Asymmetric Catalysis: Ligand Design and …kth.diva-portal.org/smash/get/diva2:8703/FULLTEXT01.pdfAsymmetric Catalysis: Ligand Design and Microwave Acceleration Ulf Bremberg Civ. Ing

i

ISBN KTH/IOK/FR--00/59--SE ISSN 1100-7974 TRITA-IOK Forskningsrapport 2000:59

Royal Institute of Technology Department of Chemistry Organic Chemistry

Asymmetric Catalysis: Ligand Design and Microwave Acceleration

Ulf Bremberg

Civ. Ing.

Akademisk avhandling som med tillstånd av Kungl Tekniska Högskolan i Stockholm framlägges till offentlig granskning för avläggande av teknologie doktorsexamen i organisk kemi, fredagen den 19:e maj, 2000, kl. 10.00 i sal K1, Teknikringen 56, KTH, Stockholm. Avhandlingen försvaras på engelska.

ii

Asymmetric Catalysis: Ligand Design and Microwave Acceleration

Ulf Bremberg

Abstract This thesis deals partly with the design and synthesis of ligands for use in asymmetric catalysis, and partly with the application of microwave heating on metal-based asymmetric catalytic reactions.

Enantiomerically pure pyridyl alcohols and bipyridyl alcohols were synthesized from the chiral pool for future use in asymmetric catalysis. Lithiated pyridines were reacted with several chiral electrophiles, yielding diastereomeric mixtures that could be separated without the use of resolution techniques. New pyridino- and quinolinooxazolines were synthesized and tested in palladium-catalyzed asymmetric allylation using 1,3-diphenyl-2-propenyl acetate and dimethyl malonate. The conformational preferences of the ligands in palladium complexes were studied with crystallography, 2D-NMR techniques and DFT calculations. Conclusions about how the chirality was transferred from the ligand to the substrate could be drawn from the conformational analysis. The effect of heating Pd- and Mo-catalyzed asymmetric allylic substitution reactions was investigated with oil bath heating and microwave irradiation. With a few exceptions, ligands with high room temperature selectivity were shown to retain their selectivity on heating. Reaction rates, catalyst stability and product selectivities of microwave-heated reactions were compared with those of reactions performed in oil bath. Palladium-catalyzed asymmetric allylation was studied with several ligand types, allylic substrates and nucleophiles. Some of the experimental procedures had to be adapted to microwave heating conditions. The procedure for asymmetric allylation catalyzed by bispyridylamide molybdenum complexes was developed into a one-pot microwave-mediated reaction. With microwaves, Mo(CO)6 could be used as an easily-handled metal source and inert conditions could be omitted. Derivatives of the bispyridylamide ligands were synthesized and tested with molybdenum as catalysts to investigate the effects of substituents on the pyridine ring. Keywords: ligand, asymmetric catalysis, pyridyl alcohols, oxazolines, conformational study, Pd-allyl, fast chemistry, microwave chemistry, Mo-allyl, bispyridylamides.

iii

Contents ♦ Abbreviations used in the text iv ♦ Introduction 1

• Scope of the Thesis 2

♦ PART I 4 • Synthesis of Ligands for Asymmetric Catalysis 4 Synthesis of Pyridyl Alcohols as Ligands for Asymmetric CatalysisI 5

• Palladium-Catalyzed Asymmetric Allylic Substitution 9 New pyridino- and quinolinooxazolinesII 12

• Structure of the Pyridylalcohol Moiety in Palladium ComplexesIII 14 X-ray Structure 15 DFT Calculations 15 NMR measurements 18 Implications on the Selectivity of the Catalytic Reaction 20 The Catalytic Reaction with 1,3-Diphenylallyl Acetate,

a Model of the Enantioselective Step 20

♦ PART II 25 • Comparison of Ordinary and Microwave Heating Methods of Accelerating

Organic Reactions 25 Theoretical Survey 26 Temperature Measurements Under Microwave Irradiation 29 Advantages of Microwave Heating 30 Effects of Heating on Asymmetric Catalysis 30

• Palladium-Catalyzed Allylic Substitution under Microwave Irradiation 33 Equipment 34 Alkylations of 1,3-Diphenyl-2-propenyl Acetate with Dimethyl

Malonate Using N,N- and N,P-LigandsIV,V 34 Pd-catalyzed Allylic Alkylation Using P,P-LigandsV,VI 37 Conclusion and Discussion 41

• Mo-Catalyzed Allylic Substitution under Microwave Irradiation 42 Molybdenum Catalyzed Allylic Substitution 42 Asymmetric Allylic Substitution Catalyzed by Molybdenum 43 Microwave-Accelerated Molybdenum-Catalyzed Allylic SubstitutionsVII 44 Steric and Electronic Variations of the Bispyridylamide ligandsVIII 46

♦ Acknowledgements 51 ♦ List of Publications 52

iv

Abbreviations used in the text rt Room temperature TBDMSCl tert-Butyldimethylsilyl chloride TBAF Tetrabutylammonium fluoride µw Microwave irradiation BSA N,O-bis(Trimethylsilyl)acetamide DBU 1,8-Diazabicyclo[5.4.0]undec-7-ene dba trans, trans-Dibenzylideneacetone equiv Molar equivalents TS Transition state tan δ Loss tangent, a measure of how efficiently a liquid absorbs microwave radiation τ Relaxation time, a measure of how quickly the dipolar molecules in a

liquid can turn. ε´´ The dielectric loss, the loss factor ε´ Dielectric constant εs Dielectric constant at static electric field ε∞ Dielectric constant at infinitely high frequency of the electric field E Enantiomeric ratio of formation, i.e. E = kR/kS

ee Enantiomeric excess, i.e. ][][][][

SRSRee

+−

=

∆G# Gibbs free activation energy ∆H# Activation enthalpy ∆S# Activation entropy ∆∆G# Difference in Gibbs free activation energy between the pathways

leading to the two enantiomers ∆∆H# Difference in activation enthalpy between the pathways leading

to the two enantiomers ∆∆S# Difference in activation entropy between the pathways leading to the two enantiomers DFT Density Functional Theory, a method of calculating the electron

density of a molecule and the corresponding energy. B3LYP An example of a DFT method. ∆E Calculated electronic energy NOESY A 2D-NMR method, used to judge distances between atoms COSYgs A 2D-NMR method, used to detect homonuclear couplings, e. g. 1H-1H HMQCgs A 2D-NMR method, used to detect heteronuclear couplings,

in this thesis 1H to 13C

Introduction

- 1 -

Introduction Each asymmetric reaction is unique in the way the chirality is transferred. A blend of electronic and steric interactions between the reactants in the transition state determines the ratio of the product stereoisomers. There is an immense amount of literature on the subject,1 whereas this thesis will only attempt to give a brief introduction in order to allow a discussion of the results presented.

The use of enantiomerically pure substances is increasing, especially in pharmaceutical, agricultural and food industries.2 As all biological systems are inherently chiral, the two enantiomers of a given molecule often have different activity. The increasing knowledge of complex biological systems through biochemical research will no doubt keep up the demand for pure enantiomers. There are a number of approaches towards the access to enantiomerically pure substances: ♦ Resolution (separation of the enantiomers of a racemate)

• Derivatization (producing diastereomeric derivatives of the enantiomers and separation of the diastereomers by crystallization, chromatography, etc.)

• Chromatographic resolution • Kinetic resolution (i.e. subjecting a racemate to reaction conditions under which

one enantiomer reacts faster than the other one and separation of the product and starting material)

• Selective crystallization (when conglomerates are possible) ♦ Asymmetric synthesis

• Diastereoselective transformations of enantiomerically pure starting materials • Use of chiral auxiliaries, which can be attached and induce chirality, and later

be removed • Use of asymmetric catalysis Synthetic asymmetric catalysts Enzymatic transformations Use of catalytic antibodies

All of these methods have their advantages, but asymmetric catalysis has by far

the highest potential. A small amount of an enantiomerically pure catalyst can transfer its chiral information to a large amount of product. There is no need for removal and recycling of derivatizing agents or chiral auxiliaries; a simple extraction or precipitation operation is often sufficient to remove all traces of the catalyst. The desired enantiomer can be produced selectively, which eliminates the waste problem of partly producing the undesired enantiomer. As only a small amount of catalyst is needed as compared with the amount of starting material, it can also be a cheap method, especially when the catalyst can be recovered.

Introduction

- 2 -

Unlike most enzymes3 and catalytic antibodies,4 metal-based asymmetric catalysts are often stable to heating and can be used in a wider range of solvents. Synthetic metal-based catalysts can also be adapted to different transformations. This is also true of catalytic antibodies, but in order to prepare a catalytic antibody it must be possible to synthesize a transition state analogue.

There are disadvantages with asymmetric catalysis, however. In contrast to chiral chromatography, which is fairly general as regards the enantiomerically pure compounds that can be produced,5 metal-based asymmetric catalysts have to be optimized for each reaction; at best it may be useful throughout a certain substrate class.

The development of new catalytic systems is still mainly based on trial and error, and the knowledge of what constitutes a selective and active catalyst is far from complete. The use of asymmetric catalysis as an efficient and general method requires much more data than we have today. The most important task of this field of research is probably to increase the understanding of how chirality is transferred from catalyst to product and learn how to apply this knowledge in designing new catalytic systems. Scope of the Thesis A few “benchmark” reactions (see Scheme 1) have emerged. 6 They have little actual synthetic value, but can be used to compare ligand performance. When a sufficient number of ligands and their “benchmark” performances have been published, conclusions about the determining factors in other systems can be drawn.

H

O OH*ZnEt2, ligand

Ph Ph

OAc Pd(0), ligandnucleophile

Ph Ph

Nu*

O OH*

Metal, ligand 2-propanol

Scheme 1. Examples of “benchmark” reactions. During research on asymmetric catalysis, time is often spent on synthesizing new and hopefully highly selective ligands. This makes the field intimately connected with the development of organic synthesis. The synthesis and “benchmark” performance of some new ligands, as well as a study of catalyst structure, will be discussed in the first part of this thesis.

Introduction

- 3 -

One aspect of asymmetric catalysis that has so far been given rather little attention is the reaction time. Some highly selective procedures require several days, which, at best, is an annoyance and at worst makes a project impossible when time is essential. One such example is combinatorial chemistry, in which chemists want to maximize the number of compounds synthesized. An instance of this is the pharmaceutical industry, in which the activity screening methodologies have developed very quickly and new synthetic strategies are needed to avoid being limiting in the effort to reach new drugs.7 Lower selectivity is often encountered when reaction mixtures are heated. However, there are situations in which the reaction time can be reduced considerably without sacrificing the selectivity, as will be shown in the second part of the thesis. 1 Examples of relevant reviews: a) Seyden-Penne, J. Chiral Auxiliaries and Ligands in Asymmetric Synthesis, John Wiley & Sons, Inc., New York, 1995. b) Houben-Weyl, Methods of Organic Chemistry – Stereoselective Synthesis, Volume E21a-f, Georg Thieme Verlag, Stuttgart, 1995-1996. c) Ojima, I. Catalytic Asymmetric Synthesis, VCH, New York, 1993. 2 a) Eliel, E. L.; Wilen, S. H. Stereochemistry of Organic Compounds, John Wiley & Sons, Inc., New York, 1994. b) Gawley, R. E.; Aubé, J. Principles of Asymmetric Synthesis, Elsevier Science Ltd, Oxford, 1996. 3 Faber, K. Biotransformations in Organic Chemistry, Springer-Verlag, Berlin, 1995. 4 a) Schultz, P. G.; Lerner, R. A. Acc. Chem. Res. 1993, 26, 391. b) Schultz, P. G., Lerner, R. A. Acc. Chem. Res. 1989, 22, 287. 5 Allenmark, S. G. Chromatographic Enantioseparation, Ellis Horwood Ltd., Chichester, 1988. 6 The term is used by e.g. Trost, B. M. and Van Vranken, D. L. in Chem. Rev. 1996, 96, 395. 7 a) Dolle, R. E.; Nelson, K. H. Jr. J. Comb. Chem. 1999, 1, 235. b) Larhed, M.; Lindeberg, G.; Hallberg, A. Tetrahedron Lett., 1996, 37, 8219.

Part I: Ligand Design

- 4 -

PART I Synthesis of Ligands for Asymmetric Catalysis The manner in which chirality is transferred from the ligand to the substrate is strongly dependent on the coordination geometry of the metal involved in bringing the reactants together. Metals can have from 2 (e.g. Zn, Hg) up to 10 (e.g. U, Th) coordination sites in geometries ranging from linear, through square planar, tetrahedral and octahedral to more complex configurations.1 The substrates may coordinate to the metal using one or several bonds. In some reactions the metal activates the reactant(s) as a Lewis acid, but it may also be redox-active, channeling electrons from one substrate to another. The catalytic properties of a metal can be altered considerably by the ligand. Even more important: ligands make up the steric environment around the catalytic center. The donor atom of the ligand can then be used to tune the electronic properties of the metal center. This is especially important if the metal is redox-active, then the ligand can be designed to stabilize the oxidation states through which the metal cycles in the catalysis. The number of possible ways for the reactants to coordinate can sometimes be reduced by a symmetric ligand. The two coordination sites are different in a C1-symmetric square planar complex, while they are identical in the complex with a C2-symmetric ligand, as shown in Figure 1.

MY X

MX Y

MX Y

MY X

C1 symmetry C2 symmetry

==

Figure 1. Illustration of the effect of ligand symmetry on the two coordination sites X and Y. With C1 symmetry the sites are diastereotopic, whereas with C2 symmetry, a 180° rotation around the axis indicated shows that the two coordination sites are identical. In order to obtain the tools for understanding the catalytic reaction, one has to determine the catalytic cycle. The catalytic cycle may be difficult to establish experimentally, however, as the intermediates not involved in the rate-determining step are rapidly consumed and may be virtually undetectable. It is often easier to detect a dead-end side-reaction intermediate than the intermediates leading to the product.2 The disparity in requirements of the ligand(s) in different asymmetric catalytic reactions makes it virtually impossible to use a general approach to ligand design. Rational designing is often impeded by lack of knowledge of the catalytic cycle and the exact role of the catalyst in the stereodetermining step. The design of successful ligands is thus a joint effort of mechanistic and structural studies of metal-catalyzed reactions and empirical ligand syntheses, followed by testing for selectivity. The mechanistic studies provide a base for


- 5 -

understanding the selectivities provided by various ligands and the synthetic development of new ligands makes the theoretical understanding useful. Synthesis of Pyridyl Alcohols for Use as Ligands in Asymmetric CatalysisI 2-(1’-Hydroxyalkyl)pyridines are interesting as ligands for a number of asymmetric catalytic reactions. They have been used for the addition of diethylzinc to aldehydes,3 for the conjugate addition of diethylzinc to α,β-unsaturated ketones catalyzed by nickel,4 and for the titanium-catalyzed epoxidation of alkenes.5 The combination of the rigid σ-donor/π-acceptor pyridine structure with the alcohol moiety, which can be deprotonated or further functionalized, constitutes a flexible framework for building a variety of ligands. Procedures for converting the pyridyl alcohols to a number of other ligands, e.g. pyridyl phosphines6 and pyridyl phosphites, have also been published.6b,c

Methods of synthesizing enantiomerically pure pyridyl alcohols are thus highly valuable. As free ligands, pyridyl alcohols tend to form an internal hydrogen bond between the pyridine nitrogen and the hydroxy proton. Alkyl ethers of pyridyl alcohols, on the other hand, have been shown to be aligned by a stereo-electronic effect in combination with electron pair repulsion, so that the alkoxymethyl group lies in the pyridine plane, with the alkoxyl unit anti to the nitrogen atom, as shown in Figure 2.7

NOMe

NOMe

Pyridine free electron pair stabilization by donation into antibonding orbital

Free electron pair repulsion

NOMe

Preferred geometry

Figure 2. The stereo-electronic effect and electron pair repulsion of 2-(alkoxymethyl)pyridines aligns the alkoxymethyl substituent in the pyridine plane, pointing away from the pyridine nitrogen. Considerable effort has been invested in finding selective syntheses and methods of separating enantiomerically pure pyridyl alcohols. Asymmetric synthetic pathways have been developed, e.g. reductions of the corresponding ketones,8 diethyl zink additions9 and Baylis-Hillman reactions,10 but most of these methods give less than full enantioselectivity, which makes separation of scalemic mixtures necessary.11 Enzymatic transformations,12 microbial reduction13 and catalysis by antibodies14 have also been used, but mostly with a maximum yield of 50% of each mirror image or with only one accessible enantiomer. Synthetic methods based on enantiomerically pure compounds from the chiral pool have advantages in that the starting materials are relatively cheap, and that the resulting diastereomeric products can be separated without resolution.15 However, low overall yields are often observed and usually only one enantiomer from the chiral pool is readily available.


- 6 -

During an earlier synthesis of ligands 1 and 23b, 15b, 46 (Figure 3) we recognized the convenience of introducing further stereocenters into the ligand structure from the chiral pool. These additional stereocenters can be quite useful: they may induce diastereoselectivity in later transformations and also enhance the chirality transfer from the ligand in the catalysis. Especially meriting is the fact that the method avoids enantiomer separation or derivatization. The method is, thus, both a cheap and a fairly general method, as almost any enantiomerically pure compound can be used, provided that it has the right functional groups.

N O

NOMe

RR´

1a R=H, R´=tBu1b R=tBu, R´=H

R´RN

2a R=H, R´=OH2b R=OH, R´=H

Figure 3. Pyridyl alcohols derived from the chiral pool and previously synthesized in our group. We investigated the introduction of structural elements from the chiral pool by reacting 2-lithiopyridine with enantiomerically pure aldehydes, esters and nitriles.3b,

15b, 46 The reaction with an aldehyde yielded the pyridyl alcohol in one step (Figure 4), while the products from reactions with esters had to be reduced in a subsequent step (Figure 5).

NO

O

OHO

OON Li N

OO

OH

3a, 45% 3b, 11%

+

Et2O -78 oC

Figure 4. The reaction of 2-lithiopyridine with an aldehyde derived from D-mannitol.16


- 7 -

N Li

MeO

O

OMe

Et2O-78 oC N Li

MeOPh

O

OMe

Et2O-78 oC

NO

OMe

N PhO

OMe

NaBH4, Et2O, MeOH –78 °C to rt.

N Ph

OMe

OH5, 49%

N

OMe

OHN

OMe

OH

+

1) NaBH4, MeOH, 0 °C to rt2) TBDMSCl, imidazole, DMF, rt3) TBAF, THF, 0 °C to rt

4a, 13% 4b, 4% Figure 5. The nucleophilic addition of 2-lithiopyridine to esters derived from (S)-lactic acid and (S)-mandelic acid, followed by reduction. The main drawback of the procedure was the low yields, especially with the pyridyl alcohols based on lactic acid (see Figure 5), as these had to be silylated to make a separation by chromatography possible. Both diastereomers could be obtained in the two examples. When the reaction with the mandelic acid derivative was performed at room temperature, a mixture of diastereomers was obtained. The temperature could be kept low to afford one main isomer, or increased to room temperature to yield a mixture. The synthetic methodology was also used to make a number of bipyridine derivatives, see Figure 6.

NBrO

O

OHNBr Li NBr

OO

OH

N

BuSiO (Me)2

O

O6a

6a, 42% 6b, 10%

8, 100%

Et2O –78 °C to rt

7, 51%

+

2

N

HO O

O

2

OO

O+

DMF, rt

TBDMSCl imidazole

DMF, 70 °C

PPh3, NiCl2 Zn

THF, RTTBAF

t


- 8 -

NBr

OMe

OSitBu(Me)2

NBr

OMe

OSitBu(Me)2

+

N

BuSiO (Me)2 2

DMF, 70 °C

PPh3, NiCl2, Zn

t

OMe

THF, rtTBAF N

HO 2

OMe

NBr Li Et2O –78 °C to rtO

OMeMeO NaBH4

MeOH0 °C to rt

NBr

OMe

OH

DMF, rt

TBDMSCl imidazole

+

9, 75 %

10a, 27 % 10b, 25 %

11, 55 % 12, 73 %

10a

N

BuSiO (Me)2

Ph2

DMF, 70 °C

PPh3, NiCl2, Zn

t

OMe

THF, rtTBAF N

HO Ph 2

OMe

NBr Li Et2O –78 °C to rt

PhO

OMeMeO NaBH4

MeOH, Et2O-78 °C to rt

NBr Ph

OMe

OH

DMF, rt

TBDMSCl imidazole

+

13, 23%

14, 31 % 15, 73% Figure 6. Syntheses of chiral bipyridines derived from D-mannitol, (S)-lactic acid and (S)-mandelic acid. The method presented gives access to a number of pyridyl alcohols without costly reagents or resolution techniques. Drawbacks are the number of steps and the low overall yields. In the near future the ligands will be tested in catalysis.


- 9 -

Palladium-Catalyzed Asymmetric Allylic Substitution Palladium is one of the most versatile transition metals in catalytic organometallic chemistry. Reactions catalyzed by Pd are the electrophilic activation of various unsaturated compounds, couplings between aryl and alkenyl halides and masked carbanions, carbonylations, hydrogenations/dehydrogenations, etc.17 The nucleophilic substitution of palladium allyl complexes was first described by Tsuji et al. as a stoichiometric procedure in 1965,18 followed by catalytic procedures in 1970 by Atkins et al. 19a and Hata et al.19b The asymmetric version was also first performed with stoichiometric amounts of palladium in 1973 by Trost and Dietsche, 20 and later developed to a catalytic procedure in 1977 by Trost and Strege.21 Since these early reports, the asymmetric palladium-catalyzed allylic substitution had been developed into a versatile transformation, able to achieve high selectivities with several classes of substrates.22 A broad range of substrates can be used: the allylic moiety can have as few or as many substituents as desired and leaving groups can be halides, carboxylates, carbonates, phosphates, sulfinates, alkoxides, hydroxide, nitrite, tertiary amines or cyanide.17a When stabilized carbon or heteroatom nucleophiles are used, their attack occurs from the face of the allyl moiety opposite to the palladium atom. Hard carbon nucleophiles, however, are thought first to coordinate to palladium and then to be reductively eliminated with the allyl group, see Figure 7.

PdL L

Nu-

PdL L

Nu

+-Nu

PdL Nu

Red. elim.+ L

PdL L

Nu

- L

Hard nucleophilesSoft nucleophiles Figure 7. The mechanistic pathways of soft and hard nucleophiles in the reaction with the palladium-allyl complex. The catalytic cycle of palladium-catalyzed allylic alkylation with soft nucleophiles is fairly well-known, as displayed in Figure 8.23


- 10 -

Pd(0)L L

Pd(II)L L

Pd(0)L L

OAc

OAc

NuH

Pd(0)L L

Nu

OAc

Nu

HOAc Figure 8. The catalytic cycle of palladium-catalyzed allylic substitution with soft nucleophiles. Nucleophiles usually attack the least substituted end of the allyl complex, but with certain ligands this preference is reversed,24 and under certain conditions the central carbon may also be attacked.25 A number of metals other than palladium are able to catalyze this reaction, although palladium still seems to be the one most employed. Mo26 and W27 can be useful when an attack on the most substituted end of the allyl moiety is desired. Ni,28 Ir29 and Pt30 are generally not as reactive as Pd, but may exhibit a different regioselectivity. Less studied but still potentially interesting metals for catalysis of this reaction are Rh,31 Ru32 and Fe.33 The coordination situation of Pd during the catalytic cycle is relatively straight-forward. Pd(II) prefers square planar geometry and Pd(0) can be trigonal or tetrahedral. As the allyl moiety uses two coordination sites, there are two left for the ligand. Ligands used for asymmetric catalysis are based on P-, N-, S-, O- and C-donor atoms of the monodentate or bidentate type (see Figure 9).

HNNHOO

PPh2 PPh2

PPh2 N

O

R

Fe Ph2 P

NN 1-Adamantyl

Figure 9. Some successful ligands in the palladium-catalyzed asymmetric allylic alkylation.34, 35, 36 Depending on the choice of substrates, the stereodetermining step can be either the formation of the allyl compex or the nucleophilic attack, as displayed in Figure 10.


- 11 -

Pd+

Nua b

O

O

Pd+

a b

Pd(0)

a b

LGLG

Pd(0)a b

a b

LG

Enantiotopic allyl carbons

Enantiotopicleaving groups

Enantiotopic facesof the alkene

Enantiotopic facesof the nucleophile

Figure 10. Illustration of different principles of asymmetric induction in the palladium-catalyzed allylic substitution. The only situation that has been studied in this thesis is the nucleophilic attack on meso-allyl complexes. It is surprising that so many ligands are successful in inducing chirality in this step, although the nucleophile approaches the metal center from the side opposite to the ligand. The steric interaction between the ligand and the nucleophile is normally so weak that the origin of selectivity has to be sought elsewhere. Many highly selective C1-symmetric ligands have different donor atoms, as exemplified in Figure 11. The allyl moiety in such complexes has been shown by 13C-NMR shifts to be electronically unsymmetric. The reactivity of an allyl carbon atom trans to a phosphorus ligand has been shown to be significantly higher than an allyl carbon trans to a nitrogen ligand37 (the reactivity ratio has been calculated38 to 24:1). Different donor atoms in a bidentate ligand could thus be used to bind the allyl group in a specific fashion, and then specifically direct the nucleophile selectively trans to one allyl carbon. Steric influence of the ligand on the allyl group can give a similar effect. If one terminal substituent of the allyl group is sterically more repelled than the other, the palladium-carbon bonds become different in length. According to one model, a 0.01 Å difference in palladium-carbon bond length will make the carbon with the longest bond approximately twice as reactive as the other one.39 Studies of the catalysis are complicated by the dynamic equilibria, in which the palladium-allyl complex is involved. Apparent allyl rotation may either occur via an η3-η1-rotation-η3 process of the allyl group,40 bond breaking to one of the ligands, rotation and reformation of the ligand-palladium bond,41 or via pseudorotation of a pentacoordinated intermediate palladium species.42 This isomerisation has been shown to be much faster than the nucleophilic attack in the catalytic cycle.43 The attack of an external Pd(0)-complex on the Pd(II)-allyl group and the reformation of the allylic starting material by nucleophilic attack of the leaving group may also quickly scramble the stereochemistry of the initial complex.44 Non-cyclic allyl groups also exhibit a rapid interconversion of anti and syn configurations.22a The situation can be simplified by using C2-symmetric ligands, as this makes the apparent allyl rotation and attack of an external Pd(0) result in identical complexes.


- 12 -

However, C2-symmetry is not a requirement for a highly selective ligand, as can be seen in Figure 11.

O

S

PPN

SBn

NH

O

NH

O

PPh2PPh2

OO

PO

N

O O

N N

O

PPh2 Ph2P

Figure 11. Recent examples of highly selective ligands for asymmetric allylic alkylation.45 New pyridino- and quinolinooxazolinesII Previous results obtained with ligands synthesized in our group46 made us interested in expanding the number of ligands based on pyridinooxazolines. The catalytic results of palladium-catalyzed allylation with some of the ligands previously synthesized in the group, are displayed in Figure 12.

Ph Ph

OAc

Ph Ph

OMe

O

MeO

O

NO

NPh

[(η3-C3H5)PdCl]2, ligand

(MeOCO)2CH2, BSA, KOAc

16 17

NO

NPh

NO

NPh

NO

NPh

NO

NPh

NO

NPh

OHtBu

OH OMe

OMe

tBu tBu

tBu

18a, 50% ee

18d, 74% ee

18b, 90% ee

18e, 95% ee

18c, >99% ee

18f, 15% ee Figure 12. Pyridinooxazolines synthesized in our group (except 18d)48 for asymmetric allylic alkylations. The reactions were run under “benchmark” conditions47 for four days and the resulting enantioselectivity of 17 is shown for each ligand.


- 13 -

The selectivities of the palladium-catalyzed reactions with 1,3-diphenylallyl acetate (16) and dimethyl malonate shown in Figure 12, can be rationalized by considering the orientation of the hydroxymethyl/methoxymethyl unit of the ligand. When comparing the enantioselectivities using ligands 18b, 18c, 18e, and 18f, it is evident that the two substituents need opposite configurations to induce the highest selectivity. The change in selectivity from >99% ee with S-(CH(OMe)tBu (18c) to 15% ee with the R-configuration (18f) is especially remarkable. We synthesized the ligands shown in Figure 13 in order to investigate the importance of the 6-substituent of the pyridine ring and in particular the function of the hydroxymethyl/methoxymethyl substituent.

NNCR

NR

O

NPh

NOH

NOH

NC

NOHO

NPh

(R)-Phenylglycinolcat. CuCl2

19, R=H20, R=OH

21a, R=H21b, R=OH

22

neat100 °C, 10 mbar

NOMe

NCNaH, MeITHF, rt

m-CPBACHCl3, rt

TMSCNCH2Cl2, rflx

Cl NMe2

O

23 24

(R)-Phenylglycinolcat. CuCl2neat, 100 °C, 10 mbar

NaOMeMeOH, rt

(R)-Phenylglycinolcat. H2SO4CH2Cl2, rflx

25a NOMeO

NPh

25b

Figure 13. Synthesis of quinolino- and pyridinooxazolines 21a-b and 25a-b. In the palladium-catalyzed allylic alkylation under “benchmark” conditions, we obtained up to 88% ee (see Figure 14).


- 14 -

NO

NPh

NOH

O

NPh

NOMeO

NPh

NOHO

NPh

21a:99% yield73% ee

25a:93% yield88% ee

21b:0% conversion

25b:99% yield82% ee

Figure 14. Results obtained after 4 days at room temperature under “benchmark” conditions.47 The almost identical enantioselectivity of the quinolinooxazoline 21a (73% ee) and the 6-methyl pyridinooxazoline 18d (74% ee) synthesized by Chelucci,48 is probably due to similar steric hindrance. The hydroxymethyl ligand 25a and methoxymethyl 25b ligands offer higher selectivities. When the hydroxy group is locked in the coordination plane, as in 8-hydroxyquinolinooxazoline 21b, catalysis is inhibited. The conclusion that can be drawn from these results is that the substituent in the 6-position has a powerful influence on the selectivity. The unsubstituted 18a, yielding 50% ee, can be used as a reference for comparison with other pyridino- and quinolinooxazolines. When the steric hindrance of the substituent in the 6-position is increased, as in ligand 18d or 21a, the selectivity is improved to ca. 75% ee. With methoxymethyl 25b or hydroxymethyl 25a the selectivities are further increased to 82 and 88% ee, respectively. These selectivities indicate that the oxygen atom may not be entirely turned away from the palladium atom. Previous results show that a bulky substituent at the methyl carbon atom can further increase the selectivity to >99% ee (using 18c), or decrease it to 15% ee (using 18f). As the opposite configurations give high selectivities with the hydroxy- and the methoxy-ligands, these ligands could be expected to have different conformations. Structure of the Pyridylalcohol Moiety in Palladium ComplexesIII The intriguing effects of the ligands bearing a hydroxy/methoxy moiety in the palladium-catalyzed allylic alkylation, prompted us to make a structural investigation of the corresponding metal complexes. We used a combination of an X-ray structure, NOESY measurements and DFT calculations. Due to problems of obtaining useful crystals for X-ray investigations of the allyl complexes, we chose to study a simplified system derived from palladium dichloride. Simplifications were also necessary in the NMR-studies, as the 1H NMR signals of the diphenylallyl moieties extensively overlapped the signals of the ligands. The palladium dichloride complexes as well as the unsubstituted allyl complexes were studied by NOESY. DFT calculations were less restricted, so we chose to study the full complexes with some simplifications to gain calculation time. With computer


- 15 -

simulations we could study the elusive Pd-olefin complex that results after the nucleophilic attack and is most important for singling out the stereodetermining factors, but rarely studied experimentally.49 X-ray Structure Crystals of the complex of ligand 25a (Figure 14) with palladium chloride proved to contain two independent crystal forms. Two distinct conformations of the hydroxymethyl unit were present in each crystal form. Thus a total of four slightly different solid state conformations were found (see Figure 15 and Table 1).

A B

Figure 15. The two independent crystal forms, displaying the conformations of the hydroxymethyl unit. The displacement ellipsoids are drawn at the 30% probability level.

Dihedral angle N-C-C-O,

Crystal form A Probability Dihedral angle N-C-C-O,

Crystal form B Probability

159.5° 91.5% 159.1° 79.4% -94.3° 8.5% -92.8° 20.6%

Table 1. Dihedral angles of the four structures found. Unfortunately the angles shown in Table 1 do not make the role of the hydroxy moiety in the catalysis evident. The less probable angles close to -90° place the oxygen (and presumably the hydroxy hydrogen) close to one of the chlorine atoms, tentatively indicating a hydrogen bond. DFT Calculations In order to understand the conformations found in the solid state, we performed DFT calculations (B3LYP) on the structure that had been analyzed by X-ray crystallography. The results are shown in Figure 16 and Table 2.


- 16 -

Pd

Cl Cl

NN

Pd

ClCl

NN

Pd

ClCl

NN

C

O OO

O

OO

H

HH

D E Figure 16. The calculated minimal energy conformations (C and D) of the palladium dichloride complex of the ligand 25a, as well as the structure of the calculated TS (E) of the process connecting C and D. Structure Dihedral angles N-C-C-O

∆E (kJ/mol)

Relative values C -73° (0) D 178° 9.2

E (TS) 149° 18.4 Table 2. Dihedral angles of the calculated structures in Figure 17 and their corresponding electronic energies, relative to structure C. The calculated gas phase complexes apparently have conformations similar to those of the experimentally determined solid phase structures. The calculated structure with the smaller dihedral angle is favored, however, a fact that can be attributed to the absence of the intermolecular hydrogen bonding interactions observed in the crystal structure, which stabilize the anti-like conformation of the hydroxymethyl substituent. The structures of direct interest, the allyl and diphenylallyl palladium complexes of a simplified ligand 25a, were also shown to have two conformational minima each, again with lower calculated energies for those having the larger dihedral angle (see Table 3).

Complex Dihedral angle

N-C-C-O ∆E (kJ/mol)

Relative values

179°

(0) -71° 11.7

178°

(0) -85° 7.1

Table 3. Calculated dihedral angles and relative heats of formation of Pd-allyl complexes. The palladium dichloride and the palladium-allyl complexes of the methoxymethyl ligand 25b were also analyzed. They showed only one minimum each, close to the dihedral angle 180°.

NHO

N

O

Pd

NHO

N

O

Pd

Ph Ph


- 17 -

It is difficult to explain the enantioselectivities observed with these calculated conformational preferences. From these results it would appear that the methoxymethyl and hydroxymethyl units have the same conformation in the stereoselecting nucleophilic attack. This is contrary to the experimental results in the asymmetric catalysis, however, where it was found that opposite configurations were necessary for optimal selectivity with hydroxymethyl and methoxymethyl ligands. In search for an explanation, calculations were performed on the products formed by the nucleophilic attack. Fluoride was chosen as a simple model of the nucleophile to facilitate the calculations.* The resulting Pd(0)-olefin complex of ligand 25a was calculated to have two conformational minima of the hydroxymethyl unit (Figure 17).

G

Pd

NN

O

OH

F

Pd

NN

O

O

H

F

F Figure 17. Conformational minima of the Pd(0)-olefin complex after the nucleophilic attack.

Dihedral angle N-C-C-O

∆E (kJ/mol) Relative values

-57 (0) 179° 9

Table 4. Calculated relative energies of the conformational minima in Figure 18. The intriguing change in preference of conformation hints at an explanation of results from the asymmetric catalysis. The lower energy of the conformation with the smaller dihedral angle in the Pd(0)-complex can be expected to have an influence also on the TS of the nucleophilic attack, depending on whether the TS is early or late. The TS of the flouride nucleophilic attack on the allyl complex (see Figure 18) was subjected to DFT calculations (B3LYP) using the DPCM method as implemented by Gaussian 9850 to approximate a dichloromethane solution. The calculated TS can be used to estimate the conformation of the hydroxy moiety as well as the rotation of the allyl group. As the dihedral angle of the TS is almost identical (-55°) to that of the product olefin complex (-57°), it appears that the TS is late, at least regarding the angle of the hydroxy group. The allyl moiety has only rotated about 4° from the starting complex, however (the allyl group is expected to rotate 30° to the product olefin). * In the calculations, fluoride was shown to give a weakly exothermic nucleophilic attack, similar in thermodynamics to the one experimentally observed.


- 18 -

Figure 18. Stereoview of the calculated TS of the fluoride nucleophilic attack on the palladium allyl complex of a simplified version of ligand 25a. The N-C-C-O dihedral angle is -55°. NMR measurements NMR investigations of the palladium dichloride complexes in CDCl3 with standard COSYgs, HMQCgs and NOESY experiments made it possible to assign all the 1H signals. The NOESY spectra indicated that the hydroxy moiety of the ligand 25a was turned towards the palladium atom (see Figure 19). The NOESY spectra of the complex of the methoxy ligand 25b were less easy to interpret (see Figure 19). The conformation predicted by DFT calculations was not the only one possible.

NN

O

Pd

H5'

H5'

H4'

H3 H4

H5H2''

H3''

H4''

H5''

H6'' Cl Cl HOHH

NN

O

Pd

H5'

H5'

H4'

OHH

H3 H4

H5H2''

H3''

H4''

H5''

s

H6''w MeCl Cl m

w w

Figure 19. NOESY correlations between the hydroxy-/methoxymethyl unit and the rest of the molecule.† The NMR-spectra of the allyl complexes 26a and 26b of ligands 25a and 25b, respectively, were difficult to interpret as rapid apparent allyl rotation made the two rotamers coalesce at about room temperature. It was possible to freeze out the two complexes, but a considerably lower temperature was needed for the methoxy than for the hydroxy complex. The different coalescence temperatures, as shown in Figure 20, made it possible to estimate the activation energies of rotation to be 7.4 kJ/mol higher for complex 26a containing a hydroxy moiety than for complex 26b containing a methoxy unit. As the only difference between the two complexes is the hydroxy/methoxy functionality, the more difficult rotation of the allyl in complex 26a

† The magnitudes of the NOESY correlations are noted in comparison with the NOESY correlation between H4-H3/H5, which can be assumed to be constant for all complexes. Signals up to half the size of this reference signal are noted as weak (w), those larger than half this signal as medium (m) and those as large or larger as strong (s).


- 19 -

can be interpreted as being due to an interaction between the hydroxy group and the palladium-allyl moiety.

26a 26b

25 °C 25 °C

-17 °C 17 °C

-19 °C 14 °C

-21 °C 10 °C

-33 °C 0 °C Figure 20. Coalescence of the oxazoline protons 5H and

'5H in complexes 26a and 26b. The apparent allyl rotation can also be seen in the non-specific NOESY correlations between both ends of the allyl group and the hydroxy/methoxy moieties. This made it impossible to assign the two rotamers of complexes 26a and 26b, and correlations from both rotamers are shown in Figure 21. NOESY crosspeaks were observed between the hydroxy proton and the pyridine, but not between the hydroxy proton and the allyl group, indicating a preferred dihedral angle N-C-C-O around 180° in complex 26a. The spectrum of the methoxy complex 26b was more difficult to interpret, as the methoxy protons showed crosspeaks to the allyl as well as to the pyridine moiety. However, a stronger correlation to the pyridine pointed to a preferred dihedral angle N-C-C-O around 180° also in complex 26b.

NN

O

Pd

H5'

H5'

H4'

H3a H1a

H1sH3s

H2

OHH

H3 H4

H5H2''

H3''

H4''

H5''

H6'' Me

NN

O

Pd

H5'

H5'

H4'

H3a H1a

H1sH3s

H2

OHHH

H3 H4

H5H2''

H3''

H4''

H5''

H6''

m

sw

w

w/m

m

m

m

sm/s

m/s

m

w

sm

26a 26b Figure 21. NOESY correlations between the hydroxy-/methoxymethyl units, the allyl group and the pyridine of complexes 26a and 26b (see note * in Figure 19 for an explanation of w, m and s).


- 20 -

Implications in the Selectivity of the Catalytic Reaction Both calculations and NMR measurements show that the methoxy group prefers to be turned away from the coordinated palladium (dihedral angles around 180°) in all palladium complexes studied. The situation of the hydroxy complexes is less clear; when an allyl moiety is coordinated, both calculations and NOESY indicate dihedral angles around 180°, but in the olefin complex resulting after nucleophilic attack, calculations show that the hydroxy group is turned partly towards palladium. The slower apparent allyl rotation in the hydroxy complex, as observed in NMR, is also an indication of an interaction between hydroxy and the metal. The systems studied are only model systems of the highly selective ligands 18c and 18e. It can be expected that the steric clash between the tert-butyl substituent and the phenyl groups of the allyl residue should change the systems. The product olefins, extrapolated from the calculated structures in Figure 17 are shown in Figure 22. In accordance with the calculated TS in Figure 18, the TS involving 18e should have a hydroxy moiety conformation similar to that in the product olefin complex. Note that both ligands show C2-topology, despite the different absolute configurations at the picolinic carbon.

N

O

NPh

tBuPd HO

Ph Ph

Nu

N

O

NPh OMePd tBu

Ph Ph

Nu Figure 22. Extrapolated structures of the 1,3-diphenylallyl palladium complexes of ligands 18c and 18e after the nucleophilic attacks. The Catalytic Reaction with 1,3-Diphenylallyl Acetate, a Model of the Enantioselective Step

The question whether the reaction caused by the nucleophilic attack has an early or a late transition state has been discussed.45a,51 Some recent reports indicate that the transition state should be late.23b Even if this is not always correct, at least part of the steric constraints of the products must be considered when the diastereomeric pathways are analyzed.

The olefin-complex produced has been shown to be trigonally planar, with the double bond in the plane.52 This means that the allyl moiety will rotate around its Pd-bond from the starting complex to the product complex. In one report a partly rotated allyl complex is described, where the product on nucleophilic attack corresponds to the one expected when the rotation is continued.53 The steric hindrance imposed on the allyl group during rotation can be used to determine in which direction the rotation is most facile, and thus, which nucleophilic attack is most favored (Figure 23).54


- 21 -

PhPh

PdR

RPdR

R

Ph

Nu

Ph

PdR

R

Ph

Nu

Ph

a b

aNu b Nu

Figure 23. The allyl rotation during a reaction catalyzed by a C2-symmetric ligand. Another determining factor can be derived from the steric repulsion between one of the phenyl groups and one R-group, that will distort the allyl moiety from symmetric binding to palladium, as shown in Figure 24.

PhPh

PdR

R

aNub Nu

PhPh

PdR

R

Steric repulsion

Figure 24. Steric repulsion distorts the symmetric binding of the allylic moiety to palladium, weakening the bond to one of the allylic termini and causing a preference of attack via the b pathway. These two factors, the preferred rotation and the steric distortion, act together by disfavoring the a pathway and favoring the b pathway. The example in Figures 24 and 25 has been drawn with a C2-symmetric ligand, but the principle works equally well to explain why ligands 18c and 18e with C2-topology in the TS are more selective than 18b and 18f. 1 a) Shriver, D. F.; Atkins, P. W.; Langford, C. H. Inorganic Chemistry, 2nd Ed., Oxford University Press, Great Britain, 1994. b) Cotton, F. A.; Wilkinson, G. Advanced inorganic chemistry, 5th Ed., Wiley Interscience, New York, 1988. 2 One example of dominating intermediates that are not directly involved in the catalysis is the hydrogenation of double bonds with Wilkinsson’s catalyst, see e.g. Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles and Application of Organotransition Metal Chemistry, University Science Books, Mill Valley, 1987, p531. 3 a) Bolm, C.; Schlingloff, G.; Harms, K. Chem. Ber. 1992, 125, 1191. Ishizaki, M.; Fujita, K.; Shimamoto, M.; Hoshino, O. Tetrahedron: Asymmetry 1994, 5, 411. b) Macedo, E.; Moberg, C. Tetrahedron: Asymmetry 1995, 6, 549. 4 a) Bolm, C. Tetrahedron: Asymmetry 1991, 2, 701. b) Bolm, C.; Ewald, M.; Felder, M. Chem. Ber. 1992, 125, 1205. 5 Hawkins, J. M.; Sharpless, K. B. Tetrahedron Lett. 1987, 28, 2825. 6 a) Jiang, Q.; Van Plew, D.; Murtuza, S.; Zhang, X. Tetrahedron Lett. 1996, 37, 797. b) Sablong, R.; Newton, C.; Dierkes, P; Osborn, J. A. Tetrahedron Lett. 1996, 37, 4933. c) Sablong, R.; Osborn, J. A. Tetrahedron Lett. 1996, 37, 4937.


- 22 -

7 a) Moberg, C.; Adolfsson, H.; Wärnmark, K. Acta Chem. Scand. 1996, 50, 195. b) Moberg, C.; Adolfsson, H.; Wärnmark, K.; Norrby, P.-O.; Marstokk, K.-M.; Møllendal, H. Chem. Eur. J. 1996, 2, 136. 8 a) Soai, K.; Niwa, S.; Kobayashi, T. J. Chem. Soc., Chem. Commun. 1987, 801. b) Hull, K. G.; Visnick, M.; Tautz, W.; Sheffron, A. Tetrahedron 1997, 53, 12405. c) Demir, A. S.; Micitoglu, I.; Tanyeli, C.; Gülbeyaz, V. Tetrahedron: Asymmetry 1996, 7, 3359. d) Bolm, C.; Ewald, M.; Felder, M.; Schlingloff, G. Chem. Ber. 1992, 125, 1169. e) Barrett, A. G. M.; Seefeld, M. A.; White, A. J. P.; Williams, D. J. J. Org. Chem. 1996, 61, 2677. f) Chen, G-M.; Brown, H. C.; Ramachandran, P. V. J. Org. Chem. 1999, 64, 721. g) Carreño, M. C.; Ruano, J. L. G.; Martín, A. M.; Pedregal, C.; Rodriguez, J. H.; Rubio, A.; Sanchez, J.; Solladié, G. J. Org. Chem. 1990, 55, 2120. 9 a) Brunner, H.; Kürzinger, A. J. Organomet. Chem. 1988, 346, 413. b) Lovel, I.; Oehme, G.; Lukevics, E. Appl. Organomet. Chem. 1998, 12, 469. 10 Barrett, A. G. M.; Cook, A. S.; Kamimura, A. J. Chem. Soc., Chem. Commun. 1998, 2533. 11 a) Elman, B.; Moberg, C. Tetrahedron 1986, 42, 223. b) Son, S. U.; Jang, H.-Y.; Han, J. W.; Lee, I. S.; Chung, Y. K. Tetrahedron: Asymmetry 1999, 10, 347. c) Habata, Y.; Bradshaw, J. S.; Young, J. J.; Castle, S. L.; Huszthy, P.; Pyo, T.; Lee, M. L.; Izatt, R. M. J. Org. Chem. 1996, 61, 8391. 12 a) Tsukube, H.; Shinoda, S.; Uenishi, J.; Hiraoka, T.; Imakoga, T.; Yonemitsu, O. J. Org. Chem. 1998, 63, 3884. b) Uenishi, J.; Hiraoka, T.; Hata, S.; Nishiwaki, K.; Yonemitsu, O. J. Org. Chem. 1998, 63, 2481. c) Fessner, W.-D.; Sinerius, G.; Schneider, A.; Dreyer, M.; Schulz, G. E.; Badia, J.; Aguilar, J. Angew. Chem. Int. Ed. Engl. 1991, 30. 555. d) Henderson, D. P.; Cotterill, I. C.; Shelton, M. C.; Toone, E. J. J. Org. Chem. 1998, 63, 906-907. e) Takemoto, M.; Achiwa, K. Tetrahedron: Asymmetry 1995, 6, 2925. 13 a) Wei, Z.-L.; Li, Z.-Y.; Lin, G.-Q. Tetrahedron 1998, 54, 13059. b) Uchiyama, M.; Katoh, N.; Mimura, R.; Yokota, N.; Shimogaichi, Y.; Shimazaki, M.; Ohta, A. Tetrahedron: Asymmetry 1997, 8, 3467. 14 Fernholz, E.; Schloeder, D.; Liu, K. K.-C.; Bradshaw, C. W.; Huang, H.; Janda, K.; Lerner, R. A.; Wong, C.-H. J. Org. Chem. 1992, 57, 4756. 15 a) Chelucci, G.; Soccolini, F. Tetrahedron: Asymmetry 1992, 3, 1235. b) Adolfsson, H.; Nordström, K.; Wärnmark, K.; Moberg, C. Tetrahedron: Asymmetry 1996, 7, 1967. c) Chelucci, G. Tetrahedron: Asymmetry 1995, 6, 811. 16 Schmid, C. R.; Bryant, J. D. Org. Synth. 1993, 72, 6. 17 a) Hegedus, L. S. Transition Metals in the Synthesis of Complex Organic Molecules, University Science Books, Mill Valley, 1994. b) Wilkinson, G.; Stone, F. G. A.; Abel, E. W. Comrehensive Organometallic Chemistry, Vol 6, Pergamon Press Ltd., New York, 1982. 18 Tsuji, J.; Takahashi, H.; Morikawa, M. Tetrahedron Lett. 1965, 49, 4387. 19 a) Atkins, K. E.; Walker, W. E.; Manyik, R. M. Tetrahedron Lett. 1970, 43, 3821. b) Hata, G.; Takahashi, H.; Miyake, A. J. Chem. Soc., Chem. Commun. 1970, 1392. 20 Trost, B. M.; Dietsche, T.J. J. Am. Chem. Soc. 1973, 95, 8200. 21 Trost, B. M.; Strege, P. E. J. Am. Chem. Soc. 1977, 99, 1650. 22 a) Trost, B. M.; Van Vranken, D. L. Chem. Rev. 1996, 96, 395. b) Frost, C. G.; Howarth, J.; Williams, J. M. J. Tetrahedron: Asymmetry 1992, 3, 1089. 23 a) Amatore, C.; Jutand, A.; Meyer, G.; Mottier, L. Chem. Eur. J. 1999, 5, 466. b) Steinhagen, H.; Reggelin, M., Helmchen, G. Angew. Chem. Int. Ed. Engl. 1997, 36, 2108. 24 a) Selvakumar K.; Valentini, M.; Pregosin, P. S.; Albinati, A. Organometallics 1999, 18, 4591. b) Blacker, A. J.; Clarke, M. L.; Loft, M. S.; Williams, J. M. J. Org. Lett. 1999, 1, 1969. c) Hilgraf, R.; Pfaltz, A. Synlett 1999, 11, 1814. d) Prétôt, R.; Pfaltz, A. Angew. Chem. Int. Ed. 1998, 37, 323.


- 23 -

25 Kadota, J.; Katsuragi, H.; Fukumoto, Y.; Murai, S. Orgametallics 2000, 19, 979. 26 a) Malkov, A. V.; Baxendale, I. R.; Dvorak, D.; Manfield, D. J.; Kocovsky, P. J. Org. Chem. 1999, 64, 2737. b) Malkov, A. V.; Davis, S. L.; Baxendale, I. R.; Mitchell, W. L.; Kocovsky, P. J. Org. Chem. 1999, 64, 2751. c) Malkov, A. V.; Spoor, P.; Vinader, V.; Kocovsky, P. J. Org. Chem. 1999, 64, 5308. d) Glorius, F.; Pfaltz, A. Org. Lett. 1999, 1, 141. e) Trost, B. M.; Hachiya, I. J. Am. Chem. Soc. 1998, 120, 1104. f) Malkov, A. V.; Davis, S. L.; Mitchell, W. L.; Kocovsky, P. Tetrahedron Lett. 1997, 38, 4899. g) Faller, J. W.; Linebarrier, D. Organometallics 1988, 7, 1670. 27 a) Lloyd-Jones, G. C.; Pfaltz, A. Angew. Chem. Int. Ed. Engl. 1995, 34, 462. b) Lehmann, J.; Lloyd-Jones, G. C. Tetrahedron 1995, 51, 8863. c) Corey, E. J.; Boaz, N. W. Tetrahedron Lett. 1984, 25, 3063. 28 a) Farthing, C. N.; Kocovsky, P. J. Am. Chem. Soc. 1998, 120, 6661. b) Billington, D. C. Chem. Soc. Rev. 1985, 14, 93. 29 Takeuchi, R.; Shiga, N. Org. Lett. 1999, 1, 265. 30 Kadota, J.; Komori, S.; Fukumoto, Y.; Murai, S. J. Org. Chem. 1999, 64, 7523, and references therein. 31 Evans, P. A.; Robinson, J. E.; Nelson, J. D. J. Am. Chem. Soc. 1999, 121, 6761. 32 a) Kondo, T.; Morisaki, Y.; Uenoyama, S.-Y.; Wada, K.; Mitsudo, T.-A. J. Am. Chem. Soc. 1999, 121, 8657. b) Morisaki, Y.; Kondo, T.; Mitsudo, T.-A. Organometallics 1999, 18, 4742. 33 Roustan, J. L.; Merour, J. Y.; Houlihan, F. Tetrahedron Lett. 1979, 3721. 34 Trost, B. M.; Van Vranken, D. L.; Bingel, C. J. Am. Chem. Soc. 1992, 114, 9327. 35 Helmchen, G. J. Organomet. Chem. 1999, 576, 203. 36 Burckhardt, U.; Baumann, M.; Trabesinger, G.; Gramlich, V.; Togni, A. Organometallics 1997, 16, 5252. 37 a) Sprinz, J.; Kiefer, M.; Helmchen, G.; Reggelin, M.; Huttner, G.; Walter, O.; Zsolnai, L. Tetrahedron Lett. 1994, 35, 1523. b) Reiser, O. Angew. Chem. Int. Ed. 1993, 32, 547. 38 Blöchl, P. E.; Togni, A. Organometallics 1996, 15, 4125. 39 Oslob, J. D.; Åkermark, B.; Helquist, P.; Norrby, P.-O. Organometallics 1997, 16, 3015. 40 a) Faller, J. W.; Tully, M. T.; J. Am. Chem. Soc. 1972, 94, 2676. b) Faller, J. W.; Thomsen M. E., Mattina, M. J. J. Am. Chem. Soc. 1971, 93, 2642. 41 Albinati, A.; Kunz, R. W.; Amman, C. J.; Pregosin, P. S. Organometallics 1991, 10, 1800. 42 Hansson, S.; Norrby, P.-O., Sjögren, M. P. T.; Åkermark, B.; Cucciolito, M. E.; Giordano, F.; Vitagliano, A. Organometallics 1993, 12, 4940. 43 Mackenzie, P. B.; Whelan, J.; Bosnich, B. J. Am. Chem. Soc. 1985, 107, 2046. 44 Hayashi, T.; Yamamoto, A.; Hagigara, T. J. Org. Chem. 1986, 51, 723. 45 a) Dierkes, P. Ramdeehul, S.; Barloy, L.; De Cian, A.; Fischer, J.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.; Osborn, J. A. Angew. Chem. Int. Ed. 1998, 37, 3116. b) Koning, B.; Meetsma, A.; Kellogg, R. M. J. Org. Chem. 1998, 63, 5533. c) Saitoh, A.; Misawa, M.; Morimoto, T. Tetrahedron: Asymmetry 1999, 10, 1025. d) Prétôt, R.; Pfaltz, A. Angew. Chem. Int. Ed. 1998, 37, 323. e) Lee, S.-G.; Lim, C. W.; Song, C. E.; Kim, K. M.; Jun, C. H. J. Org. Chem. 1999, 64, 4445. 46 Nordström, K.; Macedo, E.; Moberg, C. J. Org. Chem. 1997, 62, 1604. 47 Original procudure : Trost, B. M.; Murphy, D. J. Organometallics 1985, 4, 1143. Procedure representative of later modifications: Leutenenegger, U.; Umbricht, G.; Fahrni, C.; von Matt, P.; Pfaltz, A. Tetrahedron 1992, 48, 2143. 48 Chelucci, G.; Medici, S.; Saba, A. Tetrahedron: Asymmetry 1997, 8, 3183. 49 Steinhagen, H.; Reggelin, M.; Helmchen, G. Angew. Chem. Int. Ed. Engl. 1997, 36, 2108. 50 Gaussian 98, Revision A.3, Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, Jr., J. A.; Stratmann, R. E.;


- 24 -

Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S. and J. A. Pople, Gaussian, Inc., Pittsburgh PA, 1998. 51 a) Ramdeehul, S.; Dierkes, P.; Aguado, R.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.; Osborn, J. A. Angew. Chem. Int. Ed. Engl. 1998, 37, 3118. b) Lloyd-Jones, G. C.; Stephen, S. C. Chem. Eur. J. 1998, 4, 2539. c) Ward, T. R. Organometallics 1996, 15, 2836. d) Blöchl, P. E.; Togni, A. Organometallics 1996, 15, 4125. e) Albinati, A.; Pregosin, P. S.; Wick, K. Organometallics 1996, 15, 2419. f) Brown, J. M.; Hulmes, D. I.; Guiry, P. J. Tetrahedron 1994, 50, 4493. 52 Hodgson, M.; Parker, D.; Taylor, R. J.; Ferguson, G. J. Chem. Soc., Chem. Commun. 1987, 1309. 53 Albinati, A.; Pregosin, P. S.; Wick, K. Organometallics 1996, 15, 2419. 54 Peña-Cabrera, E.; Norrby, P.-O.; Sjögren, M.; Vitagliano, A.; De Felice, V.; Oslob, J.; Ishii, S.; O’Neill, D.; Åkermark, B.; Helquist, P. J. Am. Chem. Soc. 1996, 118, 4299.

Part II: Microwave Acceleration

- 25 -

PART II Comparison of Ordinary and Microwave Heating Methods of Accelerating Organic Reactions The first examples of organic reactions accelerated by microwave irradiation (µw) were published in 1986 by Gedye et al1 and Giguere et al.2 These and many subsequent publications reported drastically reduced reaction times and sometimes selectivity differences in relation to results obtained under conventional heating.3 Two examples of dramatic reaction time reduction1,2 and one study, in which microwave irradiation led to a different product isomer ratio than ordinary heating,4 are shown in Figure 1.

SO3HSO3HH2SO4 (aq)

+

6:1 at 25 oC1:1.3 at 200 W µw1:15 at 600 W µw

OMe

O+

72 h, rflx* : 90% yield10 min, 600 W µw : 87% yield

O

OMe

O

MeO

MeO

O

dioxane

OH

O

OMe

O

MeOH

H+ cat. 8 h, rflx* : 74% yield5 min, 575 W µw : 76% yield

Figure 1. Early examples of differences in reaction rate1,2 and selectivity,4 owing to different heating methods. (rflx*): Ordinary heating under reflux. (µw): Microwave heating. There are early reports in the literature, presenting reproducibility problems when domestic microwave ovens are used.5 Such equipment does not spread the radiation energy evenly inside the oven and the power settings do not correspond to true power levels. Instead, the maximum power is turned on and off to give the set power as an average. Naturally, this power cycling with intermittent cooling, causes severe reproducibility problems when short reaction times are used. Modifications have been presented,5 constantly using full power and having inlets for cooling to allow reflux under irradiation. However, the problem of radiation


- 26 -

inhomogeneity has only partly been solved by using multiple wave guides (several emitting points) or mode stirrers (rotating blades placed so as to reflect and spread the radiation).6,7 Further development of the microwave technology has led to a single mode cavity, in which the microwave energy is focused on the sample as a standing wave,8 yielding high and reproducible energy densities. Temperature and pressure measurements as well as stirring have been provided. This and later developments have led to equipment significantly more suitable for organic synthesis than the domestic ovens.9 One drawback with the single mode cavity is that the sample size is more or less fixed, which causes problems with scale-up. This can be avoided by using a single mode cavity as a flow reactor that continuously irradiates a hot zone and cools the outlet flow, a device with some applications in industry.10 Theoretical Survey Microwave radiation is usually defined as the frequency range between 300 MHz and 300 GHz.11 Only a few frequencies are open to general use, as cellular phones and other radio applications also use microwave frequencies. The frequency used in domestic ovens and single mode cavities is 2.45 GHz (12.2 cm wavelength) or in some countries 915 MHz (32.8 cm). When microwave radiation passes through a liquid medium consisting of molecules with dipolar moments, the electric field component of the radiation makes the molecules rotate, turning the dipoles in the direction of the electric field. However, neighboring molecules hinder each other in their rotation, setting an upper limit to the rotation frequency. The relaxation time τ is the time needed for the molecules of the liquid to orient themselves randomly after removal of the electric field. 11 The value of τ is dependent on the size and functional groups of the molecules, as well as on temperature and viscosity. A pure liquid compound absorbs most efficiently at microwave frequencies around (2πτ)-1. The molecular rotation can follow the electric field at low frequencies, but as the frequency is increased, the rotation starts to lag behind and finally stops as the field alternates too fast for the molecules to respond. The liquid changes from being an electric condenser to becoming an insulator. In the frequency domain, where the molecular rotation has started to lag behind the electric field, but has not yet stopped, the radiation energy is absorbed as internal friction, or dielectric loss. The internal friction can be described as the kinetic energy transferred when the solvent molecules push each other aside in their attempts to comply with the electric field. In this process, the kinetic energy is converted into molecular translation, rotation and vibration, i. e. heat. A measure of how much the molecular rotation is lagging behind is the loss angle δ, 0o representing rotation in phase with the induced current. Loss angles above 90o have no meaning, as the induced current would go against the electric field. The dielectric loss or the internal friction is represented by ε´´, the loss factor. At


- 27 -

frequencies far above (2πτ)-1, i. e. when the molecular rotation is no longer affected by the electric field, the permittivity of the solvent, ε´ (the dielectric constant), is a function of the electron polarizability alone. The loss tangent, ε´´/ε´ = tan δ is commonly used as a measure of the efficiency, with which a liquid absorbs microwave radiation. The values of τ and tan δ of some common solvents are shown in Table 1.11

Solvent Relaxation time

ττττ (ps)

Loss tangent,

tan δδδδ at 2.45 GHz

Water 9.04 0.123 Methanol 51.5 0.659 Ethanol 170 0.941 1-Propanol 332 0.759 2-Propanol 237 0.799 Acetonitrile 3.4 0.049 Benzonitrile 33.5 0.459 Ethyl acetate 4.41 0.059 Acetic acid 177.4 0.174 Acetone 3.4 0.045 Chloroform 8.94 0.091 Dichloromethane 3.12 0.042 THF 3.49 0.047 DMF 13.05 0.161 DMSO 20.5 0.825 Table 1. Dielectric parameters of some common solvents at 20 oC. A high value of tan δ means that the solvent absorbs microwave radiation efficiently. The closer the relaxation time τ of a solvent is to 65.0 ps (the period at 2.45 GHz), the more efficient is the absorption by the solvent.11

The frequency 2.45 GHz corresponds to a period of 65.0 ps, and the closer the relaxation time τ of the liquid is to 65.0 ps, the higher is the absorption. When the temperature is increased, the molecules can rotate faster, i. e. τ decreases. A solvent that has a τ value lower than 65 ps at room temperature will thus absorb less with increased temperature. Solvents with τ higher than 65 ps, on the other hand, show increased absorption at increased temperature, with thermal runaway as a possible effect when the solvent is irradiated continuously. With most solvents, thermal runaway is avoided as the absorption decreases when τ falls below 65 ps at further increased temperature (see Figure 2).


- 28 -

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

20 70 120 170

Temp (°°°°C)

tan

Figure 2. Tan δ of 1-propanol as a function of temperature and calculated by elaboration on Debye’s equation,11 using estimated values (εs = 19.5 and ε∞ = 6.5 (s4 A2 kg-1 m-3)). Microwave-heated reaction vessels are often closed and build up a significant pressure (like an autoclave) under irradiation. Superheating is also a common phenomenon in open microwave-irradiated reaction vessels. These are the reasons why the temperature axis is extended beyond the boiling point at 1 atm (97 °C). A second mechanism of microwave absorption, ionic conduction, is possible when dissociated ions are present in the solvent. The electric field component of electromagnetic radiation causes dissolved ions to move, just as it causes dipolar molecules to turn. The motion of the ions causes friction and heat. This heating mechanism is not as sensitive to the choice of frequency as the dielectric loss, but seems to be most efficient at frequencies lower than 2.45 GHz and is also strongly dependent on the type of salt present. Even very small quantities of certain kinds of ions (e.g. iodide) can increase the microwave absorption by several orders of magnitude.12 The question whether microwave heating differs from ordinary heating has been under debate almost as long as microwaves have been used in organic synthesis.13 Much higher reaction rates than those reached with ordinary heating have been observed in many early experiments with microwave-heated refluxing reaction mixtures. This has sometimes been attributed to a specific microwave effect. In some instances a microwave “activation” of the reacting units has been postulated.13 This has been difficult to rationalize, however, as especially vibrational transitions require higher energy with most organic molecules than that provided by 2.45 GHz. Rotational transitions of organic molecules is within reach of microwave energies, while the lowest vibrations need at least far-IR irradiation (>3 THz) for excitation.14


- 29 -

Temperature Measurements Under Microwave Irradiation Temperature measurements on irradiated refluxing reaction mixtures showed that the interior of the liquid was superheated by up to 20-30 oC (see Table 2). This corresponded to an about 4-8-fold acceleration of a typical organic reaction, explaining most of the rate increase observed with microwaves. The problem of field inhomogeneity made it even more difficult to compare ordinary heating with microwave heating in the domestic ovens used in the early investigations. Some authors have postulated “hot spots”, in which the reaction rate was so much higher that a large fraction of the reaction took place there.15 An accurate comparison of microwave heating with ordinary heating is possible only if the exact temperature is known, and if the temperature is not homogeneous, then its distribution has to be known. It is difficult to measure temperatures during microwave irradiation, as ordinary thermocouples or thermometers relying on an expanding liquid (Hg, EtOH) often absorb much more radiation energy themselves than the environment, whose temperature they should measure. A few instruments function under irradiation: thermometers relying on expanding non-polar liquids (e.g. p-xylene), the fluoroptic sensor and the IR pyrometer. The p-xylene thermometer is not very accurate; the fluoroptic sensor measures the temperature in one point and the IR pyrometer measures the average temperature of a surface layer of the liquid. The results given by the three thermometers are shown in Table 2.16

Solvent bp (oC) p-xylene thermometer (oC)

fluoroptic sensor (oC)

IR pyrometer (oC)

H2O 100 104 104 105 MeOH 65 78 84 84 2-PrOH 82 87 100 108

THF 66 79 81 103 CH3CN 81 97 107 120

Table 2. Comparison of attempts to use a p-xylene thermometer, a fluoroptic sensor and an IR pyrometer for measuring the temperature of superheated boiling solvents during microwave irradiation.16

As displayed in Table 2, the difference between the values found by the best measurement methods available is at worst 24 oC (THF), which according to the Arrhenius equation corresponds to an uncertainty in reaction rate of a factor 6 (estimated value of Ea = 84 kJ/mol).17 The fact that only local temperatures can be measured by the three methods results in large errors in rate estimations unless the temperature is uniform. As mentioned earlier, the domestic ovens give inhomogeneous temperatures. Even samples irradiated by a single mode cavity have shown significant deviation from homogeneity.18 As long as the complete temperature profile of the reaction solution cannot be determined accurately, a comparison between ordinary and microwave heating will remain difficult.


- 30 -

Advantages of Microwave Heating In spite of the difficulties described, microwave heating has some distinct advantages over ordinary heating. The reaction solution can be heated extremely rapidly, as the heat is transferred directly into the reaction medium, whereas with ordinary heating the energy must be transferred via the vessel wall.* Rapid heating is important when the reaction time is essential, e.g. in combinatorial chemistry and in syntheses utilizing radioactive isotopes with short half-lives. In irreversible reactions, where the thermodynamic product is desired rather than the kinetic one, it is also advantageous to reach high temperatures quickly.4 In addition, less energy is needed to reach a desired temperature, as only the reaction mixture itself has to be heated. Some of the differences in catalyst stability observed have been attributed to the exclusive heat transfer through the vessel walls with ordinary heating. A sensitive catalyst can be expected to deteriorate more quickly at the hot vessel wall than in the uniformly heated interior of the reaction solution in a microwave-heated vessel.* The microwave energy is especially efficiently absorbed at phase boundaries, which is an advantage with heterogeneous reaction mixtures. One such example is the transfer hydrogenation of soybean oil with an aqueous solution of ammonium formate and Pd/C as catalyst.19 There are still safety hazards, however. When a reaction is run in a closed vessel under microwave irradiation without temperature and pressure control, there is a risk of explosion. When instruments for temperature and pressure measurement are not available, the danger of explosion can be avoided by using a septum as a pressure relief device, and placing the microwave oven in an efficiently ventilated hood. Effects of Heating on Asymmetric Catalysis In most investigations on asymmetric catalysis, researchers strive to keep the temperature as low as possible, still maintaining a sufficient reaction rate. An increase in enantioselectivity is often observed on temperature decrease, especially if the initial selectivity is low.20 The kinetics of reactions catalyzed by enzymes has been more deeply investigated than that of reactions catalyzed by metal complexes.21 With sufficient data, thermodynamic constants can be calculated, and from these, something about the origin of selectivity may be learned.22 The following is a brief repetition of the necessary thermodynamic approximations, starting with the Arrhenius equation:23

* When the reaction vessel is heated in an oil bath, all of the heat is transferred through a thin layer next to the glass wall. In this layer only laminar flow is possible and hence, heat is transferred only via conduction, which is much slower than convection, which operates in the rest of the solution. This layer is called the thermal boundary layer and has a considerably higher temperature than the bulk of the solution until equilibrium has been reached. (McCabe, W. L.; Smith, J. C. Unit Operations of Chemical Engineering, 3rd Ed., 1976, McGraw-Hill Inc.)


- 31 -

−

= RTE

obs

a

Aek or

∆−

= RTG

obs eh

kTk#

The activation energy can then be split into entropy and enthalpy:

### STHG ∆−∆=∆ A combination with the Arrhenius equation gives the Eyring equation:

∆−

∆

= RTH

RS

obs eh

kTk##

Relevant for this discussion is the proportion of the rates of formation of the two enantiomers, in enzyme chemistry usually represented by E:24

∆∆−

∆∆−

∆∆

∆−∆−

∆−∆

∆−

∆

∆−

∆

===== RTG

RTH

RS

RTHH

RSS

RTH

RS

RTH

RS

R

S eee

eh

kT

eh

kT

kkE

RSRS

RR

SS

#######

##

##

This equation can be used to assess the importance of the terms contributing to the selectivity. The most selective situation occurs when ∆∆S# and ∆∆H# have opposite signs (see curve a in Figure 3). This occurs when the TS leading to one enantiomer is significantly less ordered and at the same time is lower in enthalpy than the other TS. The entropy-term alone is not sufficient to give satisfactory values of E (see curve d in Figure 3), but when combined with the enthalpy-term it becomes important (compare curves a and c in Figure 3).

0

50

100

150

200

250

300

0 50 100 150 200 250

Temp (°C)

E

Figure 3. The graph shows how E depends on temperature when a) ∆∆H# and ∆∆S# have opposite signs, b) ∆∆S# is negligible, c) ∆∆H# and ∆∆S# have the same sign and d) ∆∆H# is negligible (E=0.3). The activation parameters used to calculate this graph are representative values in enantioselective enzyme catalysis22 (∆∆H#=10 kJ/mol, ∆∆S#=-10 J/(mol K)).

a

b c

d


- 32 -

To my knowledge, these activation parameters have not been determined for any asymmetric catalytic organometallic reaction. In a recent article, Cainelli et al. describe a mechanism change on increased temperature due to solvation effects. 25 Thus it may be dangerous to draw too wide conclusions from thermodynamic data collected in a limited temperature interval. In most studies published on the temperature dependence of selectivity in catalytic reactions, the authors simply state that the selectivity decreases with increased temperature.20 However, a reaction may still be quite selective at elevated temperatures, as will be shown in the following section. Let us consider an asymmetric catalyst, in whose activation energy the ∆∆S#-term is negligible. Figure 4 shows the effects of heating on selectivity in three reactions with selectivities of 99.6, 98.0 and 90.5% ee, respectively, at 0 °C.

70.0

75.0

80.0

85.0

90.0

95.0

100.0

0.0 50.0 100.0 150.0

Temp (°C)

ee (%

)

Figure 4. Plots of ee as a function of temperature in three systems, in which the ∆∆S#-term is assumed to be negligible. The most selective example in Figure 4 still gives an excellent selectivity (96.4% ee) at 150 oC. When the selectivity is only 90.5% ee at 0 °C the temperature dependence is more dramatic, yielding 74.8% ee at 150 oC. This is easier to see when the difference in activation energies is considered. The enantioselectivity is usually represented as ee, but as can be seen in Figure 5, E is more suitable for describing differences in selectivity when ee>99%.26

eeeeE

−+

=11

; EEee

+−

=11


- 33 -

Figure 5. E and ee as functions of the difference in activation energy ∆∆G#, displayed at three temperatures (a=20 °C, b=100 °C and c=180 °C). The room temperature selectivity when ∆∆G#=15 kJ/mol is 99.6% ee as in Figure 4. If there is a difference in activation volume (∆∆V#≠0), the selectivity should change when the pressure inside the closed reaction vessel is increased on heating. To my knowledge, this parameter has not yet been determined for any asymmetric catalyzed reaction. There are a few examples of changed regioselectivity with increased pressure, however.27 As the equipment used cannot measure the pressure, the following discussion of selectivities will neglect any influence of pressure. Experimental results of asymmetric catalytic reactions under heating, both by convential means and microwaves, will be discussed in the following section. Palladium-Catalyzed Allylic Substitution under Microwave Irradiation Hallberg’s research group has published good results of microwave-accelerated couplings of the Stille, Suzuki and Heck types.28 We started a collaboration where we brought our asymmetric catalytic reactions and Hallberg et al. brought their expertise on Pd-catalyzed reactions under microwave irradiation. Initial promising results29 led us to a deeper investigation, in which several types of ligands, substrates and nucleophiles were tested.30,31 The irradiation times and effects shown below are partially optimized. The results displayed in the tables below were obtained with three goals in mind. The first one was to reach full conversion at different effect levels in order to see variations in enantioselectivity and yield. The second goal was to test the stability of the systems by determining the maximal power before the catalytic systems broke down. The third

0.0 50.0

100.0 150.0 200.0 250.0 300.0 350.0 400.0 450.0 500.0

0.0 5.0 10.0 15.0 ∆ ∆∆ ∆∆ ∆∆ ∆ G# (kJ/mol)

E

0.00 96.08 98.02 98.68 99.00 99.20 99.34 99.43 99.50 99.56 99.60

ee (%)

c

b

a


- 34 -

goal was to minimize the reaction time at the maximal effect to determine the conditions of the fastest reaction possible. Equipment The single-mode cavity MW10 from Personal Chemistry (former Labwell), Uppsala, Sweden, was used in this investigation. The MW10 is a single-mode microwave cavity that focuses 0-500 W of 2.45 GHz microwave radiation into a sample of up to 5 mL. The reaction vessels were Duran tubes (microwave transparent) with screw caps provided with septa as pressure relief devices. At the time of the last Mo-catalysis investigation the oven had been modified to provide magnetic stirring. Alkylations of 1,3-Diphenyl-2-propenyl Acetate with Dimethyl Malonate Using N,N- and N,P-LigandsIV,V The Pd-catalyzed allylic alkylation of 1,3-diphenyl-2-propenyl acetate with dimethyl malonate is an example of a “benchmark” reaction (see Introduction), commonly used to test ligand performance. The three N,N-ligands, pyridino- and quinolinooxazolines (1-3) represent a range of selectivities from 73 to >99% ee.

NO

NPh

NOMeO

NPh

NO

NPh

OH

1 2 3 An earlier study showed ligand 1 to be easily synthesized and moderately selective (73% ee) at room temperature.29 Ligands 2 and 3 were shown to be highly selective at room temperature (95 and >99% ee, respectively),32 but gave low yields and decreased enantioselectivities even under mild microwave irradiation (Table 3).

Ph Ph

OAc

Ph Ph

OMe

O

MeO

O

[(η3-C3H5)PdCl]2, ligand 1(MeOCO)2CH2, BSA, KOAc

4 5

Ligand Solvent Temp (oC) (oil bath)

Time Yield of 5 (%) ee (%)

1 CH3CN 23 3 days 99 77 1 CH3CN 100 19 min 97 62 1 CH3CN 140 6.3 min 93 60 1 CH3CN 180 4.5 min 93 56 2 CH2Cl2 23 4 days 81 95 3 CH2Cl2 23 4 days 97 >99


- 35 -

Ligand Solvent Effect (W) Time (min) Yield of 5 (%) ee (%)

1 CH3CN 35 15.0 99 65 1 CH3CN 70 7.5 99 64 1 CH3CN 120 3.5 99 63 1 CH3CN 250 3.0 99 65 1 CH3CN 500 2.0 99 65 2 CH2Cl2 35 15.0 <5 86 3 CH2Cl2 25 15.0 <5 89

Table 3. Yields and enantioselectivities of 5 obtained by using ligands 1-3.29,31 These results showed us that microwave-accelerated Pd-catalyzed allylic substitution was indeed feasible, although some ligands were not suitable. In the reaction with 1, the drop in enantioselectivity (from 77% ee at room temperature to 65% ee at 500 W) and the rate increase (over 2000-fold) were consistent with an average reaction temperature of around 140-150 oC (assuming that the ∆∆S#-term could be neglected). A ligand class giving almost total enantioselectivity at room temperature is the phosphinooxazolines (6a, 6b and 6c) developed by von Matt and Pfaltz,33 Sprinz and Helmchen34,35 and Dawson et al.36

PPh2 N

O

R'R

6a: R=H, R'=tBu6b: R=H, R'=iPr6c: R=Ph, R'=H

The original conditions36 had to be modified to give maximum rate increase, exchanging the low-boiling dichloromethane for acetonitrile. Exposure to microwave conditions gave most gratifying results; we obtained retained selectivities and yields even with high radiation effects, as shown in Table 4.

Ligand Temp (oC) Time Yield of 5 (%) ee (%) 6a 29 (rt) 6 h (1 h) 56 (94) >99 (95) 6b (rt) (1 h) (98) (98) 6c (rt) (1 h) (99) (99)


- 36 -

Ligand Effect (W) Time (min) Yield of 5 (%) ee (%) 6a 5 5.0 min 15 >99 6a 10 5.0 min 99 >99 6a 30 2.0 min 99 >99 6a 90 40 s 65 >99 6a 90 1.0 min >99 >99 6a 120 20 s 66 >99 6a 120 30 s 97 >99 6a 500 15 s 98 93 6b 90 1.0 min >99 >99 6b 500 15 s 85 97 6c 90 1.0 min 95 97

Table 4. Yields and enantioselectivities of 5 obtained by using phosphinooxazolines 6a-c as ligands under ordinary conditions and microwave heating.30 Values within brackets are values from a publication by von Matt and Pfaltz.33

20

40

60

80

100

120

140

0 30 60 90 120 150 180 210 240 270 300 330Time (s)

Tem

pera

ture

(deg

rees

Cel

cius

)

500 W

120W90 W

30 W

10 W

5 W

Figure 6. Temperature measured by means of a fluoroptic sensor at the bottom of the reaction vessel† during irradiation. The reactions are some of those reported in Table 4. The fully retained selectivity (>99% ee) in the reactions using ligands 6a and 6b with quantitative yields in 1 min reaction times at 90 W may appear strange in the light of the selectivity decrease observed with ligands 1-3. However, this could be explained

† The purpose of placing the fluoroptic sensor at the bottom of the reaction vessel was to increase the reproducibility of the temperature measurements. However, later investigations revealed a significant temperature inhomogeneity, in that a layer about 1 cm above the bottom could be up to 100 oC hotter than the bottom layer when effects >120 W were employed.18 The temperatures in this graph should, therefore, be seen as lower limits.


- 37 -

by an exceptionally great ∆∆G# (see Figures 4 and 5) and the accompanying low decrease in selectivity on heating. It is difficult to recognize a very high ∆∆G# with conventional methods22at room temperature; e.g. an increase of ∆∆G# of 4 kJ/mol corresponds to the minute increase in ee from 99.5 to 99.9%. Pd-Catalyzed Allylic Alkylation Using P,P-LigandsV,VI For our early P,P-ligand investigations we chose the well-known ligand R-BINAP (7), that had been used extensively in many studies of asymmetric catalysis. 37 A more thorough study was then performed with a ligand that had been developed by Trost et al. specifically for Pd-catalyzed allylic alkylation,38 (1R, 2R)-bis-[2’-(diphenylphosphino)benzamido]-1,2-cyclohexane (8).

PPh2PPh2

HNNHOO

7 8

PPh2 PPh2

The combination of dimethyl malonate and 1,3-diphenyl-2-propenyl acetate was the only one tested with BINAP (7), as enantioselectivities of at most 90% were expected39 (see Table 5). The Trost ligand (8), however, had been shown to transfer its chirality most efficiently to a number of allylic substrates and nucleophiles.40 We attempted to use some highly selective published procedures in the microwave oven (see Tables 4, 5 and 6). In some instances the original conditions could be used, but the solvent often had to be changed and sometimes also the source of the nucleophile.

Ph Ph

OAc

Ph Ph

OMe

O

MeO

O

[(η3-C3H5)PdCl]2, ligand 7(MeOCO)2CH2, BSA, KOAc

4 5

Effect (W) Temp (oC) Time Yield (%) ee (%) - 25 1.0 h 97 87

20 - 2.0 min 94 83 20 - 1.5 min 96 85 40 - 1.0 min 95 83

Table 5. Pd-catalyzed allylic alkylation of 1,3-diphenyl-2-propenyl acetate with dimethyl malonate.29

Previously published experiments39 had been made under the same conditions, and yielded 85% of 5 with 90% ee in 4 days at room temperature. Modifications of that process were not necessary, as full conversion and almost quantitative yields and good


- 38 -

selectivities were reached in the microwave oven. The time reduction on heating was moderate (up to 60 times) as the ligand catalyzed the reaction rapidly at room temperature. When we attempted to increase the effect beyond 40 W, the yields decreased, presumably due to decomposition of the catalyst. The relatively small drop in enantioselectivity from 87% ee at room temperature to 83% ee with the highest possible effect (see Table 5) could be expected because of the relatively high selectivity and the relatively small increase in rate.

OCO2Et+

Pd2dba3.CHCl3, ligand 8CH2Cl2

O

OMe

HO

OMe

Effect (W) Temp (oC) Time Yield (%) ee (%) - 100 5.0 min 91 94

35 - 5.0 min 91 95 70 - 40 s 43 96 70 - 2.0 min 92 95 120 - 1.0 min 96 95 150 - 1.0 min 91 94 300 - 45 s 88 92 500 - 30 s 79 92

Table 6. Pd-catalyzed allylic alkylation of 4-methoxyphenol with ethyl 3-cyclohexenyl carbonate.31

0

20

40

60

80

100

120

140

0 100 200 300 400 500

Tim e (s)

Tem

p (d

eg C

)

500 W300 W

150 W 70 W35 W

Figure 7. Temperature measured at the bottom of the reaction vessel using a fluoroptic sensor during irradiation for some of the reactions in Table 6 (see also note † in Figure 6).‡ Previously published results had been obtained with the same Pd source and solvent, yielding 88% product with 97% ee after 4 hours at room temperature.41 Using microwave conditions, however, we could omit inert conditions. All reactions in Table

‡ After irradiation, the reaction vessel was allowed to cool inside the microwave cavity for 60 s to equilibrate thermally before being cooled in a water-bath. When the temperature had reached its highest point, it slowly decreased and then dropped quickly as the reaction vessel was cooled in a water-bath.


- 39 -

6 were run under air; experiments run under nitrogen showed the same yields and selectivities. All examples with heated conditions showed slightly lower selectivities than the room temperature experiment. An effect of 500 W for 30 s gave 79% yield with 92% ee. The use of 120 W for 1.0 min, however, resulted in 96% yield with 95% ee, i. e. only marginally lower selectivity than under the conditions previously published. The experiment with oil bath heating corresponded closely to the 35 W irradiation experiment in reaction time, yield as well as a slightly decreased selectivity.

OCO2Et+

[(η3-C3H5)PdCl]2, ligand 8CH3CN

HN

O

O

N

O

O

Effect (W) Temp (oC) Time Yield (%) ee (%)

- 20 24 h 18 96 - 140 10 min 51 95

35 - 15 min 40 98 70 - 2.0 min 57 96 70 - 3.0 min 87 95 70 - 4.0 min 79 (80) 97 100 - 1.0 min 64 98 100 - 1.5 min 87 96 200 - 1.0 min 75 96 500 - 30 s 68 94

Table 7. Pd-catalyzed allylic alkylation of phthalimide with ethyl 3-cyclohexenyl carbonate.31

0

20

40

60

80

100

120

140

160

180

0 200 400 600 800 1000 1200

Tim e (s)

Tem

p (d

eg C

)

500 W200 W

100 W

70 W35 W

Figure 8. Temperature measured at the bottom of the reaction vessel using a fluoroptic sensor during irradiation. The reactions are those described in Table 7 (see notes † and ‡ from Figures 6 and 7). A procedure previously published by Trost and Bunt,42 yielded 81% product with 98% ee after 18 h at room temperature. In our hands, 99% yield (GC yield) with 99% ee after 130 min at room temperature was achieved. These results are not directly


- 40 -

comparable with those of our investigation, however, as Trost’s method used an excess of tetra-N-hexylammonium bromide to solubilize the anionic nucleophile potassium phthalimide in dichloromethane. It had been shown that the counter-ion was essential for the selectivity in the Pd-catalyzed allylic substitution,43 but unfortunately the high ion concentration created problems under microwave irradiation. Heating by ionic conduction within the reaction mixture became too efficient because of the high ion strength. We attempted to decrease the amount of salt to 0.1 equiv, whereby overheating was avoided. This gave a highly selective reaction (99% ee) at room temperature in toluene or (trifluoromethyl)benzene, but as we attempted to increase the rate using microwaves, the selectivity dropped to 35% ee. Conditions with 18-crown-6 in catalytic amounts in acetonitrile as well as the polar aprotic solvents DMF and DMSO were also attempted, but only starting material was observed after several days at room temperature. The reaction conditions with 4-methoxyphenol (Table 6) gave us the idea not to use an external base. The leaving group, a carbonate, have been shown to decompose to CO2 and alkoxide, which may act as a base during the reaction.44 The use of phthalimide itself as a nucleophile in acetonitrile worked well under heating, but as can be seen in Table 7, the room temperature experiment failed to reach full conversion in 24 hours. This was probably due to the relatively low solubility of phthalimide in acetonitrile at room temperature. The low conversion after an extended irradiation at 35 W might be explained by the assumption that the reaction mixture had not reached a high enough temperature to dissolve sufficient amounts of the nucleophile. Just as when 4-methoxyphenol was used as a nucleophile, inert atmosphere could be omitted without loss in selectivity or yield in the experiments using microwave irradiation. An oil bath experiment at 140 oC for 10 min gave a selectivity similar to that observed in the experiments with microwave heating, but the yield dropped to 51% with full conversion.

OCO2Et+

[(η3-C3H5)PdCl]2, ligand 8

BSA, CH3CNMeO OMe

OOMeO O

O

OMe

Effect (W) Time (min) Yield (%) ee (%) 50 6.0 83 94 100 3.0 69 ≥95 100 4.0 83 ≥95 200 3.0 65 ≥95 500 1.5 83 ≥95

Table 8. Pd-catalyzed allylic alkylation of dimethyl malonate with ethyl 3-cyclohexenyl carbonate.31


- 41 -

0

50

100

150

200

250

0 100 200 300 400 500 600

Tim e (s)

Tem

p (d

eg C

)500 W

200 W100 W

50 W

Figure 9. Temperature during irradiation, as measured at the bottom of the reaction vessel using a fluoroptic sensor. The measurements were made during some of the reactions presented in Table 8 (see notes † and ‡ from Figures 6 and 7). In a report published by Trost and Bunt,42 dichloromethane was employed as a solvent in combination with an excess of tetra-N-hexylammonium bromide, used to solubilize the sodium salt of the nucleophile. The problems encountered earlier − high ion concentration and accompanying too rapid heating − forced us to find an alternative strategy. An attempt to omit the external base yielded starting material only. The base BSA proved not to be strong enough, and an extremely slow reaction resulted. When DBU was used as base, however, the reaction proceeded smoothly. As in the other reactions using ligand 6, similar results were obtained under air and under inert conditions. We also attempted to use a sulfur nucleophile, sodium benzenesulfinate, which according to a procedure published by Trost et al. yielded 87% product with 98% ee.45 This procedure used an excess of ammonium salt, which resulted in problems similar to those occurring in the reaction with phthalimide as nucleophile. A decrease of the amount of salt to 0.1 equiv gave a racemic product. Methanol as solvent made it possible to perform the reaction in the microwave oven, but the selectivity was low. In analogy with the phthalimide procedure, the benzenesulfinic acid was tested as a non-ionic nucleophile source. When one equiv of trifluoroacetic acid was added to the sodium salt to form the sulfinic acid, the selectivity increased to 70% ee. Further improvement was obtained with acetonitrile as solvent, yielding 73% product with 88% ee, but this reaction was unreliable and sometimes gave no product. This was thought to be caused by thermal instability of the sulfinic acid. Conclusion and Discussion The results presented in the previous sections show that microwave acceleration is compatible with asymmetric catalysis. It is interesting that some sensitive reactions,


- 42 -

which are traditionally run under inert atmosphere, can be performed without such precautions in the microwave oven. We have no direct explanation of this, but the wall effect discussed earlier might be the responsible factor. When the hottest part of the reaction mixture is the interior and hence most of the reaction occurs there, catalyst degradation on the wall surface or reactions with oxygen from the gas above the solution might be suppressed. We also have indications that microwaves can increase the turnover number of the catalyst. An example is the reaction with the phosphinooxazolines, where full conversion sometimes was difficult to achieve at room temperature, whereas the same amount of catalyst readily consumed all of the starting material in the microwave oven. A deeper investigation is needed for a full explanation of this observation. The problems encountered with samples with high ion concentration could appear troublesome. With the equipment available, the samples were irradiated continuously. With low ion strength the absorption decreased with increasing temperature, but with high ion strength this did not seem to happen. Therefore, it was difficult to find a power/time combination that would give full conversion, and at the same time avoid explosion. As this is written, new equipment with stirring as well as temperature and pressure control is to be installed in our lab. The stirring will eliminate the problem of temperature inhomogeneity. We will also be able to set a specific temperature that will be reached in a short time, irrespective of the absorption of the reaction mixture. There are still too few data to identify the factors that make a ligand suitable for high-temperature asymmetric catalysis, but a few observations might be extrapolated. Reactions with N- and P-donor ligands have been shown to function well under microwave conditions, with the exception of the reactions with the highly selective N,N-ligands. A possible reason for this is that the Pd-N bond is weaker than the Pd-P one. A decrease in selectivity close to what could be expected with a negligible ∆∆S# (see the section “Effects of Heating on Asymmetric Catalysis”) was observed in all of the reactions except those using phosphinooxazolines. As mentioned earlier this might be explained by an exceptionally high ∆∆G#. Accurate kinetic measurements on these systems would be desirable, as they might give a deeper understanding of the factors determining the selectivity. Mo-Catalyzed Allylic Substitution under Microwave Irradiation Molybdenum Catalyzed Allylic Substitution Allylic substitution catalyzed by molybdenum is similar to the catalysis with palladium as regards the types of substrates that can be used. One general difference between the two types of catalysts is the regioselectivity (see Figure 10).


- 43 -

OMe

O

MeO

O

OCO2MeOMe

O

MeO

O

"Pd-catalyst" "Mo-catalyst"

Major isomer Major isomer Figure 10. Regioselectivity displayed by most Pd- and Mo-catalysts in allylic substitution. Molybdenum has been less studied, however, probably because of the lower reactivity. Catalyst loadings of 10-15% molybdenum and long reaction times at reflux are usually needed to obtain full conversion,46 in contrast to palladium catalysis, in which 1-2% is commonly used to complete the reaction in a few hours at room temperature. The exact role of the catalyst is still subject to debate. It seems reasonable to assume that molybdenum-allyl complexes are intermediates, as they have been characterized by e.g. crystallography,47 and as isolated molybdenum allyl complexes react with nucleophiles to give the same products as under catalytic conditions.48 Recent articles using various Mo(II)49- and Mo(IV)50-species as catalyst precursors, suggest that the metal acts partly as a Lewis acid by coordinating to the leaving group before oxidative addition. The Lewis acid character of the catalyst is further corroborated by the SN1-type reactivity of substrates, e.g. substrates with the leaving group at a tertiary carbon react faster than those with a primary leaving group.51 Lewis basic ether solvents have also been shown to slow the reaction down.51, 52

This pre-coordination of the metal to the leaving group results in retention of configuration at ionization, but as the nucleophilic attack on the molybdenum allyl complex also proceeds with retention,53 the net stereochemistry is the same as in the palladium catalysis, in which the same steps give inversion-inversion. Various nucleophiles and the use of BSA have been shown to alter the stereochemistry of the nucleophilic attack, however.46 Some of the isomerizations occurring in palladium allyl chemistry have also been observed with molybdenum, although the apparent allyl rotation seems to be less facile.54 Somewhat surprisingly, molybdenum does not move to adjacent double bonds in conjugated allyl systems.55 Asymmetric Allylic Substitution Catalyzed by Molybdenum Faller et al. and others have studied the asymmetric nucleophilic addition to molybdenum allyl complexes in a stoichiometric fashion since 1983,56 and a catalytic procedure has recently been developed by Trost and Hachiya.57 Impressive regio- and enantioselectivity is achieved with the system displayed in Figure 11.


- 44 -

HNNHOO

NNOCO2Me OMe

OO

MeOOMe

O

MeO

O

+

Mo(CO)6 (CH3CH2CN)3

49 : 1

9

10a

11a, 99% ee 11b

THF, rt

Preformedcatalyst

+

Figure 11. Catalysis results reported by Trost and Hachiya.57

The linear substrate 9 and branched isomer rac-methyl 1-phenyl-2-propenyl carbonate react to give slightly different selectivities. When the racemic branched allylic substrate was used, a 32:1 regioselectivity and 97% ee of the product was obtained, i.e. somewhat lower selectivities than for the linear substrate 9. Heteroaromatic substrates were also shown to yield good to excellent regio- and enantioselectivities with substrate 9. It is noteworthy that a high regioselectivity was not necessarily accompanied by a high enantioselectivity, and vice versa. Glorius and Pfaltz reported similar results with bisoxazolineamide ligands 12 and 13 (Figure 12).58 Their system showed a more pronounced difference in selectivity between the linear and the branched allylic substrate than that of Trost and Hachiya.57 With 13, the linear substrate 9 yielded 98% ee, while the branched isomer gave 84% ee with unchanged regioselectivity (6:1).

NH HNOO

ON N

ONH HN

OO

O N N O

Catalysis results: 99% ee of (R)-product11a:11b 14:1

98% ee of (R)-product11a:11b 8:1

12 13

Figure 12. Catalysis results reported by Glorius and Pfaltz.58 Microwave-Accelerated Molybdenum-Catalyzed Allylic SubstitutionsVII The impressive results described above made us interested in investigating whether the molybdenum catalyst systems would be stable towards microwave heating. We could soon demonstrate that they were, and also that Mo(CO)6 could be used instead of the unstable pre-catalyst Mo(CO)3(CH3CH2CN)3, when the reaction mixture was heated


- 45 -

with microwaves. A two-step procedure analogous to the one of Trost and Hachiya57 was developed, yielding almost unchanged enantioselectivity (98% ee) but somewhat decreased regioselectivity (11:1), as displayed in Figure 13. The procedure was unreliable, however, full conversion being reached in only one out of three attempts.

OMe

O

MeO

O

11a, 98% eeup to 85% yield11a:11b 11:1

Mo(CO)6NaHC(CO2Me)2, 9

"Mo-catalyst"CH3CN

120 W µw, 5 minTHF

250 W µw, 1 min

Ligand 10a

Figure 13. Two-step procedure with microwave heating. In our search for reproducible conditions, we found that the use of BSA as a base in combination with a catalytic amount of NaH resulted in a cleaner and more reliable reaction. We then managed to simplify the experimental procedure by omitting the catalyst formation step, mixing all components and applying microwave heating. It was also observed that inert atmosphere was unnecessary: air or nitrogen gave the same results, while a pure oxygen atmosphere decreased the yields considerably. In addition, it was possible to decrease the amount of catalyst from the 10% used by Trost and Hachiya57 to 4%. Lower amounts resulted in incomplete conversion. The results are displayed in Table 9.

OMe

O

MeO

OMo(CO)6, Ligand 10a

NaHC(CO2Me)2 : H2C(CO2Me)2 (1:10) BSA

THF, µw, air

OCO2Me

11a

(+ non-chiral isomer 11b)

Effect (W) Time (min) Yield (%) Regioselectivity (11a:11b)

ee (%)

60 9.0 86 11:1 98 90 6.0 86 11:1 98 120 5.0 86 12:1 98 250 5.0 87 12:1 98 500 4.0 85 12:1 95

Oil bath temp

(°°°°C) Time Yield (%) Regioselectivity

(11a:11b) ee (%)

22 20 days <1 n.d. n.d. 80 2 days 11 11:1 n.d. 165 6.0 min 59 7:1 98 180 6.0 min 70 10:1 98

Table 9. One-step molybdenum-catalyzed allylic substitution under air with microwave or oil bath heating.


- 46 -

20

80

140

200

0 2 4 6 8 10Time (min)

Temperature (oC)

250 W x 5 min

120 W x 5 min90 W x 6 min

165 oC x 6 min[a]

60 W x 9 min

90 W x 6 min (Pure THF)

180 oC x 6 min[a]

Figure 14. Temperatures measured during the reactions in Table 9 with a fluoroptic sensor, placed at the bottom of the reaction vessel (see notes † and ‡ in Figures 6 and 7). [a]Oil bath heated examples. Experiments with oil bath heating were also performed for comparison with the use of microwave heating. It is noteworthy that the fluoroptic sensor showed a higher temperature in every point in the experiment with an oil bath at 180 °C than in the experiment with 120 W microwave irradiation (see Figure 14). Even though the conventionally heated mixture was allowed to react for a longer time, the microwave-heated one gave a higher yield (Table 9). Steric and Electronic Variations of the Bispyridylamide ligandsVIII As our group had earlier experience of bispyridylamides,59 the idea to try derivatives in the molybdenum catalyzed allylic substitution was close at hand. We synthesized 10c via the acid chloride and 10b, 10d and 10e using the Mukaiyama reagent.60

HNNHOO

NNR

R' R'

R

10a R=R'=H10b R=H, R'=Me10c R=tBu, R'=H10d R=NO2, R'=H10e R=OMe, R'=H


- 47 -

We then used these ligands in the one-step molybdenum catalysis previously developed. The results are displayed in Table 10. Entry Ligand Effect (W) Time (min) Yield (%)

11a/11b ee (%)

1 10a 200 6.0 82 11:1 98 2 10a 200 5.0 71 11:1 98 3 10b 200 5.0 25 7:1 79 4 10c 200 5.0 40 7:1 98 5 10d 200 5.0 7 9:1 97 6 10d 200 8.0 27 9:1 97 7 10d 150 15.0 32 9:1 97 8 10e 200 4.0 >95 23:1 >99 9 10e 200 5.0 >95 23:1 >99

Table 10. Results from the molybdenum-catalyzed allylic substitution of methyl 3-phenyl-2-propenyl carbonate.

020406080

100120140160180200

0 200 400 600 800 1000Time (s)

Entry 7Entry 6

Entry 1

Entry 8

Entry 4

Entry 3

Figure 15. Temperatures measured with an IR pyrometer during the reactions presented in Table 10. The IR pyrometer had been calibrated against a fluoroptic sensor, making the temperature curves comparable with those in Figure 14. The temperature curves of entries 2, 5 and 9 are omitted for clarity, as they are quite similar to entry 3 and 4. All ligands except 10b retained the excellent enantioselectivity of the parent ligand 10a. Added steric hindrance, i.e. 6-methyl (10b) or 4-tbu (10c), gave overall decreased performance, evident in lower yield and regioselectivity. A tentative explanation of this could be suggested by postulating a monomeric molybdenum catalytic complex with the ligand essentially in one plane. The 6-methyls of 10b would then stericly collide and disturb the complexation. The role of the 4-tbu in 10c was less clear, but seemed to be more related to a steric than an electronic effect.

Attaching an electron-withdrawing substituent (10d) in the 4–position results in a very slow catalytic system having slightly lower regio- and enantioselectivity than 10a. However, with the electron-donating methoxy group in the 4-position (10e), the molybdenum ligand complex becomes more efficient than the parent system, attaining distinctly higher yield, regio- and enantioselectivity in a shorter reaction time. These results suggest that molybdenum does not primarily act as a Lewis acid in the rate-determining step, since the electron-withdrawing nitro substituent should then


- 48 -

have resulted in a faster and the electron-donating methoxy group a slower catalyst. This is contrary to the observed reactivity. The fact that the reactivity increase is followed by an increase in enantio- and regioselecitvity is noteworthy, but it is difficult to explain this finding at the present stage of the investigation. 1 Gedye, R.; Smith, F.; Westaway, K.; Ali, H.; Baldisera, L.; Laberge, L.; Rousell, J. Tetrahedron Lett. 1986, 27, 279. 2 Giguere, R. J.; Bray, T. L.; Duncan, S. M.; Majetich, G. Tetrahedron Lett. 1986, 27, 4945. 3 Synthetically oriented reviews : a) Langa, F.; de la Cruz, P.; de la Hoz, A.; Diaz-Ortiz, A.; Diez-Barra, E. Contemp. Org. Synth. 1997, 373. b) Caddick, S. Tetrahedron, 1995, 51, 10403. c) A. Loupy, A. Petit, J. Hamelin, F. Texier-Boullet, P. Jacquault, D. Mathé, Synthesis, 1998, 1213. 4 Stuerga, D.; Gonon, K.; Lallemant, M. Tetrahedron, 1993, 49, 6229. 5 Problems with domestic ovens and their possible modifications are covered in reference 3b. 6 Strauss, D. R.; Trainor, R. W. Aust. J. Chem. 1995, 48, 1665. 7 Mingos, D. M. P.; Baghurst, D. R. Chem. Soc. Rev. 1991, 20, 1. 8 This development is discussed in reference 7 and in: Strauss, C. R.; Trainor, R. W. Aust. J. Chem. 1995, 48, 1665. 9 a) Stone-Elander, S.; Elander, N.; Thorell, J.-O.; Solås, G.; Svennebrink, J. J. Label. Cmpds. Radiopharm., 1994, 10, 949. b) Raner, K. D.; Strauss, C. R.; Trainor, R. W.; Thorn, J. S. J. Org. Chem. 1995, 60, 2. 10 C. R. Strauss Aust. J. Chem 1999, 52, 83. 11 Gabriel, C.; Gabriel, S.; Grant, E. H.; Halstead, B. S. J.; Mingos, D. M. P. Chem. Soc. Rev. 1998, 27, 213. 12 Zijlstra, S.; de Groot, T. J.; Kok, L. P.; Visser, G. M.; Vaalburg, W. J. Org. Chem. 1993, 58, 1643. 13 The history of microwave effects and the theroretical background is discussed in: Stuerga, D. A. C.; Gaillard, P. J. Microwave Power and Electromagn. Energy 1996, 31, 87 and in reference 3a. 14 Willlard, H. H.; Merrit, L. L., Jr.; Dean, J. A.; Settle, F. A., Jr. Instrumental Methods of Analysis, Wadsworth Publishing, USA, 1988. 15 For a discussion on “Hot spots” see reference 3a and: Stuerga, D. A. C.; Gaillard, P. J. Microwave Power and Electromagn. Energy 1996, 31, 87 16 Whittaker, A. G.; Mingos, D. M. P. J. Microwave Power Electromagn. Energy 1994, 29, 195. 17 A typical activation energy of organic reactions, according to Moore, W. J. Basic Physical Chemistry, Prentice-Hall International, 1983, Chapter 15. 18 Reaction mixtures heated by the MW10 at effects above 120W have been shown to form a hot spot about 1.5 cm above the bottom of the vessel. Kaiser, N.-F., Unpublished results 19 Leskovsek, S.; Smidovnik, A.; Koloini, T. J. Org. Chem. 1994, 59, 7433. 20 For examples of allylic substitutions at different temperatures, see: a) Evans, P. A.; Brandt, T. A. Tetrahedron Lett. 1996, 51, 9143. b) Sawamura, M.; Ito, Y. Chem. Rev. 1992, 92, 857. c) Hayashi, T.; Kanehira, K.; Hagihara, T. Kumada, M. J. Org. Chem. 1988, 53, 113. 21 Wong, C.-H,; Whitesides, G. M. Enzymes in Synthetic Organic Chemistry, Elsevier Science Ltd., Oxford, 1995. 22 Overbeeke, P. L. A.; Ottosson, J.; Hult, K; Jongejan, J. A.; Duine, J. A. Biocatal. Biotrans. 1999, 17, 61. Overbeeke, P. L. A. Ph.D. Thesis, Technical University of Delft, The Netherlands, 1999.


- 49 -

23 Lowry, T. H.; Schueller Richardson, K. Mechanism and Theory in Organic Chemistry, 3rd Ed., HarperCollinsPublishers, New York, 1987. 24 Chen, C.-S,; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am. Chem. Soc. 1982, 104, 7294. 25 Cianelli, G.; Galletti, P.; Giacomini, D.; Orioli, P. Angew. Chem. Int. Ed.. Engl. 2000, 39, 523. 26 Chen, C.-S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am. Chem. Soc. 1982, 104, 7294. 27 a) Kunugi, S.; Kawade, T.; Kabata, H.; Nomura, A.; Komiyama, M. J. Chem. Soc., Perkin Trans. 2 1991, 5, 747. b) Makimoto, S.; Keizo, S.; Taniguchi, Y. J. Phys. Chem. 1982, 86, 4544. c) Gerard, J.; Papadopoulos, M. Tetrahedron Lett. 1996, 37, 1417. 28 a) Larhed, M.; Hoshino, M.; Hadida, S.; Curran, D. P.; Hallberg, A. J. Org. Chem. 1997, 62, 5583. b) Larhed, M.; Lindeberg, G.; Hallberg, A. Tetrahedron Lett. 1996, 37, 8219. c) Larhed, M.; Hallberg, A. J. Org. Chem. 1996, 61, 9582. 29 Bremberg, U.; Larhed, M.; Moberg, C.; Hallberg, A. J. Org. Chem. 1999, 64, 1082. 30 Kaiser, N.-F. K.; Bremberg, U.; Larhed, M.; Moberg, C.; Hallberg, A. J. Organomet. Chem. 2000, in print. 31 Bremberg, U.; Lutsenko, S.; Kaiser, N-F. K.; Larhed, M.; Hallberg, A.; Moberg, C. Synthesis 2000, in print. 32 Nordström, K.; Macedo, E.; Moberg, C. J. Org. Chem. 1997, 62, 1604. 33 von Matt, P.; Pfaltz, A. Angew. Chem. Int. Ed. Engl. 1993, 32, 566. 34 Sprinz, J.; Helmchen, G. Tetrahedron Lett. 1993, 34, 1769. 35 Helmchen, G. J. Organomet. Chem. 1999, 576, 203. 36 Dawson, G. J.; Frost, C. G.; Williams, J. M. J. Tetrahedron Lett. 1993, 34, 3149. 37 a) Sho, S. Y.; Shibasaki, M. Tetrahedron Lett., 1998, 39, 1773. b) Miyashita, A., Takaya, H.; Souchi, T.; Noyori, R. Tetrahedron 1984, 40, 1245. c) Miyashita, A.; Yasunda, A.; Takaya, H.; Toriumi, K.; Ito, T. J. Am. Chem. Soc. 1980, 102, 7932. 38 Trost, B. M.; Van Vranken, D. L.; Bingel, C. J. Am. Chem. Soc. 1992, 114, 9327. 39 Brown, J. M.; Hulmes, D. I.; Guiry, P. J. Tetrahedron 1994, 50, 4493. 40 Trost, B. M.; Van Vranken, D. L. Chem. Rev. 1996, 96, 395. 41 Trost, B. M.; Toste, F. D. J. Am. Chem. Soc. 1998, 120, 815. 42 Trost, B. M.; Bunt, R. C. J. Am. Chem. Soc. 1994, 116, 4089. 43 Trost, B. M.; Bunt, R. C. J. Am. Chem. Soc. 1998, 120, 70. 44 Tsuji, J.; Shimizu, I.; Minami, I.; Ohashi, Y. Tetrahedron Lett. 1982, 23, 4809. 45 Trost, B. M.; Organ, M. G.; O’Doherty, G. A. J. Am. Chem. Soc. 1995, 117, 9662. 46 Trost, B. M.; Lautens, M. J. Am. Chem. Soc. 1987, 109, 1469. 47 Hunger, M.; Limberg, C.; Kircher, P. Organometallics 2000, 19, 1044. 48 a) Pearson, A. J.; Neagu, I. B.; Pinkerton, A. A.; Kirschbaum, K.; Hardie, M. J. Organometallics 1997, 16, 4346. b) McCallum, J. S.; Sterbenz, J. T.; Liebeskind, L. S. Organometallics 1993, 12, 927. c) Rubio, A.; Liebeskind, L. S. J. Am. Chem. Soc. 1993, 115, 891. 49 a) Malkov, A. V.; Baxendale, I.; Mansfeld, D. J.; Kocovsky, P. Tetrahedron Lett. 1997, 38, 4895. b) Malkov, A. V.; Davis, S. L.; Mitchell, W. L.; Kocovsky, P. Tetrahedron Lett. 1997, 38, 4899. 50 a) Malkov, A. V.; Spoor, P.; Vinader, V.; Kocovsky, P. J. Org. Chem. 1999, 64, 5308. b) Malkov, A. V.; Davis, S. L.; Boxendale, I. R.; Mitchell, W. L.; Kocovsky, P. J. Org. Chem. 1999, 64, 2751. c) Malkov, A. V.; Boxendale, I. R.; Dvorak, D.; Mansfield, D. J.; Kocovsky, P. J. Org. Chem. 1999, 64, 2737. 51 Trost B. M.; Lautens, M. Tetrahedron 1987, 43, 4817. 52 Trost, B. M.; Lautens, M. J. Am. Chem. Soc. 1982, 104, 5543. 53 Faller, J. W.; Linebarrier, D. Organometallics 1988, 7, 1670.


- 50 -

54 Faller, J. W.; Lamber, C. Tetrahedron 1985, 41, 5755. 55 Trost, B. M.; Hildbrand, S.; Dogra, K. J. Am. Chem. Soc. 1999, 121, 10416. 56 a) Faller, J. W.; Chao, K.-H. J. Am. Chem. Soc. 1983, 105, 3893. b) Faller, J. W.; Chao, K.-H. Organometallics 1984, 3, 927. c) Faller, J. W.; Mazzieri, M. R.; Nguyen, J. T.; Parr, J.; Tokunaga, M. Pure Appl. Chem. 1994, 66, 1463. 57 Trost, B. M.; Hachiya, I. J. Am. Chem. Soc. 1998, 120, 1104. 58 Glorius, F.; Pfaltz, A. Org. Lett. 1999, 1, 141. 59 a) Adolfsson, H.; Moberg, C. Tetrahedron: Asymmetry 1995, 6, 2023. b) Moberg, C.; Adolfsson, H.; Wärnmark, K. Acta Chem. Scand 1996, 50, 195. 60 Bald, E.; Saigo, K; Mukaiyama, T. Chem. Lett. 1975, 1163.

- 51 -

Acknowledgements My supervisor Prof. Christina Moberg for your never failing commitment in the process of making a Doctor of me. Everybody in the group Robert Stranne (your taste for good wines is almost as good as your taste for dangerous chemicals), Fredrik Lake (our ambassador in Germany), Fredrik Rahm (don’t get too cocky just because you are the Best in badminton!), Kristina Hallman (didn’t we have a really great collaboration - you, me and the DMX500?), Christina Nordlund (you really made our office seem much brighter), Per Renström (for being interested in discussing the really deep issues), Oscar Belda (heir to the MW10), Christian Linde (subgroup Linde), Sacha Legrand (our own Champagnoise!), Sergey Lutsenko (A whole bottle of whiskey, just for us…). Former group members Hans Adolfsson (your calm was really an inspiration!), Kerstin Nordström, Magnus Cernerud, Simon Dunne, Emmanuel Macedo (for being a good hood-pal), Lars Hagberg (thanks for all the good suggestions on synthesis) and all our exchange students for creating a dynamic atmosphere. Mats Larhed for showing me how to turn my “bad luck” with chemistry during my first two years by focusing on what really needed to be done to get useful results. Nils-Fredrik Kaiser and Prof. Anders Hallberg for an interesting and fruitful collaboration. Personal Chemistry (former Labwell) for kindly lending us one of their ovens. Not to forget interesting discussions with Åke Pilotti, Jacob Westman and Pelle Lidström. Gunhild Erdtman for generously sharing your insights into the English language. All my badminton-pals from the department – without you I would weight a lot more! Peter Piispanen for all our deep pizza-discussions about chemistry, life and everything. Malin Svedberg for support, discussions and a good eye for the errors in this thesis. My friends for dragging me out of the lab to other joys in life. Johan Andersson and Mats Svensson – what would I have done without your flying computer support? Ulla Jacobsson for showing me the inside of the Bruker machines, I still hope I wasn’t too pushy though… Krister Zetterberg for many interesting discussions (and helping me out with this thesis’ title!) All who work at the department. You can’t guess how boring it would have been to do research here and write this thesis without you to discuss with! www.mp3.com, − my supply of energy to write when decent people should be asleep.

http://www.mp3.com/

- 52 -

List of publications I Preparation of Chiral Enantiopure 2-(Hydroxyalkyl)pyridine Derivatives. Use of the Chiral Pool. Rahm, F.; Stranne, R.; Bremberg, U.; Nordström, K.; Cernerud, M.; Macedo, E.; Moberg, C. J. Chem. Soc., Perkin Trans. 1, 2000, In press II Palladium-catalyzed allylic alkylation using pyridino- and quinolino-oxazolines as ligands – influence of steric factors. Bremberg, U.; Rahm, F.; Moberg, C. Tetrahedron: Asymmetry 1998, 9, 3437. III (Hydroxyalkyl)pyridinooxazolines in Palladium-Catalyzed Allylic Substitutions. Conformational Preferences of the Ligand. Svensson, M.; Bremberg, U.; Hallman, K.; Csöregh, I.; Moberg, C. Organometallics 1999, 18, 4900. IV Rapid Microwave-Induced Palladium-Catalyzed Asymmetric Allylic Alkylation. Bremberg, U.; Larhed, M.; Moberg, C.; Hallberg, A. J. Org. Chem. 1999, 64, 1082. V Microwave-mediated palladium-catalyzed asymmetric allylic alkylation; an example of highly selective fast chemistry. Kaiser, N.-F. K.; Bremberg, U.; Larhed, M.; Moberg, C.; Hallberg, A. J. Organomet. Chem. 2000, In Press VI Rapid and Stereoselective C-C, C-O, C-N and C-S Couplings via Microwave-Accelerated Palladium-Catalyzed Allylic Substitutions. Bremberg, U.; Lutsenko, S.; Kaiser, N.-F. K.; Larhed, M.; Hallberg, A.; Moberg, C., Synthesis 2000, In Press VII Fast, Convenient and Efficient Molybdenum-Catalyzed Asymmetric Allylic Alkylation under Non-Inert Conditions: An Example of Microwave Promoted Fast Chemistry. Kaiser, N.-F. K.; Bremberg, U.; Larhed, M.; Moberg, C.; Hallberg, A. Submitted for publication VIIIElectronic and Steric Effects in Allylations Catalyzed by Mo-Pyridylamide Complexes Belda, O.; Kaiser, N.-F. K.; Bremberg, U.; Larhed, M.; Moberg, C.; Hallberg, A., Preliminary manuscript

Documents

Asymmetric Catalysis: Ligand Design and …kth.diva-portal.org/smash/get/diva2:8703/FULLTEXT01.pdfAsymmetric Catalysis: Ligand Design and Microwave Acceleration Ulf Bremberg Civ. Ing