened against the same trisubstituted substrates (ee  99%, entries 3 and 5).

Tetrasubstituted olefins have proven difficult to hydrogenate catalytically, but several of them underwent hydrogenation (100 bar, 40 °C, 72 h) with >99% diastereomeric excess (de) with these complexes (entries 7–9). Complex VIa hydrogenated a tetrasubstituted fluorinated cinnamate ester more selectively than its trisubstituted analogue (57 vs. 29 % ee, entries 7 and 1). Complex IIa was surprisingly less selective for the tetrasubstituted esters than the trisubsti- tuted ones, though it was much more selective for a tetrasubstituted allylic alcohol than for its trisubstituted analogue (entry 9 vs. entry 6). Oddly, IIa did not catalyze the reduction of Z-8 (entry 1), whereas it hydrogenated the tetrasubstituted equivalent Z-6 (entry 7) in 30% conversion.

The absolute configuration of the hydrogenated product from substrate E-7 (entry 9) was determined by comparison to literature data65 and the relative configurations of the products from the hydrogenations of E- or Z-6 (entries 7 and 8) were determined by reducing the ester to the alcohol and comparing its 1H NMR spectrum to literature data. E/Z conformations of olefins 6 and 7 were confirmed by 1H-19F NOE experiments. Combining these data showed that hydrogenations of tetrasubstituted olefins clearly proceeded via a clean syn addition of H2 across the double bond.

25 Table 2. Hydrogenation of Fluorine-Containing Olefins.a

R1 F F H2 (20-100 bar) F R1 R1 R1 Ir catalyst (0.5-2 mol %) or R R R R CH Cl 40 oC 2 2 X Y Complex VIa Complex (R)-IIa Conv. Ratio eeb Abs. Conv. Ratio eeb Abs. Entry Substrate (%) X:Y (%) Conf.b (%) X:Y (%) Conf.b

1 F Z-8 99 98:2 29 (R) 0 - - - COOEt

2c COOEt E-8 97 95:5 7 (S) 15 63:37 8 (S) F

3 F Z-10 78 88:12 87 (R) 82 95:5 >99 (R) Ac O

Ac 4d O E-10 88 95:5 34 (R) 28 89:11 10 (R) F

5 F Z-9 99 94:6 80 (R) 97 100:0 99 (R) OH

6e OH E-9 99 97:3 28 (R) 99 93:7 6 (S) F

(-)- 7 F Z-6 21 100:0 57 30 53:47 rac - (2R*,3S*) COOEt

(+)- 8 COOEt E-6 25 100:0 74 25 70:30 rac - (2S*,3S*) F

(+)- (+)- 9 OH E-7 24 71:29 90 69 91:9 82 (2S,3S) (2S,3S) F a) General conditions: 0.5–2.0 mol % catalyst, room temp to 40 °C, dry CH2Cl2, 20–100 bar 1 H2. Ratios and conversions were determined by H NMR. b) Details are given in Supporting Information. (Paper I) c) Conditions: 1 mol % catalyst, 40 °C, 100 bar H2. d) Conditions: Complex VIa, 1 mol % catalyst, r.t., 30 bar H2, 12 h; complex IIa, 1 mol % catalyst, r.t., 100 bar H2, 66 h. e) Conditions: Complex VIa, 1 mol % catalyst, r.t., 30 bar H2, 12 h; complex IIa, 1 mol % catalyst, r.t., 30 bar H2, 18 h.

26 The many good results achieved with catalysts IIa and VIa encouraged us to continue the search for new substrates. Trans--methyl-stilbene is a standard substrate commonly used in hydrogenation. Therefore, the -fluorinated analogue 26 was tested with limited success (Table 3, entries 1 and 2). Cata- lyst VIa did not reduce either the E or Z olefin at all and catalyst IIa offered very low conversion even with higher catalyst loading (1 mol %), high hy- drogen pressure (100 bar) and long reaction times (72 h). Unfortunately, this substrate was also frequently defluorinated and was hydrogenated in low ee values.

So far, the hydrogenation of esters that have fluorine alpha to the ester group, and their derivatives had been the focus of interest. Placing the fluo- rine at the benzylic position might also give good results. Apart from being hard to synthesize, the esters E- and Z-15 (Table 3, entries 3 and 4) also gave very low conversions in the hydrogenations, high loss of fluorine and very low ee values. Interestingly, they both resulted in the same enantiomer of product. The corresponding alcohol 28 (entry 5) was completely reduced, but resulted in a complex mixture. This behavior is not uncommon in the hydrogenation of allylic alcohols, as will be seen later in the hydrogenation of 1,1-diaryl olefins 75-77.

The naphthalene derivative 29 (entry 6), also synthesized because it resem- bled a standard substrate (4-methyl-1,2-dihydronaphthalene), was not hydro- genated by catalyst VIa. Substrates 30-33 (entries 7–9), all low-cost, com- mercially available substrates, were inert to hydrogenation by catalyst VIa, even under forcing conditions (100 bar, 72 h). This might be because the substrates themselves are inert to hydrogenation under these conditions, or they may interact with and deactivate the catalyst.

Racemic samples of the desired fluorinated alkanes are usually synthesized via hydrogenation using achiral catalysts like Pd/C or the Crabtree or Wil- kinson catalysts. For the fluorine-containing substrates, the first attempt to synthesize a racemic product often failed. The achiral hydrogenations of the fluorostilbenes 26 (Table 3, entries 1 and 2) were especially demanding. Using 5% Pd/C at room temperature with 1 atm H2 pressure displaced the fluorine efficiently. Crabtree’s and Wilkinson’s catalysts catalyzed neither the displacement of fluorine nor the hydrogenation of the olefin, as no reac- tion was detected after 18 h at room temperature and under 50 bar H2 pres- sure. The racemate of catalyst IIa was somewhat effective, but yielded low conversion, making the product difficult to purify.

27 Table 3. Hydrogenation of other fluorine-containing olefins.a

F H2 (100 bar) F R'' Ir catalyst (1 mol %) R'' R'' R' R' R' R R R CH2Cl2 X Y Entry Substrate Catalyst Conv. Ratio X:Y eeb (%)

VIa 0 - - 1 E-26 IIa 25 75:25 28 F VIa 0 - - 2 Z-26 F IIa 16 69:31 52

3 Z-15 IIa 2 25:75 60 F O O

O 4 O E-15 IIa 14 18:82 47 F

5 Z-28 VIa 99 Complex mixture F OH F 6 29 VIa 0 - -

F NH 7 30 VIa 0 - - O N O H F N 8 31 VIa 0 - - H2N N O H O F H N

9 N O 32 VIa 0 - -

O

F F

F F 10 F 33 VIa 0 - - F F F a) Conditions: 1 mol % catalyst, 72 h, r.t., dry CH2Cl2, 100 bar H2. Ratios and conversions were determined by 1H NMR spectroscopy. b) GC-MS (B-DM, 90 °C 30 min, 1 °C/min 130 °C, 20 °C/min 175 °C, 14.5 psi, 1.5 ml/min); entry 1: tR = 71.1 min (major), 71.3 min (minor); entry 2: tR = 71.1 min (minor), 71.3 min (major); entry 3: tR = 33.8 min (major), 34.5 min (minor); entry 4: tR = 33.8 min (major), 34.5 min (minor).

28 The dimer [IrCODCl]2 was also tried as a catalyst at 50 bar H2 and room temperature, but fully saturated both aromatic rings. Thus the racemate had to be synthesized by other means.77

2.4.3 Evaluation of catalysts with imidazole-based ligands for the hydrogenation of fluoroolefins

Previous work in our group lead to the chiral N,P-donating oxazole- phosphinite ligands (A),37 which were further developed to the thiazole phosphines (B);34 these were better than the oxazole ligands in terms of both selectivity and versatility.9 As a natural next step, we wanted to examine a ligand that had the heteroatom of the aromatic ring replaced with a second nitrogen atom. However, changing O or S to nitrogen without further modi- fications would result in a structure that possessed annular isomers, and would not be suitable as a ligand. Therefore, structure C was chosen as a ligand scaffold (Figure 9).

PAr2 H N PAr PAr PAr O 2 2 2 Ph N N N N + Ph R Ph O S N PAr2

ABC N Ph N H Not suitable as ligand, due to formation of annular isomers Figure 9. Oxazole-phosphinite (A) and thiazole-phosphine (B) ligands serve as models for the new class of imidazole-phosphine (C) ligands.

To visually compare the Ir complexes of these ligands, the optimized geo- metries of the Ir-ligand-ethene complexes were calculated in the Jaguar pro- gram78 using the B3LYP hybrid density theory functional79,80 and the LACVP pseudopotential basis set81. The optimized structures are similar, in fact almost overlapping for the three complexes (Figure 10).

29

Figure 10. The optimized geometries of the Ir-ligand-ethene complexes (B3LYP/LACVP). I (oxazole): light grey, IVb (thiazole): black, Va (imidazole): dark grey.

As the structures are almost overlapping, but contain different heteroaro- matic groups in the ligand, their behavior in hydrogenation was worth com- paring. Bearing in mind that the pKa of the ligand is suspected to impact defluorination, the complex of ligand C, with the most basic nitrogen bonded to iridium, was of high interest. This was especially true after it was revealed that not only defluorination and conversion, but also stereoselectivity in the hydrogenation of substrate 8 (Table 4) is correlated to pKa.

It has been found that titanium and aluminum complexes, especially those with electronegative ligands attached, will coordinate C–F.50 Therefore, it is tempting to suggest that the defluorination in the iridium-catalyzed hydro- genation is promoted by more electronegative ligands at iridium, as these should encourage C–F coordination to the iridium cation. This would explain the higher defluorination observed with catalyst I than with IVb and Va. Because oxygen (in the linker) is more electronegative than carbon, the ligand in catalyst I promotes a more electrophilic Ir centre, which promotes defluorination. This would also explain why catalyst IIa, having nitrogen (more electronegative than carbon) in the linker, displaces fluorine more efficiently than IVb. If the more pronounced loss of fluorine is ascribed to fluorine coordination to iridium, the lower ee values could be explained by the fluorine coordination competing with other interactions to steer the out- come of the reaction.

30 Table 4. Effect of pKa values of the ligand backbone nitrogen in (S)-I, (S)-IVb, and (R)-Va on the defluorination ratio, conversion and enantiomeric excess of the prod- uct after hydrogenation of E/Z-8.a

O O O H2 (100 bar) O O F Ir catalyst F + O CH2Cl2 8 XY E:Z 10:1

Ph2 Ph2 Ph2 BAr P F P BArF P BArF O Ir Ir Ir N N N Ph Ph Ph O S N (S)-I (S)-IVb (R)-Va

b pKa 0.8 2.5 7.0

X:Y 60:40 68:32 79:21 Conv. (%) 43 31 16 ee (%) 4 (S) 12 (R) 72 (S) a) Conditions: 100 bar H2, 72 h, 40 °C in CH2Cl2. 82 b) pKa values correspond to the unsubstituted heterocycles.

As the complex of the imidazole ligand, which has the most basic N donor, gives less defluorination and more enantioselectivity in the hydrogenation of 8 than the complexes with oxazole and thiazole ligands, we wanted to evalu- ate this ligand class with other fluoroolefins (Table 5). The best results (86% ee), were achieved with the Z isomers of the allylic alcohols and acetates 9 and 10 (Table 5, entries 4 and 6). These substrates also showed little de- fluorination. Remarkably, catalyst (R)-Va and (S)-Vb hydrogenated both E- and Z-10 to the same enantiomer of product (Table 5, entries 3 and 4): this was also true of catalyst (S)-Vb. The hydrogenations of the unsaturated es- ters 8 (Table 5, entries 1 and 2) by both (R)-Va and (S)-Vb gave better enan- tioselectivities than had previously been published, and E-8 gave higher ee than Z-8. The tetrasubstituted vinylfluorides 6 were hydrogenated with mod- est ee values (Table 5, entries 7 and 8).

31 Table 5. Hydrogenation of fluorine-containing olefinsa

R' F H2 (100 bar) R' F R' Ir catalyst F R' or + R R CH2Cl2 R R XY

Ph2 3,5-Me2(C6H3) P BArF P BArF Ir Ir N N Ph Ph N N (R)-Va (S)-Vb

Conv. ee Abs. Conv. ee Abs. Entry Substrate X:Y X:Y (%) (%) Conf. (%) (%) Conf.

COOEt 1 E-8 16 83:17 72 (S) 82 72:28 55 (R) F E:Z 10:1

2 F Z-8 99 66:34 46 (R) 93 65:35 29 (S) COOEt

Ac 3 O E-10 99 50:50 17 (R) 99 57:43 11 (S) F

4 F Z-10 99 84:16 85 (R) 99 87:13 85 (S) Ac O

5 OH E-9 99 43:57 52 (S) 99 51:49 60 (R) F

6 F Z-9 99 92:8 80 (R) 99 93:7 86 (S) OH

O (+)- 7 O E-6 - - - - 97 58:42 59 (2S*,3S*) F

F (+)- 8 Z-6 - - - - 35 46:54 25 (2R*,3S*) O O a) General conditions: 0.5–1.5 mol % catalyst, room temp to 40 °C, dry CH2Cl2, 30– 100 bar H2.

32 Table 6. Results of asymmetric hydrogenation with complexes Va-c.a

o-tol Ph2 2 3,5-Me2(C6H3) P BArF P BArF P BArF Ir Ir Ir N N N Ph Ph Ph N N N (R)-Va (R)-Vc (S)-Vb

Conv.b eec Conv.b eec Conv.b eec Entry Substrate (%) (%) (%) (%) (%) (%)

1 34 >99 93 (R) >99 96 (R) >99 98 (S)

2 35 >99 94 (R) >99 89 (R) >99 92 (S)

O

CO2Et 3 36 >99 93 (R) >99 93(R) >99 95 (S)

CO2Et 4 37 >99 66 (R) >99 33 (R) >99 84 (S)

5 38 >99 70 (S) >99 31 (S) >99 72 (R)

6 39 >99 90 (R) >99 89 (R) >99 86 (S)

OH Complex Complex 7 40 >99 92 (R) Mixture mixture

OAc 8 41 23 37 (R) >99 14 (R) 20 18 (S)

OH 9 42 25 61 (R) 50 84 (R) 50 75 (S)

OAc 10 43 68 63 (R) >99 70 (R) >99 73 (S)

a) Reactions conditions: 0.25 M substrate in CH2Cl2, 50 bar H2, 0.5 mol % Ir- complex, r.t., overnight. b) Determined by 1H NMR spectroscopy. c) Determined by chiral GC/MS or chiral HPLC.

33 Complexes Va-c were also used to hydrogenate the trisubstituted olefins 34- 43, which are standard olefins commonly used to evaluate new catalysts (Table 6). The selectivities were comparable to those obtained with com- plexes of ligands A and B (Further reading, Paper II). Enantioselectivities varied from good to excellent for unfunctionalized olefins 34, 35, 38 and 39 (Table 6, entries 1, 2, 5 and 6). High ee values (93–95%, Table 6, entry 3) were achieved for the hydrogenation of ethyl trans--methyl-cinnamate 36, whereas low ee values were obtained for ethyl trans--methyl-cinnamate 37 (33–84%, Table 6, entry 4). To our surprise, the hydrogenation of allylic alcohol 40 (Table 6, entry 7) proceeded smoothly with complex Vc, but gave mixtures of products when catalyzed by complexes Va and Vb.

We conclude that fluoroolefins are difficult to synthesize, often difficult to separate diastereomerically, often poorly reactive, demand harsh conditions, might lose fluorine during hydrogenation and may give enantiomeric prod- ucts that are impossible to separate. However, as discussed before, they are very interesting substances, and even though every reaction needed care and thought, we have discovered new ways to produce some fluoroalkanes in high conversion with little defluorination, and almost enantiomerically pure.

2.5 Hydrogenation of <-chlorocinnamic ester and the corresponding alcohol

The behavior of fluorine-containing olefins in hydrogenation using several different catalysts has been thoroughly investigated. Therefore, evaluating the hydrogenation of other halogenated olefins were of interest and an E/Z mixture of substrate 44 was synthesized via a literature procedure (Scheme 8).83

Ph O Ph P+ COOEt 1) NCS, CH Cl O 1) DIBAL, 0 oC, Et O Cl C- 2 2 2 Ph H 2) K2CO3, benzaldehyde Cl 2) Flash chromatography OH 44 Z-45 Z:E 87:13 Scheme 8. The synthesis of -chlorocinnamic ester 44 and its reduction to the corre- sponding alcohol 45.

Attempts to hydrogenate E/Z-44 utilizing 0.5 mol % of complexes I, IVa, IVb, VIb and VIc with 20–80 bar H2 in CH2Cl2 at room temperature for up to 5 h yielded no trace of product. The experience from the fluorinated sub- strates was that the allylic alcohol achieved from reducing a cinnamic ester is hydrogenated more readily than the ester itself. Therefore substrate 45 was

34 synthesized via standard DIBAL reduction and was attempted in hydrogena- tion reaction using 0.5 mol % complex VIa, 80 bar H2 and elevated tempera- ture (40 °C) in CH2Cl2 for 72 h. This reaction did not result in any detectable conversion (1H NMR spectroscopy).

The question arose whether the substrate itself or impurities in the sample were interfering with the catalyst. Therefore, standard substrate 35 was mixed with 44 in a vial and these were hydrogenated using 0.5 mol % cata- 1 lyst VIb under 80 bar H2 at room temperature in CH2Cl2 for 1 h. H NMR spectroscopy showed full conversion for olefin 35, but still no trace of hy- drogenation for ester 44, indicating that the catalyst is not destroyed by 35. The hydrogenation of chlorinated olefins was not examined further.

2.6 Further studies on the new catalyst (R)-II

The catalyst (R)-IIa was first designed to be evaluated in the hydrogenation of fluorine-containing olefins. The amine functionality in the linker also offered an opportunity to vary the substituents close to the phosphorus, which could permit the steric bulk around the phosphorus to be easily tuned. Therefore, both complexes (R)-IIa and (R)-IIb were synthesized by our group.

The catalysts differed significantly in the steric bulk of the substituents at the nitrogen in the linker, which would be expected to affect the hydrogenation behavior of the catalysts. The initial studies on the hydrogenations of two standard substrates, 36 and 39, yielded high conversions (98–99%) and sur- prisingly identical ee values with the two catalysts (36: 90% and 39: 95%).

The similarities in the results were somewhat explained upon examining the X-ray structure of complex (R)-IIa (Figure 11, left), as the methyl group at the nitrogen in the linker points away from the reactive center. Because the X-ray structure is measured on crystals and not in solution, where the hydro- genation takes place, the resulting structure is sometimes considered insuffi- cient for drawing conclusions. Therefore, the X-ray structure was compared with its counterpart from calculations (in vacuo). The optimized geometry (B3LYP/LACVP) and X-ray structure of IIa were almost overlapping (Fig- ure 11, right).

35

Figure 11. The X-ray structure (left) and an overlay of the X-ray structure and the optimized geometry (right, B3LYP/LACVP) of catalyst (R)-IIa. The anion has been omitted from the X-ray structure for clarity.

The fact that the substituent at the linker amine points away from the reac- tive center may make it an appropriate site for attaching the ligand to a solid phase, making it recyclable, but this has not yet been examined.

36 3 Hydrogenation of trifluoromethyl- substituted olefins (Paper III)

3.1 Introduction

As previously mentioned, fluorine-containing chiral substances are useful in applications ranging from agrochemicals to pharmaceuticals. Fluorination is also essential for the characteristics of some substances used in materials science, for example in the field of liquid crystal displays (LCDs).56,58,84-88 Chiral fluorine dopants (for example substance 48, Scheme 9) enable the fast-response ferroelectric crystals that make “flat-screen” technology possi- 88 ble. The CF3-containing alcohol 57 (Table 11, entry 3) has been used as a building block in the search for potent drugs such as calcium-channel activa- tors,89 metastasis inhibitors for cancerous cells,90 hypertension-reducing drugs91 and prolyl-endopeptidase inhibitors for the treatment of nerve system degenerative diseases, spinal injury, small strokes, and more.92 Iridium- catalyzed asymmetric hydrogenation proved useful in the making of chiral building blocks similar to these (Figure 12).

Figure 12. The asymmetric hydrogenation of CF3-substituted olefins can provide building blocks for a wide range of applications.

The existing syntheses of CF3-containing chiral centers rely mostly on the chemical or biocatalytic resolution of racemates, selective fluorination of chiral nonfluorinated substrates, or enzymatic or biological methods.55,93-105 For example, the chiral dopant 48, used in LCD screens, is synthesized via rac-46a through the resolution of diastereomeric salts in a five-step synthesis

37 106 (Scheme 9). A few CF3-substituted olefins, all containing a coordinating atom, have been hydrogenated by chiral rhodium or ruthenium catalysts in 2 – >95% ee,6 but iridium catalysts, which are useful for both coordinative and less-coordinative substrates, had not been employed. The methods for the highly stereospecific synthesis of compounds with CF3-bearing stereocenters are still very few.

OH Br + - H3N O C6H13 H2 C H 5 11 O CF3 CF CF3 3 Pd/C C H rac-46a 6 13 C6H13 48 E/Z- 46 CF3 47 85 % ee Scheme 9. Synthetic route to 48, for use in LCDs, via rac-46a.

3.2 Hydrogenation

Evaluating catalysts I, III, IV, V and VI for the asymmetric hydrogenation of Z-46 revealed wide variations in conversions and enantioselectivities among different catalysts (Table 7), as was seen in the case of fluoroolefin hydrogenation.

Table 7. Overview of the conversions and enantioselectivities in the hydrogenation of Z-46 by various catalysts.a

Ir catalyst H2 (100 bar)

CH2Cl2 CF3 46 CF 46a r.t., o.n. 3 Entry Catalyst Conv. (%) ee (%)

1 I 27 60 2 III 95 87 3 (R)-IVa 31 18 4 (R)-IVb 88 96 5 (S)-IVc 92 47 6 Va 88 95 7 VIa 79 27 8 VIb 8 47 a) Reaction conditions: 1.0 mol % catalyst, 16 h, r.t., dry CH2Cl2, 100 bar H2. Conversions were determined by 1H NMR spectroscopy and ee values were determined by chiral GC/MS.

38 The complexes with thiazole- (Table 7, entry 4) and imidazole-based ligands (Table 7, entry 6) provided the best enantioselectivity, whereas catalysts I, IVa, IVc and VI were less selective.

The dependance of the reaction on pressure and solvent were briefly exam- ined. Raising the hydrogen pressure from 50 to 100 bar did not affect the selectivity using complex IVb, but increased the ee value from 85% to 95% using complex Va. Therefore, most of the hydrogenations were performed at the higher pressure (100 bar H2).

To see how the catalyst would manage different chain lengths, a series of related olefins with alkyl chains of various lengths (Table 8, entries 1–4) were synthesized and hydrogenated using catalyst (R)-IVb. These reactions proceeded well with ee values up to 96%, although longer alkyl chains re- sulted in lower conversion. When the alkyl chain was substituted with an - phenyl group (substrate 52), the reactivity decreased further; 52 was hydro- genated in very good ee, but with low conversion, using complex (R)-III (Table 8, entry 5). A second aryl group directly on the olefin reduced the reactivity even more (entry 9). In this case, E-56 was attempted with a vari- ety of catalysts. Only (S)-IVc yielded product, and even then only traces were produced. Z-56 was difficult to purify, an E/Z mixture was hydrogen- ated to see if the Z diastereoisomer reacted more readily. This mixture also yielded only traces of product, suggesting that it was not worth purifying further. The sensitivity of the hydrogenation reaction to changes in the aro- matic moiety was also investigated using analogues of 46 with two different para-substitutiuents (entries 6 and 7); these were hydrogenated with similar enantioselectivities to the unsubstituted olefin 46 (entry 3). Placing a cyclo- hexyl group geminal to the CF3 group produced an olefin that was reduced to high conversion and with good ee using catalyst (S)-IVc (Table 8, entry 8). Previously, our group has had success in hydrogenating enol phosphi- nates,44,45 and one substrate from this class was therefore included in the present study. Full conversion and excellent selectivity were obtained.

39 a Table 8. Asymmetric hydrogenation of CF3-substituted olefins.

R R' Ir catalyst R R'

CH2Cl2 F3C F3C H2 (100 bar) 49-56 49a-56a

Entry Olefin Catalyst Conv. (%) ee (%)

1 Z-49 (R)-IVb 94 95 ()

F3C Me

2 Z-50 (R)-IVb 87 92 ()

F3C Pr

3 Z-46 (R)-IVb 88 96 ()

F3C Pentyl

4 Z-51 (R)-IVb 85 95 ()

F3C Octyl

5 Z-52 (R)-III 21 90 ()

F3C CH2CH2Ph F

6 Z-53 (R)-III 84 81 ()

F3C Pentyl

7 Z-54 (R)-III 92 84 ()

F3C Octyl

8 E- 55 (S)-IVc 96 74 ()

F3C VIa 0 - (R)-IVa 0 - 9 E- 56 (S)-IVc Trace 25 F3C (R)-III 0 -

VIa 0 - (S)-IVc Trace 0 10 E/Z- 56 F3C (R)-III 0 - E:Z 56:44 a) Reaction conditions: 0.5–1.0 mol % catalyst, 72 h, r.t., dry CH2Cl2, 100 bar H2. Conver- sions were determined by 1H NMR spectroscopy and ee values were determined by chiral HPLC or chiral GC/MS. For details see Supporting Information (Paper III). Optical rotations are given in parentheses. To date, the absolute configurations have not been correlated with optical rotations. b) The reaction was run at 50 bar H2.

40 a Table 9. Asymmetric hydrogenation of an isomeric mixture of CF3-substituted olefins.

Pentyl Hexyl Ir catalyst CF3 CF3 100 bar H2 CH2Cl2 E/Z- 46 46a Entry Olefin Conv. (%) ee (%)

1 E-46 4 0

2 Z-46 95 87 ()

3 E:Z-46 (1:1) 56 83 () a) Reaction conditions: 1 mol % catalyst, r.t., dry CH2Cl2, 100 bar H2, 72 h.

We observed that the cis and trans isomers of CF3-substituted olefins reacted with different rates and selectivities using catalyst (R)-III. Iseki and co- workers observed this phenomenon for the hydrogenation of E/Z-2- (trifluoromethyl)undec-2-en-1-ol by ruthenium catalysts but it has not been previously reported for iridium catalysts.107,108 In fact, hydrogenating an E/Z mixture of olefin 46 gives almost as high ee as hydrogenating the pure Z isomer (Table 9). Evidently, the E isomer reacts much slower than the Z isomer. In this case, it reacts so much slower that only the E isomer is visible in 1H NMR spectrum after the reaction. This is interesting because it offers the possibility to hydrogenate the Z component in a mixture of cis and trans olefins, leaving the E isomer available for further functionalization.

Intrigued by the results of hydrogenating isomeric mixtures of 46, we began further studies (Figure 13). The E and Z isomers of 46 were hydrogenated with different catalysts, revealing that catalyst (S)-IVc converts both isomers to yield low ee values. Catalyst VIa and (R)-IVa both discriminated between the two isomers, but hydrogenated the reacting olefin with low selectivity.

(R)-III showed the best results in the asymmetric hydrogenation of 46, and was therefore used in the hydrogenation of similar olefins (Table 10). For two of the isomeric mixtures examined (entries 3 and 18), only one of the isomers remained after the reaction. In most cases, the ee values of the hy- drogenated mixtures were below, but only slightly below the ee values ob- tained from hydrogenating the isomerically pure olefin. In all cases where the product was not racemic, the absolute configuration of the hydrogenated E isomer was opposite to the product from the Z isomer, but the conversion and the ee were lower for the E isomer.

41 Ir catalyst H (100 bar) 2 CH2Cl2 CF E/Z- 46 3 r.t., 72 h CF3 46a

Ph o-tol Ph o-tol P P BArF BAr Ir Ir F N N N i-Pr Ph Ph O S VIa III E: Conv = trace, ee = 15% E: Conv = 4%, ee = 0% Z: Conv = 79%, ee = 27% Z: Conv = 95%, ee = 87%

Ph o-tol Ph o-tol P P Ir BArF Ir BArF N N H S S IVa IVc E: Conv = 0%, ee = - E: Conv = 40%, ee = 19% Z: Conv = 31%, ee = 18% Z: Conv = 92%, ee = 47% Figure 13. Hydrogenation of E- and Z-46 with different catalysts.

The hydrogenation of a few other substrates was evaluated and further broadened the usefulness of this transformation. Adding additional func- tional groups to the substrate could make them more useful as building blocks for further synthesis. The enol phosphinate 58 (Table 11, entry 1) gave the best results; full conversion and 96% ee were obtained after apply- ing 50 bar H2 and 0.5 mol % catalyst for only 12 h. The esters 59 and 61 (entries 2 and 5) reacted most slowly and had also the poorest ee values. The alcohols 57 and 62 (entries 3 and 6) and the acetates 60 and 63 (entries 4 and 7) were more promising, with up to 71 % ee (entry 3). Neither reactivity nor selectivity could be gained by protecting the alcohol as an acetate, as both parameters were lower for the acetates. As mentioned previously, alcohol 57 is useful as a building block in many applications; this could be hydrogen- ated in full conversion and with higher ee using catalyst (S)-IVb.

42 Table 10. Extended study of the hydrogenation of isomeric mixtures of CF3- substituted olefins.

R' R'' H2 (100 bar) R' R'' (R)-III F3C F3C CH2Cl2

Entry R' R'' E:Z before E:Z after Conv. (%) ee (%)

1 Ph Me 49 E 13 63 (+) 2 Ph Me 49 Z 93 88 ()

3 Ph Me 49 51:49 100:0 57 73 () 4 Ph n-Pr 50 E 4 0 5 Ph n-Pr 50 Z 95 87 ()

6 Ph n-Pr 50 62:38 94:6 44 74 ()

7 Ph CH2OH 57 E 17 0

8 Ph CH2OH 57 Z 83 53 ()

9 Ph CH2OH 57 50:50 78:22 45 49 () p-fluoro 10 n-Pentyl 53 E 8 15 (+) phenyl p-fluoro 11 n-Pentyl 53 Z 84 81 () phenyl p-fluoro 12 n-Pentyl 53 53:47 74:26 37 78 () phenyl 13 Ph n-Octyl 51 E 9 17 (+) 14 Ph n-Octyl 51 Z 99 85 ()

15 Ph n-Octyl 51 63:37 94:6 44 81 ()

p-tolyl 16 n-Octyl 54 E 8 15 (+) phenyl p-tolyl 17 n-Octyl 54 Z 92 84 () phenyl p-tolyl 18 n-Octyl 54 67:33 100:0 39 87 () phenyl a) Reaction conditions: 1 mol % (R)-III, r.t., dry CH2Cl2, 100 bar H2, 72 h.

43 Table 11. Hydrogenation of CF3-substituted olefins containing additional functional groups.a

R R' Ir catalyst R R'

CH2Cl2 F3C R'' F3C R'' H2 (100 bar) Entry Olefin Catalyst Conv.b (%) eec (%)

1 Ph2(O)PO 58 (R)-III >99 96 (+)

F3C

2 59 (R)-III 49 22 (-)

F3C O O (S)-IVb 99 71 (-) 3 57 (R)-III 83 53 (-) F3C OH

4 60 (R)-III 93 68 (-)

F3C OAc

O F3C O 5 61 (R)-III 10 0

F3C OH 6 62 (S)-IVb >99 65 (+)

F3C OAc 7 63 (S)-IVb 99 59 (-)

a) Conditions: Entry 1: 0.5 mol % catalyst, 12 h, r.t., CH2Cl2, 50 bar H2; entries 2–7: 1.0 mol 1 % catalyst, 72 h, r.t., CH2Cl2, 100 bar H2 b) Conversions are determined by H NMR spec- troscopy. c) Determined by chiral GC analysis.

44 4 Formation of 1,1-diarylmethine stereocenters (Paper IV)

4.1 Introduction

Diarylmethine chiral centers can be found both in marketed drugs, such as Tolterodine109,110 and Sertraline111 (Figure 14), and in natural products (e.g. podophyllotoxin112).

NHMe

Me N

OH Cl Cl (R)- Tolterodine Sertraline Figure 14. Two examples of pharmaceuticals bearing diarylmethine stereogenic centers.

Diaryl stereocenters can be produced in several ways,113-120 and both Tol- terodine and Sertraline have been made using several synthetic meth- ods.113,119,121-129 However, the existing approaches to diarylmethine chiral centers are often limited in substrate scope. The synthesis of unfunctional- ized, chiral gem-diarylalkanes is challenging and few reports have been pub- lished in this area.130-134 Existing methods involve asymmetric alkyla- tion,130,132 carboselenylation,131 deracemization133 or cross-coupling reactions of chiral alkylsilanes with aryltriflates,134 and generally yield rather low ee values. Carreira et al. recently reported a Rh-catalyzed decarbonylation of optically pure aldehydes to chiral diarylethanes with excellent selectiv- ities,135 but new methods to diarylmethine stereocenters are still needed.

The aim was to use Ir-catalyzed asymmetric hydrogenation to form di- arylmethine stereocenters and increase the scope of known optically pure diarylmethine compounds. At first glance, this task appeared difficult; two aryl groups often have very similar steric and electronic properties. When reducing olefins bearing almost identical groups at the prochiral carbon, we

45 rely on other steric interactions to induce selectivity. In this study, diarylme- thine chiral centers were made from substrates whose geminal olefin sub- stituents differed very little. In one example, these differed by only a single atom in the para-position. These results support the hypothesis that steric hindrance in one specific part of the substrate plays a key role in the hydro- genation,6,136-139 which is highly enantioselective for many substrates. The enantiodiscrimination is discussed further in Chapter 5.

4.2 Hydrogenation

Two substrates were initially chosen for a catalyst screening to get an idea of which catalysts performed best with these types of bulky substrates. One unfunctionalized olefin and one thiophene- and ester-substituted olefin were chosen to represent two different substrate classes. The asymmetric hydro- genation of the ,-unsaturated ester 63 required heating, and both conver- sion and enantioselectivity were very catalyst-dependent. The best ee value was achieved using the thiazole catalyst (R)-IIa (Table 12, entry 4). The reduction of unfunctionalized olefin 64 proceeded well without heating and was most effective with the thiazole-based catalyst (S)-IVb and the imida- zole-based catalyst (S)-Vb; both fully reduced 64 with excellent enantiose- lectivities (Table 12, entries 1 and 3).

46 Table 12. Catalyst screening in the asymmetric hydrogenation of 1,1-diaryl- substituted olefins 63 and 64.a

Ar'' R Ir catalyst Ar'' R H2 Ar' CH2Cl2 Ar' 63 or 64

Me S

CO2Me Ph Ph Ph 63 64

Entry (N,P)* Cat. Conv.b (%) eec (%) Conv.b (%) eec (%)

PPh2 >99 1 N (S)-IVb 29 71 (+) d >99 (-) Ph (>95) S PPh O 2 2 N (S)-I 38 30 (+) 11 94 (-) Ph O P(3,5-Me2C6H3)2 3 N (S)-Vb 51 67 (+) >99 >99 (+) Ph N Me PPh N 2 4 N (R)-IIa 83 87 (-) 57 >99 (+) Ph S PPh N 2 5 N VIc 66 68 (-) 8 50 (-) O P(o-Tol) N 2 6 N VIa >99 77 (-) 81 80 (-) O a) Reaction conditions: 25 mg of 63 or 64, 1 mol % catalyst, 1 mL CH2Cl2. Hydrogenation of 63: 40 °C, 100 bar H2, 24 h. Hydrogenation of 64: 25 °C, 50 bar H2, 24 h. b) Conversion to alkane determined by 1H NMR spectroscopy. c) Determined by chiral HPLC. For details see Supporting Information (Paper IV). d) Isolated yield.

The effects of solvent, temperature and hydrogen pressure on the asymmetric hydrogenation of 63 by catalyst (S)-IVb were also studied (Table 13). The reaction time was limited to 24 h to keep conversions comparable. The ee values did not differ very much when using different solvents, but the best ee value was achieved in <,<,<-trifluorotoluene solution (entry 1). The conver- sions, on the other hand, differed a lot. The best conversion was obtained in dichloromethane (entry 5) whereas the reaction run in 2,2,4-trimethylpentane gave the lowest conversion (entry 2). This might be due to decreased solubil- ity of the catalyst complex in the less-polar solvent, though the phenomenon was not investigated further. No conversion was observed when the reaction was run at room temperature and 50 bar H2 pressure (Table 13, entry 4).

47 When either the reaction temperature (25  40 °C) or hydrogen pressure (50  100 bar) was raised, the conversion followed (entries 3 and 5).

Table 13. The effects of solvent, temperature and hydrogen pressure on the asym- metric hydrogenation of 63 with catalyst (S)-IVb.a

S S Ir catalyst (S)-IVb CO Me CO2Me 2 H2 Ph Ph 63 63a Hydrogen Temperature Conv.b eec Entry Solvent Pressure (°C) (%) (%) (bar)

1 ,,-Trifluorotoluene 100 25 39 82 (+)

2 2,2,4-Trimethylpentane 100 25 17 78 (+)

3 CH2Cl2 100 25 31 75 (+) d 4 CH2Cl2 50 25 0 -

5 CH2Cl2 50 40 46 71 (+) a) Reaction conditions: 25 mg 63, 1 mol % catalyst, 1 mL solvent, 24 h. b) Determined by 1H NMR spectroscopy. c) Determined by chiral HPLC. See Supporting Information for details (Paper IV). d) Not applicable.

Based on the optimization study and catalyst screening, catalysts (S)-IVb and (R)-IIa were applied to further substrates. We reasoned that the catalyst that performed best with each of the two substrates should be chosen, as the olefins represent two different substrate classes. Thus, catalyst (R)-IIa was chosen because it performed best with olefin 63. Although both catalyst (S)- IVb and (S)-Vb showed excellent results when reducing substrate 64, (S)- IVb was chosen because of its slightly higher ee values in the reduction of 63. The best reaction conditions from Table 13, i.e. ,,-trifluorotoluene or CH2Cl2 solution, 25 °C and 100 bar hydrogen pressure, were used for the subsequent studies and optimized further when needed.

4.2.1 Unfunctionalized olefins

Unfunctionalized olefins, which had either alkane or phenyl substituents at the non-prochiral olefin terminus, were studied first (Table 14). Overall, these were reduced highly selectively. Placing an electron-withdrawing CF3 substituent on one aromatic ring caused the reaction to proceed slowly, re- quire heat and give moderate enantioselectivity (entry 3). Substituents with more electron-donating substituents (-Me, -OMe), were reduced completely at room temperature (entries 1 and 2) to excellent ee values. Interestingly,

48 another substrate bearing an electron-withdrawing substituent, the bromo- substituted 71, gave an excellent ee of 95% (entry 7); this is particularly interesting because the resulting optically active product can be used as an intermediate in cross-coupling reactions.140

Table 14. Ir-catalyzed asymmetric hydrogenation of unfunctionalized 1,1-diaryl- substituted olefins.a

Ar'' R Ir catalyst Ar'' R H2 Ar' CH2Cl2 or PhCF3 Ar' 65-72 65a-72a Hydrogen Conv. Entry Substrate Cat. T (°C) Solvent ee (%) Pressure (bar) (%)

Me

1 65 (S)-IVb 25 50 CH2Cl2 >99 >99 (-)

Ph Ph MeO

2 66 (S)-IVb 25 50 CH2Cl2 >99 95 (-)

Ph Ph F3C

3 67 (R)-IVb 40 100 CH2Cl2 48 76 (-)

Ph Ph Me

Me 4 68 (R)-IVb 25 50 CH2Cl2 >99 >99 (+)

Ph Ph

Me 5 69 (R)-IIa 80 100 PhCF3 35 92 (+) Ph Ph Me

6 pentyl 70 (R)-IIa 25 100 PhCF3 >99 97 (-)

Ph Br 95 (+) 7 71 (R)-IIa 25 100 PhCF >99 3 (S)b Ph Me Ph

8 72 (R)-IIa 25 100 PhCF3 99 99 (+)

Ph pentyl a) Reaction conditions: 0.2–0.3 M substrate in CH2Cl2 or ,,-trifluorotoluene, 1 mol % catalyst, 24 h. Conversions were determined by 1H NMR spectroscopy and ee values were determined by chiral HPLC or chiral GC. For details, see Supporting Information (Paper IV). Optical rotations are shown in parentheses. b) Absolute configuration assigned after derivati- zation.

49 The effect of substitution in different positions on one aryl moiety was in- vestigated by placing a methyl group in the ortho, meta and para positions on one phenyl group. Both the meta,meta-dimethyl-substituted substrate 68 and para-substituted substrate 65 gave excellent enantioselectivities (99 and 97% ee) and full conversions; whereas the ortho-substituted substrate 69 required extensive heating to react, possibly due to steric hindrance near the carbon-carbon double bond. Interestingly, the enantioselectivities were not strongly affected by the steric properties of the third olefin substituent, R. Also worth noticing is that para substitution with a phenyl group (substrate 72), which is much bigger than a methyl group also results in a substrate that is hydrogenated to high conversion and with an excellent ee value (entry 8).

4.2.2 Esters, alcohols and acetates

Substrates with ester, alcohol, or acetate substituents on the non-prochiral carbon atom of the olefin, gave more varied results in terms of conversions and enantioselectivities (Table 15). As suggested by the initial studies, the asymmetric hydrogenation of ,-unsaturated esters 73 and 74 required slightly elevated temperatures to reach full conversion. When the hydrogena- tion of the allylic alcohols 75-77 was attempted using complex (R)-IIa, com- plicated product mixtures resulted. As complex (R)-Vc can successfully hy- drogenate a related allylic alcohol,17c it was used in the hydrogenation of diaryl-substituted allylic alcohols and an acetate, giving high ee values and conversions (Table 15, entries 3–6). Comparing the hydrogenations of alco- hols 75 and 76 (entries 3 and 4) reveals that, contrary to the results obtained with unfunctionalized olefins (Table 14), electron-withdrawing -CF3 and electron-donating -OMe groups in the para-position of an aryl group do not influence conversion or enantioselectivity much.

50 Table 15. Ir-catalyzed asymmetric hydrogenation of functionalized 1,1- diarylolefins.a

Ar'' R Ir catalyst Ar'' R H2 Ar' Ar' 73-78 73a-78a Conv. Entry Substrate Catalyst T (°C) ee (%) (%)

MeO

1 73 (R)-IIa 40 >99 69 (+) (R)b CO2Me Ph

O

2 CO2Me 74 (R)-IIa 40 >99 80 (-) Ph MeO

c 3 OH 75 (R)-Vc 25 >99 92 (-) (R)

Ph

F3C

4 OH 76 (R)-Vc 25 >99 92 (+)

Ph

S 5 OH 77 (R)-Vc 25 >99 90 (-) Ph MeO

c 6 OAc 78 (R)-Vc 25 >99 95 (-) (R)

Ph a) Reaction conditions: 0.2–0.3 M substrate in ,,-trifluorotoluene, 1 mol % catalyst 100 1 bar H2, 24 h. Conversions were determined by H NMR spectroscopy and ee values were determined by chiral HPLC or chiral GC. For details, see Supporting Information. Optical rotations are shown in parentheses. b) Absolute configuration assigned by comparing reten- tion order in HPLC with literature data.113 c) Absolute configuration assigned after converting the chiral product from substrate 73 into alcohol and acetate.

The fact that high selectivity can be achieved with olefins having very simi- lar aryl groups opens up new possibilities. For example, it should be possible to obtain high ee values in the reduction of substrates with aryl groups differ- ing only in isotopic substitution. The question, however, is how to evaluate the results of such a reaction; separating enantiomers can be difficult even for enantiopairs having very different substituents.

51 5 Enantiodiscrimination

5.1 Introduction

To further understand the enantiodiscrimination in iridium-catalyzed asym- metric hydrogenation, known absolute configurations for some of the prod- ucts were required. The absolute configurations of many of the fluorine- containing and diaryl products were known or determined by us, but the configuration of the trifluoromethylated products were hard to determine, as the data published on these are limited.

General selectivity models for the Ir-catalyzed asymmetric hydrogenation of olefins have been proposed by Brandt,138 Andersson136,137 and Burgess and Hall.139 Their DFT calculations suggest that the olefin binds to the metal in the same equatorial plane as the chiral ligand, cis to the nitrogen atom. The olefin binds preferably with its smallest substituent oriented towards the bulky group of the ligand (Figure 15). The absolute configuration of the resulting chiral hydrogenation product can be rationalized by considering the sterics about the Ir center from the perspective of the olefin. Separating the space into four quadrants, labelled I–IV as in a Cartesian coordinate system, showed that quadrant III were occupied by a phenyl group and therefore sterically hindered. The two aryl substituents on the phosphorus atom partly occupy quadrant I, which is therefore semi-hindered. Quadrants II and IV, which are free from bulky groups, are open. Thus the smallest substituent of the olefin (-H) should be pointing towards the hindered quadrant III in order to minimize steric repulsion in the catalyst-substrate complex.

II I + H R''' H R'' PAr N 2 Semi- R'' Open R''' Ir Ir Hindered R H H2 Bulk R' R'

IIIHindered Open IV H = the hydride in the migratory insertion step Figure 15. The substituent, R, on the heterocycle of the ligand, blocks one quadrant, which contributes to the selectivity in the reaction.

52 The electronic properties of the olefin also play a key role in directing the transformation. If an electron-withdrawing group is conjugated with the ole- fin, the initial hydride addition occur preferably at the  position, which will favour the electronically matched orientation (Figure 16).

+ C-EWG 

Ir Ir

 + C-EWG II I II I II I C-EWG H H H Semi- Semi- Semi- Open Hindered Open Open Ir Ir Hindered IrIr Hindered

Bulk Bulk Bulk C-EWG

III IV Hindered Open III Hindered Open IV IIIHindered Open IV Electronic mis-match Empty catalyst Electronic match Figure 16. An olefin possessing a conjugated electron-withdrawing group (c-EWG) will prefer addition of the initial hydride in the -position, which will favour the orientation to the right (electronically matched).

The quadrant representation for the (R)-form of a specific complex will be horizontally flipped compared to its (S)-form (Figure 17). The quadrant pro- jection of one (R) complex does not necessarily have to be equal to the one of another (R) complex. For complexes I-V the (S) catalysts are overlapping, and therefore have the same quadrant projection. The catalysts based on the aza-norbornyl ligand have the same quadrant projection as (R)-I; therefore, the complexes behaving like a (R)-catalyst are called (“R”)-catalysts in the figure and vice versa. As a majority of the catalysts used in these studies are of (“R”)-type, this representation is most frequently used, even though the (“S”)-quadrant representation model is most common in the literature.

("S") catalyst ("R") catalyst

II I I II

H H Semi- Semi- Open Hindered Open Ir Hindered Ir

Bulk Bulk

IV Hindered III III Hindered Open IV Open Figure 17. The quadrants for the catalysts behaving like (S)-I, called (“S”) catalyst, are horizontally flipped compared to the (“R”) catalyst.

53 5.2 Fluorine- and trifluoromethyl-substituted olefins

Based upon the assumption that the orientation of an olefin at the metal cen- ter is limited as proposed (Figure 15), the substrates from the studies of fluo- rine-containing olefins were arranged into the quadrant diagram (Table 16). If olefin Z-8 is oriented with the fluorine atom towards quadrant III (orienta- tion i, sterically unfavoured), the addition of the initial hydride occurs to the carbonyl (electronically favoured, assuming that their electronic proper- ties are similar to the non-fluorinated counterpart). If the olefin adopts orien- tation ii, the insertion of the hydride should be electronically unfavoured, but sterically matched (Table 16, entry 1). It is hard to predict how the fluorine atom will affect the electronics directing the reaction. A recent computa- tional study, concluded that adding two electron-withdrawing trifluoro- methyl groups to the position will reverse the regioselectivity to yield pre- dominantly  addition to the double bond (Figure 18).141

- ? Nu Nu-

+ F C CF H H H CF3 3 3 Ir Ir Ir  C-EWG F C-EWG C-EWG

- Nu Nu- ? Figure 18. The addition of additional electron-withdrawing groups on the double bond are predicted to affect the regioselectivity.

The study also reveals that attaching a single fluorine atom at C will have the opposite effect, stabilizing -addition. However, attaching a second fluo- rine at C promotes -addition. Chatfield et al. ascribed these countervailing effects to the fluorine atom’s dual character as both weak -donor and - withdrawing group,141 but arguments for the lack of -donating character of the fluorine in for example substrate Z-15 have been published.50 In any case, Chatfield et al. found fluorine to be less effective than other electron- withdrawing groups in promoting  addition. Another report handling calcu- lated atomic partial charges indicated that attaching a single fluorine atom at C interchanges the partial positive and negative charges at C and C, mak- 142 ing C more positive (Figure 19). These findings make it difficult to pre- dict the preferred orientation of fluoroolefins in the hydrogenation reaction.

54 0.01 -0.02 + -

O - O F + + H  +  0.18  0.18 0.06 -0.07 O O Figure 19. Adding a fluorine at C changes the properties of the olefin significantly, 142 interchanging the partial charges at C and C.

Given that a phenyl group in quadrant III and fluorine group in I is sterically unfavoured, E-8 is assumed to coordinate in orientation ii (Table 16, entries 1 and 2). This correctly predicts the stereochemical outcome of its hydro- genation. Note the initial  addition of the hydride. Considering that Z-8 and E-8 give opposite enantiomers for all catalysts tried, it is reasonable to as- sume that the non-prochiral carbon is setting the selectivity and it is there- fore reasonable to think that both Z- and E-8 arrange in orientation ii. The acetates (entries 3 and 4), all give the (R) product (with varying ee), regard- less of olefin diastereomer used. This suggests, contrary to the results with the esters, that the prochiral termini of olefins Z-10 and E-10 are setting the selectivity, which implies that the fluorine is small enough to be directed toward quadrant III (sterically unfavoured). The same pattern is seen for alcohols Z- and E-9 using catalyst VIa. However, the hydrogenation of alco- hols Z-9 and E-9 apparently occurs from the more favourable orientation ii when catalyzed by (R)-IIa, (R)-Va, (S)-Vb (entries 5 and 7), which again hydrogenate the two diastereomeres to opposite enantiomers. The outcome of the hydrogenation of tetrasubstituted substrates E-6 and Z-6 are difficult to rationalize because only the relative configurations of the products are known, but the outcome of the hydrogenation of Z-7 would be correctly pre- dicted by arranging it in orientation ii. However, the likeliness of the methyl group of pointing right toward quadrant III is questionable.

To apply the selectivity model to these substrates is difficult. The possibility that the small fluorine atom may point towards the bulky quadrant, and per- haps the protean electronic properties of fluorine, make the situation even more complex.

55 Table 16. The proposed olefin orientations that result in the observed stereoisomerical outcome of the reactions. Orientation (“R”)-VIa (R)-IIa (R)-Va (S)-Vb

ee Abs. ee Abs. ee Abs. ee Abs. Entry I ii (%) Conf. (%) Conf. (%) Conf. (%) Conf.

I II I II H H F COOEt 1 Ir Ir Z-8 29 (R) - - 46 (R) 29 (S) EtOOC F Bulk Bulk IV III IV III

I II I II H H EtOOC F

2 Ir Ir E-8 7 (S) 8 (S) 72 (S) 55 (R) EtOOC F Bulk Bulk IV III IV III

I I II II Ac O H H F 3 Ir Ir Z-10 87 (R) >99 (R) 85 (R) 85 (S) F Bulk Ac O Bulk IV III IV III

I II I II Ac O H H F

4 Ir Ir E-10 34 (R) 10 (R) 17 (R) 11 (S) F O Bulk Bulk Ac IV III IV III I II I II H H OH F 5 Ir Ir Z-9 80 (R) 99 (R) 80 (R) 86 (S) F OH Bulk Bulk IV III IV III I II I II OH H H F

6 Ir Ir E-9 28 (R) F OH Bulk Bulk IV III IV III

I II I II H OH H F 7 Ir Ir E-9 6 (S) 52 (S) 60 (R) F OH Bulk Bulk IV III IV III I II I II H H F COOEt (-)- (+)- 8 Ir Ir Z-6 57 * * rac - - - 25 EtOOC F (2R ,3S ) (2R*,3S*) Bulk Bulk IV III IV III I II I II H H EtOOC F (+)- (+)- 9 Ir Ir E-6 74 * * rac - - - 59 EtOOC F (2S ,3S ) (2S*,3S*) Bulk Bulk IV III IV III I II I II OH H H F (+)- (+)- 10 Ir Ir E-7 90 82 - - - - F (2S,3S) (2S,3S) OH Bulk Bulk IV III IV III

56 The trifluoromethylated olefins are also difficult, although their electronic effects can be accounted for by considering only their  character. There are few reports of this kind of products, and reports containing known absolute configurations for them are even fewer. We do observe that the E and Z di- astereomers of the trifluoromethyl-substituted substrates are hydrogenated to give opposite enantiomers. Therefore, the selectivity is presumably induced by the non-prochiral carbon of the olefin, similarly to the case of the 1,1-diaryl olefins (Figure 20). The selectivity is higher for the Z diastereo- mers (having the larger groups trans on the double bond).

I II

CF3 H Ar

Ir

R Bulk IV III Figure 20. The assumed preferred orientation of trifluoromethyl-substituted olefins.

5.3 1,1-Diaryl olefins

Arranging any of the 1,1-diarylolefins 71, 73, 75 or 78 on the diagram ac- cording to the selectivity model, i.e. with its smallest substituent (-H) point- ing toward the bulky group of the ligand, allows the correct prediction of the observed stereochemical outcome of that olefin’s hydrogenation (Figure 21).

Figure 21. Schematic diagram showing the critical enantiodetermining substrate- ligand interactions that occur when a trisubstituted 1,1-diarylolefin coordinates to the catalyst.

Convincing support for this binding mode leading to the major product comes from the hydrogenation of the cis/trans isomer pairs E- and Z-71 and

57 E- and Z-72. The two isomers gave opposite enantiomers in almost equal ee values (Figure 22).

Br Br

Ir catalyst (R)-IIa H2 PhCF3 Ph Me Ph Me E-71 or Z-71 E-71: 95% ee (S) Z-71: 99% ee (R)

Ph Ph

Ir catalyst (R)-IIa H2 PhCF3 Ph pentyl Ph pentyl

E-72 or Z-72 E-72: 99% ee (+) Z-72: 99% ee (-)

Figure 22. The asymmetric hydrogenation of the cis/trans isomer pairs 71 and 72.

The selectivity model can easily predict the sense of enantioselectivity in the hydrogenation of 1,1-diaryl olefins, and probably also of CF3-substituted olefins (we can not evaluate this because the absolute configuration is un- known for the products), but predictions for fluorinated olefins is compli- cated by the size of the fluorine atom (larger than a hydrogen) and the pro- tean electronic properties of fluorine substituents.

58 6 Conclusions and Outlook

We have shown that iridium complexes can catalyze the asymmetric hydro- genations of fluorine-, trifluoromethyl- and 1,1-diaryl-substituted olefins. The loss of fluorine initially seemed like a problematic side reaction, but further studies revealed several catalysts that showed little of this reaction.

These studies gave among the best results published to date for the asymmet- ric hydrogenation of fluorine substrates, yielding high conversion and very high ee values (up to 99%) with very little fluorine elimination in some cases. The trials also revealed that tetrasubstituted fluorine-containing ole- fins can be hydrogenated highly enantioselectively (up to 90 % ee), despite that tetrasubstituted olefins have usually shown low reactivity in this reac- tion type.

We have also developed a new, highly enantioselective (up to 96% ee) syn- thetic route to building blocks with CF3-groups at their chiral centers. These are useful in applications ranging from pharmaceuticals to ferroelectric crys- tals. This synthetic route does not rely on additional functionality to direct its stereochemical course, and is therefore useful in the formation of chiral fluo- rine-containing molecules for a wide range of applications. The hydrogena- tion of an isomeric mixture of olefins gave the hydrogenation product highly enantioselectively (87% ee).

We have also created diarylmethine stereogenic centers from trisubstituted olefins with aryl groups having a variety of electronic and steric properties. The enantioselectivies and conversions were high, even when the differences between the aryl groups on the prochiral carbon were very small. These re- sults also strengthen the hypothesis that steric hindrance in one specific quadrant of the catalyst plays a key role in this transformation.

Looking at all of these results together, one potential further refinement seems obvious. Comparing the hydrogenation of the fluoro-substituted alco- hol Z-28 using catalyst VIa (Table 3, entry 5), the hydrogenation of 40 (Ta- ble 6, entry 7) and the hydrogenation of 75-77 (Table 16, entries 3–5) reveals that catalyst (R)-Vc could be successful in the hydrogenation of Z-28.

The possibility of making a recyclable solid-phase catalyst for asymmetric hydrogenation by attaching ligand (R)-II to a solid support was mentioned

59 earlier (Chapter 2.6). This would allow a column to be filled up with solid- supported asymmetric catalyst, and olefins could be applied in solution. This way, the batchwise hydrogenations used today could potentially be replaced by more facile continuous processes (Figure 23).

L33

H2 R1V501

L33V101

L33V100 L33V501 Reactor 1 Product tank 1

M1 P1V501

L45V501 M1V501 Figure 23. Process Flow Diagram describing a continuous process for hydrogenation.

- A further extension could be to attach the BArF anion to a solid phase and add the catalytic iridium-N,P-ligand complex directly to the column via ion exchange (Figure 24).

- Figure 24. The BArF anion situated on a solid support (polystyrene) and flanked by the active N,P-ligated iridium catalyst.

This procedure would have several advantages. When the catalyst is no longer active, adding new catalyst via ion exchange would easily regenerate the column. Because of the simple regeneration, the filling material would not have to be changed as often as when a ligand is added to solid support. The synthesis of catalyst does not have to be done directly on the solid sup- port, as only the anion is covalently bound to the column. In either case, the lifetime of the catalyst might be prolonged if its immobilization as on a sur- face prevents the formation of trimers, deactivating the catalyst.

60 Acknowledgements

I would like to express my sincere gratitude to a lot of people who have made this journey possible and more enjoyable:

My supervisor Prof. Pher G. Andersson for all the help and support I have received during my stay at the department and for always being available for sharing knowledge and thoughts.

My second supervisor, Prof. Henrik Ottosson for keeping me on track both concerning research and fighting mosquitoes and tics on the running path.

The Swedish Research Council (VR), European Cooperation in Science and Technology (COST) and Swedish Foundation for Strategic Research (SSF) for financial support.

Tamara, Diana, Henrik, Julius and Pher for proofreading and commenting my thesis. Your help has been invaluable.

Past and present members of the PGA group; Alexander “Is it possible to take that apparatus apart?” Paptchikhine, Johan “Chemicals don’t grow on trees - yet” Verendel, Päivi “Super-efficient” Tolstoy, Dr. Tamara “The language genius” Church, Dr. Jarle “Because it’s together, people will re- member us” Diesen, Dr. Pradeep “Always in the hood next to me” Cheruku, Jia-Qi “Ballade for Adeline” Li, Taigang “Always happy” Zhou, Sadiq “HELLO!” Mohammad, (for keeping my attention on my surroundings and not only on writing the thesis), Maryam, Dr. Klas Källström, Dr. Stefan Modin, Dr. Christian Hedberg, Dr. Anna Trifonova, Dr. Iveta Kmentova, Dr. Ian Munslow, Dr. Eskender “It’s true!” Mume, Muhammad Ali and Muhammad Ali, Anna, Andreas, Tobias R., Sebastian V., Katrin A., Abby, Declan and Martin. It has been a pleasure working with you all!

The whole Prof. Mikael Widersten group: Especially Diana Lindberg and Ann Gurell, thanks for lending me you HPLC and for all nice discussions and support. Thanks for moving up.

The previously neighboring Prof. Henrik Ottosson group: Especially Julius “Findus” Tibbelin and Dr. “Whyy” Tong.

61 Henrik “Fellurium, said Tille” Johansson, your presence adds splendour to coffeebreaks and long HPLC-sessions.

Gunnar Svensson for always wanting to help and for untiringly fixing new broken parts on the espresso-mashine. Thomas Kronberg and Lilian Fors- berg for all support during the undergraduate teaching. Eva, Johanna and Inger for helping with administrative issues.

Prof. Jonas Bergquist for help with many HRMS samples.

Prof. Adolf Gogoll for the guidance in 1H-19F NOE.

Dr. Torben Rasmussen for getting me started in the world of Maestro, Jaguar and written commands.

Stiftelsen Fredrik Lindströms minne for all support during my studies.

Mom and Dad

My own little family, Petra och Zion, my angels. Thank you for filling my life with action and love. No one makes it better than the two of you.

Tusen tack!

Mattias

Uppsala, March 2009

62 Summary in Swedish

Kemo- och enantioselektiva hydrogeneringar Kampen för att vidga substratbredden för iridiumkata- lyserad asymmetrisk hydrogenering

Neurosedynkatastrofen under 1960-talet är ett typexempel på konsekvenser- na av att en molekyl kan existera i två former som är varandras spegelbilder. Dessa s.k. enantiomerer kan ha helt olika egenskaper. I det nämnda fallet hade den ena önskad farmaceutiskt effekt, medan den andra formen gav upphov till mycket svåra fosterskador. Ett annat exempel på dessa enantio- merers olika interaktion med vår kropp är de två formerna av limonen, där den ena enantiomeren smakar och luktar citron medan den andra smakar apelsin (Bild 1). Dessa två exempel påminner oss om varför det ibland är viktigt att ha verktyg för att selektivt kunna tillverka bara den ena av de två.

H H

(S)-Limonen (R)-Limonen

Citron Apelsin Bild 1. De två formerna av limonen luktar och smakar olika.

De två enantiomerena kan liknas vid en höger och en vänsterhand. Båda händerna har fem fingrar, men de sätts samman i olika ordning. Det asym- metriska elementet (något förenklat) - tummen - sitter på olika sidor, vilket resulterar i att händerna blir spegelbilder av varandra.

De båda spegelbilderna kommer bete sig på samma sätt i miljöer som inte är kirala, d.v.s. de som inte innehåller något som är asymmetriskt, men väsent- ligen olika i en miljö som är kiral. Eftersom vår kropp består av en massa kirala element, kommer höger- och vänsterformen av t.ex. ett administrerat

63 läkemedel kunna ha helt olika effekt i kroppen. Detta kan verka underligt, men är egentligen inte konstigare än att en högerhand passar bättre in i en högerhandske.

Ett verktyg för att kemiskt tillverka enantiomerer är asymmetrisk hydrogene- ring. Ordet asymmetrisk betyder att reaktionen inte producerar lika många av de två enantiomererna, vilket i sin tur betyder att vi kan få en selektivitet mot den av de två vi önskar. Med hydrogenering menar man introduktion av väte i en molekyl, i vårt fall vätgas. Asymmetrisk hydrogenering kan alltså be- skrivas som en reaktion med väte som inte bildar en 50/50-blandning av produkterna, d.v.s mer av den ena av dem.

I centrum av denna reaktion står i vårt fall oftast en iridiumjon. Runt denna har man koordinerat in molekyler för att skapa en omgivning, en ficka, som ser till att ursprungsmolekylen som ska reagera - något förenklat - bara kan koordinera in på ett sätt. Vätgasen som ska reagera med ursprungsmolekylen koordinerar också in till metallen. När alla dessa förutsättningar är uppfyllda sker reaktionen. Eftersom koordinationsmöjligheterna för ursprungsmoleky- len är begränsade, kan vi få mer av den ena formen av produkten.

Innehållet i den här avhandlingen beskriver arbetet för att öka användbarhe- ten av iridiumkatalyserad hydrogenering genom att finna nya substrat som låter sig reagera med hög selektivitet. Metoder har utvecklats för att reagera tre nya substratklasser.

Eftersom inkorporerandet av fluor i molekyler på senare år har ökat inom många olika områden är tillverkningen av fluorinnehållande kirala center högintressant. Kiralt rena substanser med trifluorometylgrupper kan använ- das till allt från läkemedel till LCD-skärmar. Föreningar med två arylgrupper i stereocentrat dyker även de upp i läkemedel. I den här studien har vi lyck- ats reagera alla dessa tre substansklasser med mycket hög selektivitet (mer än 99,5 % av den ena enantiomeren). Dessa resultat demonstrerar flera nya användningsområden för verktyget iridiumkatalyserad asymmetrisk hydro- genering som jag härmed stolt lägger ner i den kemiska verktygslådan.

64 References

1. L. Pasteur, Compt. Rend. Acad. Sci. 1848, 26, 535. 2. D. M. Burley; W. Lenz, The Lancet 1962, 279, 271-272. 3. W. Lenz; R. A. Pfeiffer; W. Kosenow; D. J. Hayman, The Lancet 1962, 279, 45-46. 4. J. M. Brown, Hydrogenation of functionalized carbon-carbon double bonds. In Comprehensive Asymmetric Catalysis I-III, Jacobsen, E. N.; Pfaltz, A.; Yamamoto, H., Eds. Springer-Verlag: Berlin, 1999; Vol. 1, pp 121-182. 5. X. Cui; K. Burgess, Chem. Rev. 2005, 105, 3272-3296. 6. T. L. Church; P. G. Andersson, Coord. Chem. Rev. 2008, 252, 513-531. 7. S. J. Roseblade; A. Pfaltz, C. R. Chim. 2007, 10, 178-187. 8. S. J. Roseblade; A. Pfaltz, Acc. Chem. Res. 2007, 40, 1402–1411. 9. K. Källström; I. Munslow; P. G. Andersson, Chem.- Eur. J. 2006, 12, 3194-3200. 10. R. H. Crabtree, Acc. Chem. Res. 1979, 12, 331-337. 11. A. Lightfoot; P. Schnider; A. Pfaltz, Angew. Chem., Int. Ed. 1998, 37, 2897-2899. 12. M. Schrems; E. Neumann; A. Pfaltz, Heterocycles 2008, 76, 771-781. 13. J. Zhou; J. W. Ogle; Y. Fan; V. Banphavichit; Y. Zhu; K. Burgess, Chem.- Eur. J. 2007, 13, 7162-7170. 14. M. N. Cheemala; P. Knochel, Org. Lett. 2007, 9, 3089-3092. 15. T. Bunlaksananusorn; K. Polborn; P. Knochel, Angew. Chem., Int. Ed. 2003, 42, 3941-3943. 16. W. Tang; W. Wang; X. Zhang, Angew. Chem., Int. Ed. 2003, 42, 943-946. 17. M. C. Perry; K. Burgess, Tetrahedron: Asymmetry 2003, 14, 951-961. 18. P. G. Cozzi; F. Menges; S. Kaiser, Synlett 2003, 833-836. 19. D.-R. Hou; J. Reibenspies; T. J. Colacot; K. Burgess, Chem.- Eur. J. 2001, 7, 5391-5400. 20. J. Blankenstein; A. Pfaltz, Angew. Chem., Int. Ed. 2001, 40, 4445-4447. 21. P. G. Cozzi; N. Zimmermann; R. Hilgraf; S. Schaffner; A. Pfaltz, Adv. Synth. Catal. 2001, 343, 450-454. 22. S. Kaiser; S. P. Smidt; A. Pfaltz, Angew. Chem., Int. Ed. 2006, 45, 5194-5197. 23. G. Xu; S. R. Gilbertson, Tetrahedron Lett. 2003, 44, 953-955. 24. F. Menges; M. Neuburger; A. Pfaltz, Org. Lett. 2002, 4, 4713-4716. 25. N. End; C. Stoessel; U. Berens; P. di Pietro; P. G. Cozzi, Tetrahedron: Asymmetry 2004, 15, 2235-2239. 26. C.-J. Wang; X. Sun; X. Zhang, Angew. Chem., Int. Ed. 2005, 44, 4933-4935. 27. A. Pfaltz; J. Blankenstein; R. Hilgraf; E. Hörmann; S. McIntyre; F. Menges; M. Schönleber; S. P. Smidt; B. Wüstenberg; N. Zimmermann, Adv. Synth. Catal. 2003, 345, 33-44. 28. W. J. Drury, III; N. Zimmermann; M. Keenan; M. Hayashi; S. Kaiser; R. Goddard; A. Pfaltz, Angew. Chem., Int. Ed. 2004, 43, 70-74.

65 29. M. Diéguez; J. Mazuela; O. Pàmies; J. J. Verendel; P. G. Andersson, J. Am. Chem. Soc. 2008, 130, 7208-7209. 30. M. Diéguez; O. Pàmies; J. J. Verendel; P. G. Andersson, Chem. Commun. 2008, 2008, 3888-3890. 31. P. Kaukoranta; M. Engman; C. Hedberg; J. Bergquist; P. G. Andersson, Adv. Synth. Catal. 2008, 350, 1168-1176. 32. P. Cheruku; A. Paptchikhine; M. Ali; J.-M. Neudörfl; P. G. Andersson, Org. Bio- mol. Chem. 2008, 6, 366-373. 33. J. J. Verendel; P. G. Andersson, Dalton Trans. 2007, 47, 5603-5610. 34. C. Hedberg; K. Källström; P. Brandt; L. K. Hansen; P. G. Andersson, J. Am. Chem. Soc. 2006, 128, 2995-3001. 35. A. Trifonova; J. S. Diesen; P. G. Andersson, Chem.- Eur. J. 2006, 12, 2318-2328. 36. A. Trifonova; J. S. Diesen; C. J. Chapman; P. G. Andersson, Org. Lett. 2004, 6, 3825-3827. 37. K. Källström; C. Hedberg; P. Brandt; A. Bayer; P. G. Andersson, J. Am. Chem. Soc. 2004, 126, 14308-14309. 38. Y. Zhu; K. Burgess, J. Am. Chem. Soc. 2008, 130, 8894-8895. 39. Y. Zhu; K. Burgess, Adv. Synth. Catal. 2008, 350, 979-983. 40. S. Li; S.-F. Zhu; C.-M. Zhang; S. Song; Q.-L. Zhou, J. Am. Chem. Soc. 2008, 130, 8584-8585. 41. S.-M. Lu; X.-W. Han; Y.-G. Zhou, Adv. Synth. Catal. 2004, 346, 909-912. 42. N. S. Goulioukina; T. M. Dolgina; G. N. Bondarenko; I. P. Beletskaya; M. M. Ilyin; V. A. Davankov; A. Pfaltz, Tetrahedron: Asymmetry 2003, 14, 1397-1401. 43. P. Cheruku; T. L. Church; A. Trifonova; T. Wartmann; P. G. Andersson, Tetrahe- dron Lett. 2008, 49, 7290-7293. 44. P. Cheruku; J. S. Diesen; P. G. Andersson, J. Am. Chem. Soc. 2008, 130, 5595-5599. 45. P. Cheruku; S. Gohil; P. G. Andersson, Org. Lett. 2007, 9, 1659-1661. 46. M. Engman; J. S. Diesen; A. Paptchikhine; P. G. Andersson, J. Am. Chem. Soc. 2007, 129, 4536-4537. 47. K. Källström; I. J. Munslow; C. Hedberg; P. G. Andersson, Adv. Synth. Catal. 2006, 348, 2575-2578. 48. D. Seebach, Angew. Chem., Int. Ed. 1990, 29, 1320-1367. 49. A. Bondi, J. Phys. Chem. 1964, 68, 441-451. 50. D. O'Hagan, Chem. Soc. Rev. 2008, 37, 308-319. 51. S. N. Pieniazek; K. N. Houk, Angew. Chem. Int. Ed. 2006, 45, 1442-1445. 52. H. Plenio, ChemBioChem 2004, 5, 650-655. 53. T. Ritter; M. W. Day; R. H. Grubbs, J. Am. Chem. Soc. 2006, 128, 11768-11769. 54. S. Purser; P. R. Moore; S. Swallow; V. Gouverneur, Chem. Soc. Rev. 2008, 37, 320–330. 55. K. Mikami; Y. Itoh; M. Yamanaka, Chem. Rev. 2004, 104, 1-16. 56. A. M. Thayer, Chem. Eng. News 2006, 84, 15-24. 57. Quotation by M. Schlosser in the article “Fabulous Fluorine” by Thayer, A. M. Chem. Eng. News 2006, June 5, 15–24. 58. G. A. Patani; E. J. LaVoie, Chem. Rev. 1996, 96, 3147-3176. 59. H. J. Böhm; D. Banner; S. Bendels; M. Kansy; B. Kuhn; K. Müller; U. Obst- Sander; M. Stahl, ChemBioChem 2004, 5, 60. J. L. Grem, Invest. New Drugs 2000, 18, 299-313.

66 61. J.-A. Ma; D. Cahard, Tetrahedron: Asymmetry 2004, 15, 1007-1011. 62. D. Y. Kim; E. J. Park, Org. Lett. 2002, 4, 545-547. 63. L. Hintermann; A. Togni, Angew. Chem., Int. Ed. 2000, 39, 4359-4362. 64. N. Shibata; T. Ishimaru; T. Nagai; J. Kohno; T. Toru, Synlett 2004, 10, 1703-1706. 65. Y. Huang; A. M. Walji; C. H. Larsen; D. W. C. MacMillan, J. Am. Chem. Soc. 2005, 127, 15051-15053. 66. Y. Hamashima; T. Suzuki; Y. Shimura; T. Shimizu; N. Umebayashi; T. Tamura; N. Sasamoto; M. Sodeoka, Tetrahedron Lett. 2005, 46, 1447-1450. 67. M. P. Brochu; S. P. Brown; D. W. C. MacMillan, J. Am. Chem. Soc. 2004, 126, 4108-4109. 68. T. D. Beeson; D. W. C. MacMillan, J. Am. Chem. Soc. 2005, 127, 8826-8828. 69. T. D. Nelson; M. H. Kress; S. W. Krska; J. V. Mitten; Y. Sun Process for prepara- tion of chiral piperidines via asymmetric hydrogenation of dehydropiperidines us- ing metal chiral phosphine catalyst complexes. WO 2006069287 A1, CAN 145:103564, 20051221., 2006. 70. M. Hudlicky, J. Fluorine Chem. 1979, 14, 189. 71. M. Hudlicky, J. Fluorine Chem. 1989, 44, 345. 72. T. Allmendinger; C. Dandois; B. Walliser, Tetrahedron Lett. 1991, 32, 2735-2736. 73. M. Engman; P. Cheruku; P. Tolstoy; J. Bergquist; S. F. Völker; P. G. Andersson, Adv. Synth. Catal. 2009, 351, 375-378. 74. S. Peng; F.-L. Qing; Y.-Q. Li; C.-M. Hu, J. Org. Chem. 2000, 65, 694-700. 75. D. Chevrie; T. Lequeux; J. P. Demoute; S. Pazenok, Tetrahedron Lett. 2003, 44, 8127. 76. X. Lei; G. Dutheuil; X. Pannecoucke; J.-C. Quirion, Org. Lett. 2004, 6, 2101-2104. 77. P. Coe; T. Waring; C. Mercier Process for the preparation of fluoro compounds from the corresponding amines. PCT/GB1997/001201, WO/1997/041083, 1997. 78. Jaguar, version 6.5; Schrodinger LLC: New York, 2005. 79. A. D. Becke, J. Chem. Phys. 1993, 98, 5648. 80. P. J. Stevens; F. J. Devlin; C. F. Chablowski; M. J. Frisch, J. Phys. Chem. 1994, 98, 11623. 81. J. P. Hay; W. R. Wadt, J. Chem. Phys. 1985, 82, 299. 82. T. Eicher; S. Hauptmann, The Chemistry of Heterocycles. 2nd edn. ed.; Wiley- VCH: Weinheim, 2003. 83. M. M. Kayser; J. Zhu; D. L. Hooper, Can. J. Chem. 1997, 75, 1315-1321. 84. M. Schlosser, Angew. Chem., Int. Ed. 1998, 37, 1496-1513. 85. B. E. Smart, J. Fluorine Chem. 2001, 109, 3-11. 86. M. Hird; J. W. Goodby; R. A. Lewis; K. J. Toyne, Mol. Cryst. Liq. Cryst. 2003, 401, 115-132. 87. R. Dabrowski, Mol. Cryst. Liq. Cryst. 2004, 421, 1-21. 88. W. R. J. Dolbier, J. Fluorine Chem. 2005, 126, 157-163. 89. M. N. Romanelli; F. Gualtieri; R. Mannhold; A. Chiarini, Farmaco 1989, 44, 449-464. 90. H. Kurihara; S. Yoshida; T. Tsuruoka; H. Yamamoto; S. Fukuyasu Jpn. Kokai Tokkyo Koho JP02306962 A19901220 CAN 114:185939 AN 1991:185939., 1990. 91. M. Vincent; G. Remond; M. Laubie Ger. Offen. DE2731299 19780119 CAN 88:136657 AN 1978:136657, 1978. 92. M. Vincent; G. Remond; B. Portevin; Y. Herve; J. Lepagnol Fr. Demande FR 91- 11893 19910927 CAN 119:139121 AN 1993:539121, 1993. 93. P. Bravo; G. Resnati, Tetrahedron: Asymmetry 1990, 1, 661-692.

67 94. T. Kitazume; T. Yamazaki, Enzymatically Controlled Reactions of Organofluorine Compounds. In Organofluorine Chemistry - Techniques and Synthons, Burton, D. J.; Chambers, R. D., Eds. Springer-Verlag: Berlin, 1997; pp 91-130. 95. K. Iseki, Tetrahedron 1998, 54, 13887-13914. 96. J.-A. Ma; D. Cahard, Chem. Rev. 2004, 104, 6119-6146. 97. M. Shimizu; T. Hiyama, Angew. Chem. Int. Ed. 2005, 44, 214-231. 98. M. Kimura; T. Yamazaki; T. Kitazume; T. Kubota, Org. Lett. 2004, 6, 4651-4654. 99. K. Kundu; J. V. McCullagh; A. T. Morehead, J. Am. Chem. Soc. 2005, 127, 16042-16043. 100. K. Tojo; Y. Aoki; M. Yasutake; T. Hirose, J. Fluorine Chem. 2006, 127, 620-626. 101. R. Motoki; D. Tomita; M. Kanai; M. Shibasaki, Tetrahedron Lett. 2006, 47, 8083-8086. 102. T. Konno; T. Takehana; M. Mishima; T. Ishihara, J. Org. Chem. 2006, 71, 3545-3550. 103. H. Sugimoto; S. Nakamura; Y. Shibata; N. Shibata; T. Toru, Tetrahedron Lett. 2006, 47, 1337-1340. 104. Y. Suzuki; T. Hagiwara; I. Kawamura; N. Okamura; T. Kitazume; M. Kakimoto; Y. Imai; Y. Ouchi; H. Takezoe; A. Fukuda, Liq. Cryst. 2006, 33, 1344-1349. 105. H. Nagao; Y. Yamane; T. Mukaiyama, Chem. Lett. 2007, 36, 666-667. 106. H. Nohira; S. Imamura; K. Sakai Jpn. Kokai Tokkyo Koho JP05229974 A19930907; CAN 120:91044 AN 1994:91044, 1993. 107. K. Iseki; Y. Kuroki; T. Nagai; K. Y., Chem. Pharm. Bull. 1996, 44, 477-480. 108. K. Iseki; Y. Kuroki; T. Nagai; K. Y., J. Fluorine Chem. 1994, 69, 5-6. 109. E. S. Rovner; A. J. Wein, Eur. Urol. 2002, 41, 6-14. 110. J. Wefer; M. C. Truss; U. Jonas, World J. Urol. 2001, 19, 312-318. 111. A. L. McRae; K. T. Brady, Expert Opin. Pharmacother. 2001, 2, 883-892. 112. M. Gordaliza; P. A. García; J. M. Miguel del Corral; M. A. Castro; M. A. Gómez- Zurita, Toxicon 2004, 44, 441-459. 113. S. Sörgel; N. Tokunaga; K. Sasaki; K. Okamoto; T. Hayashi, Org. Lett. 2008, 10, 589-592. 114. T. Itoh; T. Mase; T. Nishikata; T. Iyama; H. Tachikawa; Y. Kobayashi; Y. Yama- moto; N. Miyaura, Tetrahedron 2006, 62, 9610-9621. 115. N. Tokunaga; T. Hayashi, Tetrahedron: Asymmetry 2006, 17, 607-613. 116. J.-F. Paquin; C. R. J. Stephenson; C. Defieber; E. M. Carreira, Org. Lett. 2005, 7, 3821-3824. 117. T. Nishikata; Y. Yamamoto; N. Miyara, Tetrahedron Lett. 2007, 48, 4007-4010. 118. T. Nishikata; Y. Yamamoto; I. D. Gridnev; N. Miyaura, Organometallics 2005, 24, 5025-5032. 119. A. R. Jagdale; A. Sudalai, Tetrahedron Lett. 2008, 49, 3790-3793. 120. H. Matsuzawa; Y. Miyake; Y. Nishibayashi, Angew. Chem., Int. Ed. 2007, 46, 6488-6491. 121. F. Ulgheri; M. Marchetti; O. Piccolo, J. Org. Chem. 2007, 72, 6056-6059. 122. C. Hedberg; P. G. Andersson, Adv. Synth. Catal. 2005, 347, 662-666. 123. G. Chen; N. Tokunaga; T. Hayashi, Org. Lett. 2005, 7, 2285-2288. 124. C. Selensky; T. R. R. Pettus, J. Org. Chem. 2004, 69, 9196-9203. 125. P. G. Andersson; H. E. Schink; K. J. Österlund, J. Org. Chem. 1998, 63, 8067-8070. 126. A. Alexakis; S. El Hajjaji; D. Polet; X. Rathgeb, Org. Lett. 2007, 9, 3393-3395.

68 127. H. M. L. Davies; D. G. Stafford; T. Hansen, Org. Lett. 1999, 1, 233-236. 128. G. J. Quallich; T. M. Woodall, Tetrahedron 1992, 48, 10239-10248. 129. W. M. Welch; A. R. Kraska; R. Sarges; B. K. Koe, J. Med. Chem. 1984, 27, 1508-1515. 130. J. A. Wilkinson; S. B. Rossington; S. Ducki; J. Leonard; N. Hussain, Tetrahedron 2006, 62, 1833-1844. 131. K. Okamoto; Y. Nishibayashi; S. Uemura; A. Toshimitsu, Angew. Chem., Int. Ed. 2005, 44, 3588-3591. 132. J. A. Wilkinson; S. B. Rossington; S. Ducki; J. Leonard; N. Hussain, Tetrahedron: Asymmetry 2004, 15, 3011-3013. 133. L. Prat; G. Dupas; J. Duflos; J. Quéguiner; J. Bourguignon; V. Levacher, Tetrahe- dron Lett. 2001, 42, 4515-4518. 134. Y. Hatanaka; T. Hiyama, J. Am. Chem. Soc. 1990, 112, 7793-7794. 135. T. C. Fessard; S. P. Andrews; H. Motoyoshi; E. M. Carreira, Angew. Chem., Int. Ed. 2007, 46, 9331-9334. 136. C. Hedberg; K. Källström; P. Brandt; L. K. Hansen; P. G. Andersson, Journal of the American Chemical Society 2006, 128, 2995-3001. 137. K. Källström; C. Hedberg; P. Brandt; A. Bayer; P. G. Andersson, Journal of the American Chemical Society 2004, 126, 14308-14309. 138. P. Brandt; C. Hedberg; P. G. Andersson, Chem.- Eur. J. 2003, 9, 339-347. 139. Y. Fan; X. Cui; K. Burgess; M. B. Hall, J. Am. Chem. Soc. 2004, 126, 16688-16689. 140. A. de Meijere; F. E. Diederich, Metal-Catalyzed Cross-Coupling Reactions. Wiley- VCH: Weinheim, 2004. 141. David C. Chatfield; A. Augsten; C. D'Cunha; E. Lewandowska; S. F. Wnuk, Eur. J. Org. Chem. 2004, 2004, 313-322. 142. K. Steert; I. El-Sayed; P. Van der Veken; A. Krishtal; C. Van Alsenoy; G. D. Westrop; J. C. Mottram; G. H. Coombs; K. Augustyns; A. Haemers, Bioorg. Med. Chem. Lett. 2007, 17, 6563.

69  ,   ,-                 )  . /  0 ( /      

       /   0 ( /      1 ,--  , 1  ((  (  / (* / - -2  /3  - /   -    4-   5  3   * 1 3  (     *(      (   .   -  (  / ,--  .   /   0 ( /      2 67   8 ( 1 ++"1   3  -(* (    9 -  (  / ,--  .   /   0 ( /      :2;

.*( ) -(*  2((2 ( ) * ))(() $$&+&