ACTAUNIVERSITATIS
UPSALIENSISUPPSALA
2012
Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 984
Transition Metal Catalysisfor Selective Synthesis andSustainable Chemistry
J. JOHAN VERENDEL
ISSN 1651-6214ISBN 978-91-554-8507-8urn:nbn:se:uu:diva-182900
Dissertation presented at Uppsala University to be publicly examined in B42, BMC,Husargatan 3, Uppsala, Friday, November 30, 2012 at 10:00 for the degree of Doctor ofPhilosophy. The examination will be conducted in English.
AbstractVerendel, J. J. 2012. Transition Metal Catalysis for Selective Synthesis and SustainableChemistry. Acta Universitatis Upsaliensis. Digital Comprehensive Summaries ofUppsala Dissertations from the Faculty of Science and Technology 984. 125 pp. Uppsala.ISBN 978-91-554-8507-8.
This thesis discusses the preparation and use of transition-metal catalysts for selective organicchemical reactions. Specifically, two different matters have been studied; the asymmetrichydrogenation of carbon-carbon double bonds using N,P-ligated iridium catalysts and the metal-catalyzed transfer of small molecules from biomass to synthetic intermediates.
In the first part of this thesis, chiral N,P-ligands were synthesized and evaluated in iridiumcatalysts for the asymmetric hydrogenation of non- and weakly functionalized alkenes (PapersI & II). The new catalysts were prepared via chiral-pool strategies and exhibited superiorproperties for the reduction of certain types of alkenes. In particular, some of the catalystsshowed excellent activity and selectivity in the enantioselective reduction of terminal alkenes,and the preparation of a modular catalyst library allowed the asymmetric hydrogenation of awide range of 1,1-disubstituted alkenes with unprecedented efficiency and enantioselectivity(Paper III). Methods for the selective preparation of chiral hetero- and carbocyclic fragmentsusing iridium-catalyzed asymmetric hydrogenation as an enantiodetermining key step were alsodeveloped. A range of elusive chiral building blocks that have applications in pharmaceuticaland natural-product chemistry could thus be conveniently prepared (Papers IV & V).
The second part of this thesis deals with the catalytic decomposition of polysaccharides intosugar alcohols and the incorporation of their decomposition products into alkene substrates.Iridium-catalyzed dehydrogenative decarbonylation was found to decompose polyols intoCO:H2 mixtures that could be used immediately in the ex situ low-pressure hydroformylationof styrene (Paper VI). The net process was thus the hydroformylation of alkenes with biomass-derived synthesis gas.
Keywords: Catalysis, Transition metals, Asymmetric catalysis, Hydrogenation, Sustainablechemistry
J. Johan Verendel, Uppsala University, Department of Chemistry - BMC, Synthetical OrganicChemistry, Box 576, SE-751 23 Uppsala, Sweden.
© J. Johan Verendel 2012
ISSN 1651-6214ISBN 978-91-554-8507-8urn:nbn:se:uu:diva-182900 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-182900)
To my parents
List of Papers
This thesis is based on the following papers, which are referred to in the text by their Roman numerals.
I Development of pinene-derived N,P-ligands and their utility in cata-lytic asymmetric hydrogenation. Verendel, J. Johan; Andersson, Pher G. Dalton Transactions, 2007, 47, 5603–5610
II Biaryl phosphite-oxazolines from hydroxyl aminoacid derivatives:
highly efficient modular ligands for Ir-catalyzed hydrogenation of alkenes. Diéguez, Montserrat; Mazuela, Javier; Pàmies, Oscar; Verendel, J. Johan; Andersson, Pher G. Chem. Commun. 2008, 33, 3888–3890.
III Iridium phosphite-oxazoline catalysts for the highly enantioselective
hydrogenation of terminal alkenes. Mazuela, Javier; Verendel, J. Jo-han; Coll, Mercedes; Schäffner, Benjamin; Börner, Armin; Anders-son, Pher G.; Pàmies, Oscar; Diéguez, Montserrat J. Am. Chem. Soc. 2009, 131(34), 12344–12353.
IV Highly flexible synthesis of chiral azacycles via iridium-catalyzed
hydrogenation. Verendel, J. Johan; Zhou, Taigang; Li, Jia-Qi; Paptchikhine, Alexander; Lebedev, Oleg; Andersson, Pher G. J. Am. Chem. Soc. 2010, 132(26), 8880–8881.
V Chiral hetero- and carbocyclic compounds from the asymmetric hy-
drogenation of cyclic alkenes. Verendel, J. Johan; Li, Jia-Qi; Quan, Xu; Peters, Byron; Zhou, Taigang; Gautun, Odd R.; Govender, Thavendran; Andersson, Pher G. Chem. Eur. J. 2012, 18(21), 6507–6513.
VI Selective, metal-catalyzed transfer of H2 and CO from polyols to al-
kenes. Verendel, J. Johan; Nordlund, Michael; Andersson, Pher G. 2012, Manuscript
Reprints were made with permission from the respective publishers.
Paper I and II reproduced by permission of the Royal Society of Chemistry. Copyright 2007/2008, Royal Society of Chemistry. Paper III and IV reprinted with permission from the American Chemical Society. Copyright 2009/2010, American Chemical Society. Paper V reprinted with permission from John Wiley and Sons, Inc. Copyright 2012, Wiley-VCH Verlag GmbH & Co. Publications not included in this thesis: Chiral pyranoside phosphite-oxazolines: a new class of ligand for asymmet-ric catalytic hydrogenation of alkenes. Diéguez, Montserrat; Mazuela, Javier; Pàmies, Oscar; Verendel, J. Johan; Andersson, Pher G. J. Am. Chem. Soc. 2008, 130(23), 7208–7209. Catalytic one-pot production of small organics from polysaccharides. Veren-del, J. Johan; Church, Tamara L.; Andersson, Pher G. Synthesis, 2011, 11, 1649–1677.
Contribution report
The author wishes to clarify his contribution to the included papers. Paper I Performed all experimental work; contributed significantly to
the interpretation of the results and wrote the paper. Paper II Performed a part of the experimental work; contributed partly
to the interpretation of the results and writing of the paper. Paper III Performed a significant part of the experimental work, con-
tributed to the interpretation of the results and writing of the paper.
Paper IV Contributed to the research idea and formulation; performed a
significant part of the experimental work; contributed signifi-cantly to the interpretation of the results and wrote the paper.
Paper V Contributed significantly to the research idea and formulation;
performed a major part of the experimental work; contributed significantly to the interpretation of the results and wrote the paper.
Paper VI Contributed with the original research idea and formulation.
Performed a major part of the experimental work, contributed significantly to the interpretation of the results and wrote the paper.
Contents
1 Introduction ................................................................................................ 13 1.1 Organic Chemistry in Society ............................................................. 13 1.2 Stereochemistry and Chirality ............................................................ 13
1.2.1 Chirality ...................................................................................... 14 1.2.2 Preparation of enantiomerically enriched compounds by resolution .............................................................................................. 15 1.2.3 Asymmetric synthesis ................................................................. 18
1.3 Catalysis .............................................................................................. 20 1.3.1 Catalyst types .............................................................................. 23
1.4 Transition Metals ................................................................................ 23 1.4.1 The platinum group ..................................................................... 25 1.4.2 Iridium ......................................................................................... 25
1.5 Sustainable Chemistry ........................................................................ 27 1.5.1 Atom economy ............................................................................ 27 1.5.2 Renewable feedstocks ................................................................. 28
2 Selective synthesis ...................................................................................... 30 2.1 Hydrogenation of alkenes ................................................................... 30
2.1.1 Alkenes with coordinating functional groups ............................. 33 2.1.2 Non- or weakly-functionalized alkenes ...................................... 39 2.1.3 Chiral iridium N,P-ligated catalysts ............................................ 45
2.2 Catalyst development ......................................................................... 53 2.2.1 Selectivity in hydrogenation with chiral N,P-ligated iridium complexes ............................................................................................ 53 2.2.2 Pinene-derived ligands for asymmetric hydrogenation (Paper I)57 2.2.3 Phosphite-oxazolines as extremely modular ligands (Paper II) . 62
2.3 Method development .......................................................................... 66 2.3.1 Applications of chiral iridium-N,P complexes in synthesis ........ 66 2.3.2 Asymmetric hydrogenation of 1,1-disubstituted alkenes (Paper III) ........................................................................................................ 66 2.3.3 Catalytic asymmetric synthesis of chiral azacycles (Paper IV) . 72 2.3.4 Asymmetric hydrogenation of cyclic alkenes (Paper V) ............ 75
2.4 Conclusion .......................................................................................... 84
3 Renewable resources .................................................................................. 85 3.1 Metal catalysis for renewable feedstocks ........................................... 85
3.1.1 Hydrolysis of polysaccharides .................................................... 86 3.1.2 Formation of polyols from polysaccharides ............................... 90
3.2 Dehydrogenative Decarbonylation of alcohols .................................. 97 3.2.1 The transformation RCH2OH to RCHO + H2 ............................. 97 3.2.2 The transformation RCHO to RH + CO ................................... 100 3.2.3. Dehydrogenative Decarbonylation .......................................... 102 3.2.4 Transfer of CHOH from polyols to alkenes (Paper VI) ............ 104
3.3 Conclusion ........................................................................................ 111
4 Summary in Swedish ................................................................................ 112
5 Acknowledgements .................................................................................. 115
6 References ................................................................................................ 117
Abbreviations
Ac acetyl Ar aryl BArF tetrakis[3,5-bis(trifluoromethyl)phenyl]borate BINAP 2,2'-bis(diphenylphosphino)-1,1'-binaphthyl BINOL 1,1'-binaphthalene-2,2'-diol bipy 2,2'-bipyridine Bn benzyl Boc tert-butoxycarbonyl BPPFA (N,N-dimethyl-1-[2,1'-
bis(diphenylphosphino)ferrocenyl]ethylamine) tBu tert-butyl CBS Corey-Bakshi-Shibata Cbz benzyloxycarbonyl CI chemical ionization cod 1,5-cyclooctadiene Conv. conversion Cy cyclohexyl DAST diaminosulfur trifluoride DFT density functional theory DGDE Di(ethyleneglycol) diethylether DIOP 2,3-O-isopropylidene-2,3-dihydroxy-1,4-
bis(diphenylphosphino)butane DIPAMP 1,2-bis[(2-methoxyphenyl)(phenylphosphino)]ethane DMAP 4-dimethylaminopyridine DMF dimethylformamide DOPA 3,4-dihydroxyphenylalanine DPPP 1,3-bis(diphenylphosphino)propane DuPhos 1,2-bis[2,5-dimethylphospholano]benzene ee enantiomeric excess EI Electron impact ionization (electron ionization) ESI electrospray ionization Et ethyl GC gas chromatography HMF 5-hydroxymethylfurfural HPLC high pressure liquid chromatography mCPBA meta-chloroperoxybenzoic acid
Me methyl Ms mesityl NMR nuclear magnetic resonance N,C-ligand nitrogen, carbene - ligand N,P-ligand nitrogen, phosphorous - ligand NSAID non-steriod anti-inflammatory drug Ph phenyl PHOX phosphino oxazoline iPr iso-propyl RT room temperature SAMP (S)-(−)-1-amino-2-(methoxymethyl)pyrrolidine TFAA trifluoroacetic anhydride TFA trifluoroacetic acid TMS trimethylsilyl TOF turnover frequency TON turnover number TPP tetraphenyl porphyrine Ts tosyl UV ultraviolet (light)
13
1 Introduction
1.1 Organic Chemistry in Society Synthetic organic compounds (compounds containing carbon atoms) are fundamental for the function of the modern society. Fuels, drugs and plastics are well-defined compositions of organic molecules; whereas crops and foods are produced with the help of, and often containing additions of, tailor made organic substances. The prevalence and importance of these com-pounds necessitate efficient, clean methods for their production. Today, as higher demand is put on environmentally friendly production, the need for effective synthesis of chemicals is as important as ever.
1.2 Stereochemistry and Chirality Molecules that have the same composition of atoms are called isomers.1 For example, for the elemental formulae C7H9N and C3H6O, there are many pos-sible ways to connect the atoms together into a molecule. These molecules, having the same atomic constitution, are thus called constitutional or struc-tural isomers (Figure 1).
Figure 1. Two constitutional isomers of each of the formulae C7H9N and C3H6O.
Molecules that have the same atoms, connected in the same configuration, are called stereoisomers. As the connectivity of the atoms is the same, these isomers can only differ in their three-dimensional orientation. Stereoisomers are subdivided into diastereoisomers and optical isomers. Diastereomers are molecules that can be differentiated by some scalar property, usually the internuclear distances of selected pairs of atoms. Diastereomers include E/Z-,
NH2
NH2
H
O
HO
14
cis/trans- and conformational isomers (Figure 2). Since diastereoisomers have different energies, in cases where two diastereomers can interchange such as in the case of conformers, the equilibrium will be shifted towards one or the other.
Figure 2. Diastereoisomers.
Optical isomers, or enantiomers, are mirror images of each other that are not superimposable (Figure 3). Hence they are stereoisomers in which the inter-nuclear distances between any two atoms in one enantiomer is identical to that between the same two atoms in the other enantiomer. Enantiomers thus have the highest level of similarity of all isomeric compounds and they also have the same energy.
Figure 3. Two pairs of enantiomers.
Enantiomers are also called optical isomers because they rotate the plane of polarized light in opposite directions, + or –.
1.2.1 Chirality Closely related to enantiomers is the concept of chirality.2 Molecules that exist as pairs of enantiomers are called chiral compounds. The word chiral is derived from Greek and means “handedness”, in analogy to our hands that,
Br
Br OHOH
OH
OH
cis/trans isomersE/Z isomers Conformers
Ph
Ph
Ph
OH
Mirror
Enantiomer A Enantiomer B(S)-(–)-
1-Phenylethanol
Ph
OH
(R)-(+)-1-Phenylethanol
15
like enantiomers, are mirror images of each other and thus chiral. 1-Phenylethanol is a chiral compound that has two enantiomers, denoted S and R (Figure 3).2
Chirality is ubiquitous and immensely important in Nature, where two different enantiomers can interact very differently with other chiral mole-cules. Especially important is the fact that proteins in living organisms react differently towards the two different enantiomers of chiral molecules. (S)-Carvone for instance, smells like caraway, whereas (R)-carvone smells like spearmint (Figure 4).3
Figure 4. The two enantiomers of carvone interact differently with human olfactory receptors; thus they smell different.
By convention, the ratio of the two enantiomers in a sample of a chiral compound is expressed as the enantiomeric excess (ee) and is given in per-cent. The enantiomeric excess, as the name implies, is calculated as the ex-cess of the major enantiomer (say R) over the other (S) and is given by the formula:
A sample with a 1:1 ratio of the two enantiomers, which will have an ee of 0%, is called a racemic mixture or a racemate.
1.2.2 Preparation of enantiomerically enriched compounds by resolution The first separation of enantiomers was achieved by Pasteur in 1848.4 He assessed the phenomenon that a solution of tartaric acid salt extracted from wine rotates plane-polarized light, but that the same (racemic) compound, when synthesized in the laboratory, did not. Pasteur noted by optical inspec-tion that the laboratory-prepared tartrate formed crystals that were mirror images of each other. After manually picking out the two crystal enantio-mers, he could conclude that they, like the naturally derived tartrate, turned
(R)-Carvone
O
(S)-Carvone
O
Spearmint smell Caraway smell
(R) – (S)(R) + (S)
ee = 100 %
16
the plane of polarized light. When heating one of the pure enantiomers in basic water the compound lost its optical activity and reformed the racemate. Pasteur also noted that the two enantiomers rotated the light in equal amounts, but in opposite directions, and that a 1:1 mixture of the two did not produce any rotation.
The achievement by Pasteur illustrates an important problem in chemistry – the laboratory preparation of a chiral compound produces a 1:1 mixture of enantiomers in the absence of other optically active compounds. Pasteur was fortunate, one might say; racemic mixtures only rarely crystallize as two separate enantiomers (spontaneous resolution) so for most laboratory proce-dures, other methods for enantiomer separation have to be used.
The separation of racemic mixtures usually requires an enantiomerically enriched chiral element to be present.5 A single enantiomer of a chiral com-pound will interact differently with the two enantiomers of the racemic mix-ture, thus making separation possible; this is called chiral resolution. Resolu-tion is, in essence, the formation of two diastereomeric compounds or com-plexes that, by virtue of their different chemical properties can be separated. This can be realized by the preparation of diastereomeric pairs, which can be separated by conventional means. Scheme 1 a) shows the separation of a racemic amine via the synthesis and separation of two diastereomeric salts. Another alternative is to perform a reaction in which one enantiomer reacts faster with a chiral reagent than does the other (kinetic resolution), allowing (hopefully) easy separation of two markedly different compounds. This is illustrated in Scheme 1 b), where the two enantiomers of a racemic alkene reacts at different rates with a chiral hydrogenation catalyst. Today, a com-mon practice is resolution by chiral chromatography, in which a chiral sta-tionary phase separates enantiomers by non-bonding interactions. The inter-actions between the stationary phase and the individual enantiomers thus forms supramolecular diastereomers, one of which is more stable, so that one enantiomer is retained for longer than the other (Scheme 1 c)).
17
Scheme 1. Separation of enantiomers by a) the formation of diastereomeric com-pounds, b) kinetic resolution and c) chiral chromatography.
A major drawback of the resolution of racemates is that half of the com-pound, the undesired enantiomer, is wasted. Industrially, this is completely unacceptable and resolution is only used preparatively when the unwanted enantiomer can easily be racemized and subjected to sequential separations.
a)
b)
c)
H2N Ph
racemate
COOHPh
OH
+H3N Ph+COO-Ph
OH(S)-Mandelic acid
+H3N PhCOO-Ph
OH
Diastereomer A Diastereomer B
PhChiral catalystH2
Ph
+
Ph
Chiral cat.
Ph
Chiral cat.
Diastereomer A
Diastereomer B
Fast reaction
Slow reactionracemate
Ph
+
O
OH
NH
Chiral stationary phase
Ph Ph
RacematePh Ph
Fast elution Slow elution
S R
Ph Ph
timeChromatographic separation
= +
HPLC
O
O
Ar
HOWeaker
interactionStrongerinteraction
18
Of course, if pure enantiomers are to be used to separate racemates, one can ask how to obtain the optically pure compound in the first place, and here the phenomenon observed by Pasteur gives us a clue. The true origin of the homochirality in Nature, i.e. the reason why almost all naturally occur-ring amino acids and sugars have the same absolute configuration, has yet to be elucidated. Local fluctuations in the relative amounts of two enantiomers in a prehistoric racemic mixture, followed by amplification, for instance by autocatalysis, is a favored theory.6 Nature produces chiral compounds enan-tioselectively, so an obvious way to obtain enantiomerically enriched com-pounds is to take them from Nature and modify them. Chiral amino acids and sugars can be used as building blocks for other organic compounds, allowing their stereochemistry to be incorporated into the final product. This strategy to obtain optically active compounds is termed the ‘chiral pool ap-proach’.
1.2.3 Asymmetric synthesis7 The selective preparation of a single enantiomer of a compound has the same requirement as the methods of resolution described above, i.e. a component in the system has to be non-racemic. Thus, using a chiral substrate or reagent in a reaction can yield optically enriched products, and this is termed asym-metric synthesis. For example, the achiral ketone 1 reacts with (S)-1-amino-2-methoxymethylpyrrolidine (SAMP) under dehydrative conditions to form the chiral hydrazone 2 (Scheme 2).8 This is an example of substrate-controlled asymmetric synthesis; the chirality of the substrate becomes a part of the product. Treatment of 2 with base and an electrophile (propyl iodide in Scheme 2) gives only one enantiomer of 3 due to the influence of the chi-ral fragment. Removing the pyrrolidine moiety by ozonolysis gives a chiral ketone product 4 in very high enantioselectivity. As the pathway from 1 to 4 uses the chiral molecule SAMP only as a chiral steering agent (auxiliary), the overall process is an auxiliary-controlled asymmetric synthesis.
19
Scheme 2. Asymmetric synthesis of the chiral ketone 4 using SAMP as chiral auxil-iary.8b
When enantioselectivity is induced by a chiral reagent the process is called reagent-controlled asymmetric synthesis. Reduction of carbonyls by chiral boranes 5, as shown in Scheme 3 a), is an example of this strategy. The in-termediate adduct is not isolated but hydrolyzed at the end of the reaction. The reagent 5 is expensive, and should ideally be reused. Adding BH3 to the mixture cleaves the product–borane adduct and regenerates 5, as shown in Scheme 3 b). As 5 is not consumed in the reaction, it is now classified as a chiral catalyst and the reaction is an example of asymmetric catalysis. Effi-cient, selective asymmetric catalysis is the most powerful way to make enan-tiomerically enriched compounds, because only one enantiomer is produced, and only a small amount of catalyst is needed. Asymmetric catalysis is a major part of this thesis and will be discussed much more in detail in later sections.
Chiral fragment
NN
OMe1) Base2) PrI
NN
OMe
New chiralcenter
60% yield, >99% ee
1 2
3
H+/H2O O
O
4
H2NN
OMe
-H2O
SAMP
MeI
20
Scheme 3. Asymmetric synthesis of an alcohol by a) reagent-controlled asymmetric synthesis and b) asymmetric catalysis.
1.3 Catalysis9 Scheme 3 illustrated the asymmetric reduction of carbonyl compounds using the famous CBS (Corey-Bakshi-Shibata) reductive system, and how it is made catalytic.10 The term catalysis was coined by J. J. Berzelius when he wrote in his annual report to the Swedish Academy of Science 1835:11
“It is given then, that many, both simple and complex compounds, both in solubilized and solid form, have the ability to, without necessarily contrib-uting with its own constituents, produce transformation of other compounds. This is a new power, able to produce chemical activity, belonging to both inorganic and organic nature, which is surely more extensive than we have hitherto believed and the nature of which is hidden to us. When I call it a new power, I do not mean that it is a force independent of the electrochemi-
R R'
O
N+
B OH3B–
HPh
Ph
5
R R'
O
R R'
O
R R'
O
a)
b)
N+
B OH3B–
HPh
Ph
5
BH3
N+
B OH2B
HPh
Ph
R R'
OHH+, H2O
R R'
O
N+
B OH2B
HPh
Ph
–
–
BH2
21
cal properties of matter. On the contrary, I am unable to suppose that this is anything other than a kind of special manifestation of these, but as long as we are unable to discover their relationship, it will simplify our research to regard it as a separate power. It will also make it easier for us to refer to it if it possesses a name of its own. I shall therefore, using a derivation well-known in chemistry, call it the catalytic power of substances, and decompo-sition by means of this power catalysis, similar to how we use the word anal-ysis to denote separation of the component parts of bodies by means of ordi-nary chemical forces. The catalytic power appears to constitute the ability of substances that, just by their mere presence, and not by its reactivity, awak-en reactions that otherwise would be slumbering at the observed condi-tions.”
Today a catalyst is popularly defined as: An additive that speeds up a chemical reaction without being consumed in the process. This definition is however, inadequate, because a catalyst can enable completely different reaction pathways and cause the reaction to give different products than what would be obtained in its absence. Consider Scheme 4, in which a) shows the simple, uncatalyzed reaction between molecules A and B to form molecule AB. A and B reacts because the energy of AB is lower than the energies of A and B combined. Let’s imagine that this reaction is slow. It is slow because the energy barrier for the reaction is high. It can be sped up by, for example, heating or by adding an excess of A or B, but this requires a waste of energy or material, and frequently leads to the formation of byproducts. When a catalyst is applied to the process as in b), the catalyst typically reacts with one or both of the reagents A and B, making them more susceptible to react to form final product, because the energy barrier for the reaction between A, B and catalyst is lower than that for the uncatalyzed reaction. Upon releasing the product compound AB, the catalyst is reformed and can participate in another cycle.
22
Scheme 4. Catalysis is a kinetic effect. a) An uncatalyzed reaction has a high energy barrier and is thus slow. b) A catalyst lowers the energy barrier by reaction with the components, thereby speeding up the reaction.
A BSlowa)
b)
A B
A + B
A–B
E
Cat–AB–Cat–A
A + B
A–B
E
Cat.
Cat.
Fast
B
A
A
Cat. A
B
A B
23
Three important things should be noted: - As a catalyst is not consumed in a reaction, it can be used in substochio-metric amounts. Highly active catalysts can be used in amounts well below 1 mol%. - When A, B or both A and B react with the catalyst, they form a new com-pound with unique properties. That can make possible the steering of the reaction towards other products. - By modifying the catalyst, the reaction can be tuned to, for example, occur faster or produce other compounds.
1.3.1 Catalyst types Catalysts are usually divided into two groups; the homogeneous and hetero-geneous catalysts. In homogeneous catalysis, the catalyst is in the same phase as the reagents (usually dissolved in a solvent) whereas this is not the case for heterogeneous catalysts. Commonly, ‘heterogeneous catalysts’ are suspended solids that react with substrates in solution or in gas phase. As solids, these catalysts can easily be removed and recovered after the reaction, which is a huge advantage, especially in industrial settings. Homogeneous catalysts on the other hand, as they are in the same phase as the substrates, interacts more frequently with the substrate and usually gives products in higher selectivity. Homogeneous catalysts are well-defined, discrete com-pounds (such as the catalyst 5 in Scheme 3) and can thus more easily be modified and studied.
The effectiveness and efficiency of a catalyst are usually expressed in turnover number (TON) and turnover frequency (TOF), respectively.12 In homogeneous catalysis, which is the main concern of this thesis, the turnover number is usually defined as number of cycles (Scheme 4 b)) that the cata-lyst can undergo before it decomposes to inactive materials. The turnover frequency is defined as TON/time, that is, the number AB molecules that one catalyst molecule can produce in a given time.
1.4 Transition Metals Organic chemistry focuses mainly on the atoms carbon and hydrogen, but frequently also involves atoms from the nitrogen and oxygen groups, the halogens and other groups in the periodic table. Many organics also interact with transition metals (or transition elements). Transition metals are defined as elements whose atoms have, or can give rise to, electronic configurations containing incomplete d subshells.13 Hence atoms in groups 3–11 of the pe-riodic table are transition metals by definition. The group 12 atoms, which also belong to the d-block, have s2d10 configuration, and can form ions with snd10 (n= 1 or 2) configurations so they are not transition metals by defini-
24
tion. Only Hg has (once) been observed to form an incomplete d subshell.14 Group 12 atoms do, however, share several (structural) properties with the transition metals and are sometimes also considered as transition-metals or post transition-metals (Figure 5).
The partially filled d-orbitals of the transition elements gives rise to prop-erties that are rare or absent in the main-group elements. Three of these properties are of special importance for their ability to act as catalysts in chemical reactions: - Ability to form five or more chemical bonds. - They often have multiple accessible oxidation states (of similar energies). - The tendency to accept electron pairs, forming coordination compounds.
Figure 5. Truncated periodic table with the transition metals highlighted.
To obtain a closed (full) d shell and hence a noble-gas configuration, a tran-sition metal coordinates electron-donating ligands. The tendency of the met-al to attain a noble gas electron configuration is called the 18-electron rule,15 because 18 electrons are required to fill up the nd, (n+1)s and (n+1)p orbitals (i.e. 10 + 2 + 6 electrons). The steric and electronic properties of the ligands alters the properties of the metal atom to which they are coordinated, and given the vast array of neutral or anionic molecules that can act as a ligand in one way or another, the possibilities for changing the properties of the metal center becomes enormous.
An example is Vaskas complex, trans-Ir(CO)Cl(PPh3)2, which can act as an oxygen carrier by coordinating O2 that is bubbled through a solution of the complex (Scheme 5).16 Bubbling N2 through the solution releases the O2 again, but if the Cl– ligand is replaced by I–, the coordination of O2 becomes irreversible. This is probably due to the lesser electronegativity of the iodine atom, which allows more electron density to be donated to the metal, which in turn improves M→π* back-bonding between the metal and the O2 frag-ment, resulting in a more strongly bound O2.
Hf Ta W Re Os Ir Pt Au Hg Tl Pb Bi Po At RnBaCs
Rb Sr Y Zr Nb Mo Tc Ru Rh Pd Ag Cd In Sn Sb Te I Xe
K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Br Kr
Na Mg
Li Be
H
Al Si P S Cl Ar
B C N O F Ne
He
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
1
2
3
4
5
6
25
Scheme 5. The addition of oxygen to Vaska’s complex is reversible.
1.4.1 The platinum group The metals platinum, palladium, iridium, rhodium, ruthenium and osmium, are historically grouped together as the “platinum group metals” because they are often found in the platinum-rich residues left after other metals such as nickel and copper have been isolated.17
Although the 18-electron rule is valid for different transition metal com-plexes to different extents, and should really be regarded more as a trend than a rule, certain metals tends to be more prone to form stable complexes with other electron configurations. The platinum group elements in particu-lar can frequently form coordinatively unsaturated 16-electron complexes in addition to coordinatively saturated complexes that satisfy the 18-electron rule. This property makes the platinum group especially useful as catalysts in chemical reactions. These 16e– complexes are usually square planar and relatively electron-rich despite that they lack electrons compared to 18e– complexes. These electronic and structural characteristics give them, above all, one important property; the ability to easily engage in oxidative addition reactions, and subsequently their reverse reactions, reductive eliminations. These are pivotal reactions for a catalytic metal because they are involved in almost all catalytic reaction cycles.12
It should be noted that, although oxidative addition is observed for many transition metals, both late and early, products formed from oxidative addi-tion to the early transition metals are often quite stable owing to the electro-positive nature of these elements. Thus, reductive elimination from these elements is rare. Similarily, early transition elements often have faster mi-gratory insertions (the electron count around the metal decreases) and slower β-hydride elimination (increased electron count) than the late, platinum-group metals, which are more electronegative character.
1.4.2 Iridium Within the platinum group metals, the group 9 transition metals Rh and Ir tend to have properties between those of group 8 (Ru and Os) and group 10 (Pd and Pt) metals. Whereas Ru and Os can exist in oxidation states up to +8 (e.g. in OsO4), and Pd and Pt prefer the +2 oxidation state, Ir and Rh have not been observed in oxidation states higher than +6.17 This trend is presum-
IrPh3P ClOC PPh3
O2
Vaska's complex
IrPh3P
ClOC PPh3
OO
N2 - Bubbling-O2
26
ably due to the increasing effective nuclear charge that is present when mov-ing to the right in the period. The increasing nuclear charge better stabilizes d electrons, makes the element less prone to oxidation. Whereas ruthenium and osmium have rich oxide chemistry, rhodium and especially iridium dis-plays an equally impressive array of hydride complexes, which have been found for a range of oxidation states.18 For Rh and Ir, the most common oxi-dation state is +3 and, especially with π-acceptor ligands, +1 is frequently encountered; while +2 and +5 are less common. These metals have also, very rarely, been observed in the +4 and +6 states.
Some differences exist within group 9. Iridium, being in the sixth period (electron configuration: [Xe]6s24f145d7), can more easily achieve higher oxidation states due to its lower effective nuclear charge. The decrease in the effective nuclear charge when going down a group also results in increased polarizability or ‘softness’ of the metal. As an example, Co forms hexahalo-gen complexes only with F–, while Ir forms hexahalogen complexes with all halogens except F–. Additionally, as soft metals forms strong covalent bonds to soft ligands, ligand-metal bond strength increases down a group for most organic ligands. Another effect of going down a group is, of course, in-creased atomic size, and thus ability to coordinate more or bigger ligands. For instance, Wilkinsons hydrogenation catalyst, Rh(Cl)(PPh3)3, adds dihy-drogen to form the coordinatively saturated octahedral complex, Rh(H)2(Cl)(PPh3)3 (Scheme 6). The three triphenylphosphines are large and create a strained environment around the metal, so one of them readily dis-sociates to leave an open coordination site at which an alkene can bind and undergo hydrogenation. For iridium, which is a bigger and forms stronger bonds with the phosphine ligands, the three triphenylphosphines are much less prone to dissociate, making Ir(Cl)(PPh3)3 a poor hydrogenation catalyst.19 Co(Cl)(PPh3)3 does not act as a hydrogenation catalyst either. Cobalt, also being in group nine and certainly small enough to extrude PPh3, has much lower tendency to oxidize, and thus does not add dihydrogen oxi-datively at all.20
Scheme 6. Difference between Co, Rh and Ir in the reactivity with dihydrogen.
RhPh3P ClPh3P PPh3
H2
IrPh3P ClPh3P PPh3
H2
CoPh3P ClPh3P PPh3
H2
No reaction
Forms reversibly Forms irreversibly
RhPh3P Cl
H PPh3
H
PPh3
IrPh3P ClH PPh3
H
PPh3
27
1.5 Sustainable Chemistry Sustainable, or ‘green’, chemistry is the idea that the best ways to handle hazardous chemical waste is to not produce it in the first place. Hence, sus-tainable chemistry requires selective chemical transformations that produce little or no waste.
1.5.1 Atom economy9 Whereas chemo-, regio-, diastereo- and enantioselective reactions are crucial for efficient synthetic organic reactions, they do not, in themselves, guaran-tee the optimal use of reagents. Another important parameter is atom econ-omy (given in %), which indicates how many of the atoms that are input are found in the final product.
Atom economy is a theoretical term in that it does not take into account, for instance, the degree of conversion of the starting material. Perfect atom economy requires that all atoms of the reactants are present in the final com-pound. An illustrative example is the high-yielding Jones oxidation of di-phenyl methanol to benzophenone (Scheme 7). The reaction consumes two equivalents of chromium trioxide and three equivalents of sulfuric acid to produce three equivalents of benzophenone, one equivalent of chromium sulfate and six equivalents of water.
Scheme 7. Jones oxidation is selective and gives high chemical yields but has poor atom economy.
The atom economy of the reaction is only 52%, despite its high selectivity and yield. Additionally, significant amounts of hazardous chromium waste are produced and must be disposed of. The inherently poor atom economy obtained from reactions with stoichiometric reagents is a strong incentive for the use of catalysts in sustainable chemistry. An example highlighting the potential of metal catalysts in alcohol oxidation is shown in Scheme 8.21 The iridium-catalyzed dehydrogenation of 1-phenylethanol to acetophenone and dihydrogen is a reaction with 98% atom economy and gives yields as high as the chromic acid oxidation does. Instead of chromium waste, valuable hy-
Molecular weight of the desired product(s)Molecular weight of all reactants
atom economy = 100
quantitative yield
OH+ 2 CrO3 + 3 H2SO4
Ph Ph
O
Ph Ph+ Cr2(SO4)3 + 6 H2O
reagents waste
3 3
28
drogen gas is liberated. On the other hand, the reaction requires high temper-ature, which of course is undesirable, both from an economic and environ-mental viewpoint. Indeed, the development of more effective catalysts that operate at milder conditions is an important part of sustainable chemistry.
Scheme 8. The iridium-catalyzed dehydrogenation of alcohols is a reaction with high atom economy.
1.5.2 Renewable feedstocks In addition to not generating waste materials, sustainable chemistry advo-cates use of renewable feedstocks. Our modern society is built on petroleum, and the vast majority of the materials we use for fuels, plastics, drugs and other chemicals, come from crude oil. The cycle time for fossil fuels to be regenerated are >10 million years, making it a limited, non-renewable (with-in foreseeable future) resource.22 During the past 15 years or so, awareness of this issue has been rising, and the research efforts to find renewable sources to replace petroleum have grown into a significant area of chemical research. Renewable feedstocks are, as the name implies, continuously being regenerated. As for organic chemicals, plant biomass is the greatest potential source of renewable materials.23 The cycle time for organic materials derived from plants vary from under a year (crops) to hundreds (trees) of years, mak-ing them a potential renewable source provided that enough can be grown to cover demand. Major sources of plant biomass are wood, crops (such as wheat, maize and rice) and waste (such as agricultural waste and forestry residues) (Figure 6). As plants take up CO2 in order to grow, they are CO2 neutral, so the combustion of biomass-derived fuels can be performed with no net release of CO2.
Ph
OH
Ph
O+ H2
N Ir ClHO
0.001 6
p-xylene, refluxquantitative yield
6
29
Figure 6. Important sources of biomass. Adapted from reference 22.
Vegetable oils, such as rapeseed, palm or soybean-oil, that are used to pro-duce biodiesel, and the corn and sugarcane extracts used for bioethanol, are currently among the most important biomass materials in the fuel industry.24 Their production, however, is not “green” in every sense, as using them for chemical production competes with food production, consumes large amounts of fertilizers and often is only made possible by deforestation. Alt-hough agricultural and forestry residues remain useful sources of biomass, they are not produced in sufficient amounts to replace petroleum in the long run. Hence, woody biomass remains as a plentiful source of organic chemi-cals. The main constituents of woody biomass are polysaccharides such as cellulose and hemicellulose, which are long-chain polymers of sugars such as glucose and fructose. Cellulose is the most abundant organic compound on earth, produced in 180 billion tons annually by carbon dioxide capture in the green plants.25 For a future sustainable society, use of these raw materials in chemical manufacture will be essential.
30
2 Selective synthesis
2.1 Hydrogenation of alkenes Hydrogenation of carbon-carbon double bonds is one of the most widely used synthetic transformations both in laboratory and industrial settings. The transformation realizes the conversion of a C=C moiety to HC-CH using dihydrogen as the source of hydrogen atoms (Scheme 9). The reasons why the reaction is so widespread are that it features perfect atom economy, is often very chemoselective and operates under mild conditions.26
Scheme 9. Addition of dihydrogen across a carbon-carbon-double bond
Hydrogenation of carbon-carbon double bonds is thermodynamically fa-vored but, when mixing H2 and an alkene alone, the reaction rate is negligi-ble under normal laboratory conditions, so to achieve the transformation efficiently transition metal catalysts are used. Although the vast majority of hydrogenations that are carried out are per-formed using heterogeneous metal catalysts (Pd/C, PdO2, Pt/Al2O3 etc.), asymmetric hydrogenation is most efficiently done using homogeneous cata-lysts and since this is the subject of the chapter, heterogeneous catalysis will not be discussed in any depth.
The H-H bond is relatively strong but upon coordination to a metal, L→M σ-donation from the bonding electrons and L←M π-back-donation into the antibonding orbital results in that the bond is weakened and eventu-ally broken completely.27 In general, as long as the metal is sufficiently elec-tron rich, the cleavage will be homolytic, resulting in oxidative addition and formation of a dihydride complex, increasing the formal oxidation state of the metal by two. For electrophilic metal centres however, and especially when bases are present, heterolytic hydrogen cleavage is common (Scheme 10). Heterolytic cleavage results in the formation of a proton and a metal hydride where the oxidation state of the metal is unchanged. The proton is usually removed from the metal via ligands (as HCl for example) or by bases available in the reaction medium. In more extreme cases, η2-H2 adducts can be very acidic with pKa well below 0.28 By acid-base equilibria, heterolytic
H2+ H H
31
cleavage is hampered by acid and promoted by base.29 Throughout the rest of this thesis, the discussions will mainly concern homolytic cleavage of dihydrogen, but it is worth remembering that the mechanisms are comple-mentary and that changing the reaction conditions may favor one or the other type.
Scheme 10. Different types of dihydrogen cleavage by metal fragments.
Considering the homolytic case, for solubilized homogeneous metal com-plexes, the difference in energy between the dihydrogen and the dihydride complex is relatively small, and most often the two forms exists in equilibri-um (Scheme 10).30 By varying the degree of back-donation from the metal by adding more electron donating or electron withdrawing ligands, the equi-librium can be shifted in either direction. Especially, the ligand trans to the dihydrogen ligand will have a strong influence on the degree of H–H and H–M bonding. By adding a strong σ-donor or π-acceptor in the trans position, dihydrogen is stabilized towards oxidative addition by decrease of L→M donation or L←M back-donation respectively.31 Intermediate ligands, i.e. π-donors and weak σ-donors, tends to promote formation of the dihydride when occupying the trans position. It should be remembered thou, that for the general oxidative addition case, oxidative addition is promoted by high electron density on the metal, so strong donor ligands facilitate oxidative addition while electron withdrawing ligands counteracts it, provided that they are not in the trans position. As far as hydrogenation is concerned, the “best” hydrogen transfer catalyst should be the intermediate case, where neither the dihydride nor the dihydrogen complex is too stable. The relative stability of M–(H)2 vs. M–(H2) is also affected by the metal atom itself, where the stability of the hydridic species, at least to some extent, increases going down a group.32 For instance, in the Co–Rh–Ir triad, stoichiometric hydrogenation (under Ar) of dimethyl maleate can only be performed with [(PP3)M(H2)][PF6] (PP3 = P(CH2CH2PPh2)3) when M = Rh (Scheme 11).33 The cobalt complex does not undergo oxidative addition of H2 but instead undergoes ligand substitution upon addition of the alkene. Thus, the dihy-drogen is minimally activated if at all, the reaction products are the isomer-ized alkene and H2(g). For the iridium complex on the other hand the situa-tion is reversed and the stable compound [(PP3)Ir(H)2][PF6] is formed. Since
MH2 M
M
H
H
H HLM LH2 +
Homolytic cleavage
Heterolytic cleavage
MH
H
!2-dihydrogen complex
dihydridecomplex
32
this coordinatively saturated compound forms irreversibly due to the strong-er Ir-H bonds, the reaction fails as there are no coordination sites available for the alkene.33
Scheme 11. Properties of coordination compounds are metal-dependent. E = COOMe.
As demonstrated in the above example, alkene hydrogenation also requires coordination of the C=C double bond to the metal. This coordination is simi-lar to the one of dihydrogen in that the ligand acts as a σ-donor - π-acceptor. Contrary to the H2 case though, the alkene coordination is strongly directed by steric effects and in square-planar complexes for instance a bound ethene molecule will usually align itself vertically to minimize interactions with the other equatorial ligands.34 Another notable difference is that while the barrier to rotation around the M-L bond is small for L = H2, the size of the alkene, with its substituents, hampers rotation around the M-L σ-bond.35 The elec-tronic interactions between the alkene and the metal fragment decreases the C=C bond order and eventually results in formation of a metallacyclopro-pane. In order for the alkene to act as an electrophile towards hydrides and other nucleophiles, very electron rich metal centers should be avoided since excessive back donation results in an electron rich species of metallacyclo-propane character (Scheme 12).
CoP
PPP
H H
RhP
PPP
H H
IrP
PPP
H H
E
E
+ H2
RhP HP H
P
P
CoP HP H
P
PDoes not form
E
E+
E
E+
E
E
Forms reversibly
IrP HP H
P
P
E
E
+ No reaction
Forms irreversibly
+ +
++
+ +
33
Scheme 12. The two extremes of alkene-metal coordination.
As the η2-alkene-metal-hydride fragment has formed, the next fundamen-tal step is migratory insertion to form an (usually highly unstable) alkyl σ-complex (Scheme 13, step i). The new alkyl-metal complex initially has an agostic interaction between the newly freed up coordination site and the new C-H bond. In order to ensure that β-hydride elimination does not occur to reverse the reaction, the free coordination site should as quickly as possible be occupied by a new ligand.
Scheme 13. Two fundamental steps of alkene hydrogenation; i) Migratory insertion (forward) and β-hydride elimination (backwards). ii) Reductive elimination.
Following formation of the σ-alkyl complex is reductive elimination of the alkyl group along with another hydride ligand from the metal, to form the product alkane (Scheme 13, step ii). Being the reverse of oxidative addition, reductive elimination is promoted by electron withdrawing ligands that stabi-lize the low-valent metal. Especially for the platinum group metals, that tends to shift between octahedral 18e– (d6) (products of oxidative addition) and 16e– (d8) square planar (products of reductive elimination) complexes in catalytic cycles, the presence of bulky ligands stabilizes the square planar geometry and facilitates reductive elimination. Contrary to all other steps in homogeneous hydrogenation as described before, the reductive elimination is almost always considered to be irreversible.36
2.1.1 Alkenes with coordinating functional groups The first records on asymmetric hydrogenation of a carbon-carbon double bond dates back to 1968, only a few years after the discovery of RhCl(PPh3)3 (Wilkinson’s catalyst), the first highly active hydrogenation catalyst.37 Horn-er and Knowles independently reported enantioselective reduction of 1,1-disubstituted alkenes using complexes of the type RhCl(PRR’R’’)3, i.e. with chiral monodentate phosphines in place of the PPh3. Knowles performed the reduction of α-phenylacrylic acid using 15 mol% catalyst and (R) -
M M
!-ComplexElectron-rich metalElectron-poor alkene
"-ComplexElectron-poor metalElectron-rich alkene
MH M
H++
MH
i ii
++
MH
HMH
HH H
H
34
P(Me)(iPr)Ph as the chiral ligand. Optically active hydratropic acid was ob-tained in 15% optical yield.38 Horner reduced α-ethylstyrene to (S)-2-phenylbutane in 8% optical yield using a similar phosphine (S)-P(Me)(nPr)Ph as the chiral ligand.39 These were direct extensions of Wil-kinson’s achiral system where the triphenylphosphines had been replaced by chiral units. The real birth of asymmetric hydrogenation should perhaps ra-ther be associated with the preparation of the first chiral bidentate P,P-ligand DIOP in 1971.40 Dang and Kagan mixed DIOP (synthesized from (–) diethyl tartrate) with [Rh(coe)2Cl]2 to form the active catalyst in situ. The catalytic system reduced several alkenes with >50% enantioselectivity (Scheme 14).
Scheme 14. Asymmetric hydrogenation of dehydroaminoacids with Rh-DIOP.
Building on the work by Schrock and Osborn, that compounds of the type [Rh(diene)(phosphine)2][BF4] were viable catalyst in hydrogenation (vide infra),41 Knowles, using the chiral bidentate ligand 1,2-bis(o-anisylphenyl-phosphino)ethane (DIPAMP), could perform asymmetric hydrogenation of dehydroaminoacids such as S1, that gave the products in selectivity above 90% ee (Scheme 15).42
COOEtEtOOC
OH
OH
O
O
Ph2PPh2P
(-)-DIOP
AcHN COOH
0.02 [Rh(coe)2Cl]20.04 DIOP0.10 Et3N
r.t. 1 atm H2Benzene/EtOH
AcHN COOH*
Ph Ph
72% ee
35
Scheme 15. Asymmetric hydrogenation of dehydroaminoacids by Rh-DIPAMP.
The Rh-DIPAMP hydrogenation to produce the anti-parkinson agent L-DOPA became a commercial process (Scheme 15),43 several other P,P-ligated systems were developed and soon asymmetric hydrogenation of de-hydroaminoacids was a well established method with a broad substrate scope.
The mechanism behind the activity and selectivity of the catalysts such as [Rh(cod)(P,P*)][BF4] has been studied intensely and a deep knowledge has been acquired.44 These catalysts work best in alcoholic solvents that can stabilize the complex and separate the ion-pair in order to leave the metal cation naked for reactions.45 Upon subjection to H2, the diene is removed by hydrogenation forming [Rh(P,P*)(S)2]+ (S = Solvent) (A, Scheme 16). A then coordinates a substrate molecule through the alkene and, if available, also through a coordinating functional group in the substrate to form B. The rate-determining step then takes place as the complex oxidatively add dihy-drogen to form C which is highly reactive and quickly reacts further to form the σ-alkyl D. Reductive elimination then releases the product alkane and re-forms the species A, ready to partake in another cycle.
Ph
COOHAcNH
0.005 [Rh(cod)(L1)][BF4]3 atm H2
Ph
COOHAcNHMeOH, 50 oC, 1hS1 94% ee
DIPAMPL1
P
P
Ph
Ph
O
O
COOHNH2
HO
HO
L-DOPA
36
Scheme 16. Mechanism for the catalytic hydrogenation of alkenes by Rh-P,P sys-tems.
It is worth mentioning that the substrate coordination mode in B usually does not determine the reaction outcome in terms of absolute configuration of the product, rather it is usually a less stable conformer of B that reacts fastest with H2 to form C and subsequently D.44
In 1980, the axially dissymmetric ligand BINAP was first developed by Noyori and co-workers.46 Although it showed good selectivity in the rhodi-um-catalyzed hydrogenation of dehydroaminoacids, the true worth of this ligand was not realized until 1987, when Noyori and co-workers found that Ru(BINAP)(OAc)2 could reduce prochiral allylic and homoallylic alcohols with high selectivity.47 Such substrates had previously been impossible to reduce selectively. The same year they also presented the highly selective reduction of α,β-unsaturated carboxylic acids using the same catalytic sys-tem (Scheme 17).48
RhPP
SS*
RhPP*
2 H2
Cycloctane
NH
O
R''
R'RhPP*
NHO
R''
H2NH
O
R''
RhH
PH
P
R'
NHO
RhP
H
R'
R''R'
*P*
S S
S
2 S
A
C
+
+
+
+
+NHO
R''
R'*
2 S
BD
37
Scheme 17. Asymmetric hydrogenation of unsaturated carboxylic acids using Ru-BINAP.
The ruthenium-P,P catalytic system is significantly different from the rhodi-um one used mainly for dehydroaminoacids. While the former features a Rh(I)/Rh(III) dihydride mechanism, the ruthenium system progresses through Ru(II) throughout and is classified as a monohydride hydrogenation cycle.49 The catalytic cycle for hydrogenation of α,β-unsaturated carboxylic acids is shown in Scheme 18, the mechanism is similar for other types of alkene substrates. The acetates of the pre-catalysts are easily displaced by substrate molecules and the catalyst activated by heterolytic splitting of H2 to form the anionic species F and a proton which is taken up by the solvent. Methanol or other alcoholic solvents are preferred as they allow fast shuf-fling of protons from and to the catalyst. As in the case of rhodium, the re-versible cleavage of H2 is considered to be the turnover-limiting step. The hydride then migrates to the alkene to form G, which takes up a proton and reductively eliminates the alkane, yielding H. Substitution of the product with a new alkene forms I which progresses again through the cycle. As expected, presence of acids significantly slows the catalyst down since it counteracts the release of H+ in the rate-determining step.
COOH0.002 Ru(L2)(OAc)2
112 atm H2
MeOH, 25 oC, 24hS2 92% ee
COOH
PPh2PPh2
BINAPL2
38
Scheme 18. Ruthenium-catalyzed hydrogenation of unsaturated carboxylic acids.
In 1990, Burk and co-workers developed the DuPhos ligand which was unique in the sense that it was probably the first ligand that could be easily modified and tuned to fit a specific substrate class.50 Another new type of alkene, enol esters, could thus be reduced in high enantioselectivities using DuPhos variants.51 Using a development of the DuPhos ligand, BPE, Burk’s group was able to, for the first time, reduce simple enamides in high enanti-oselectivity.52 This had previously only been done with a limited set of cy-clic substrates using Noyori’s system.53 Other novel classes that could be reduced using their system was α-enolbenzo- and acetamido-phosphonates54 and β-acylamino acrylates.55
A rough timeline over the asymmetric hydrogenation of functionalized alkenes with the metal complexes discussed is presented in Figure 7.
RuO
O
PP
OO*
H2RuO
H
PP
O*
H+O O
R'R''
RuO
PP
O* O
OR''R'
H+RuO
O
PP
OO*
R'
R''
*
*R'
HO
R''
O
*
R'
HO
R''
O
RuO
O
PP
OO*
HOAc
R'
HO
R''
O
R' R''
RuO
O
PP
OO
R
*
H2
H+
–
–
E
F
G
H
I
39
Figure 7. Timeline over asymmetric hydrogenation of alkenes using P,P-ligated metal complexes. Only high selectivities (>80% ee) are considered.
2.1.2 Non- or weakly-functionalized alkenesWhile the rhodium and ruthenium diphosphine based catalytic systems have developed over the last 40 years to include many types of alkene substrates, these systems are very dependent on a secondary coordination to achieve high activity and enantioselectivity. Without the additional heteroatom coor-dination, the metal-alkene complex gains significantly more conformational freedom and the steric environment directly around the alkene presented by the P,P-ligands is then not enough to maintain high enantioselectivity (Fig-ure 8).
Figure 8. Non-chelating substrates allow more coordinating modes to a M-P,P fragment and thus lowers the selectivity in catalytic hydrogenation.
1960 1970 1980 1990 2000 2010
Wilkinson's catalyst BINAP
DuPhos & BPE
COOR' NHCOR''
R
DIPAMP
First asymmetric hydrogenation
OCOR'' P(O)(OR')2
RR COOH
R' R
OH
R'
R' NHCOR'
R
R' OCOR''
R
R NHCOR''
COOR'
DIOP
MP P
XO
R
*
MP PL
R
*
MP PL
R
*
MP P
XO
R
*
MP PL
R
*
MP PL
R
*
Alkene with coordinatingfunctional group
Alkene without coordinating functional group
40
Additionally, since the substrate competes with alcoholic solvent molecules for coordination to the metal, non-chelating alkenes are reduced much more slowly than alkenes bearing coordinating functional groups.
Attempts to use P,P-ligated Rh and Ru systems for asymmetric hydro-genation of alkenes lacking an adjacent coordinating group has almost ex-clusively proven unsuccessful.56 A table of published results on asymmetric hydrogenation is shown to illustrate this issue (Table 1), note that two of the most outstanding diphosphine ligands, BINAP and DuPhos were used.
Table 1. Asymmetric hydrogenation of α-substituted styrenes using P,P-ligated rhodium and ruthenium catalysts.
When discussing coordinating and non-coordinating groups, we mean that good coordinating functional groups are the acetamides and enol esters. Less powerful but still viable coordinating groups are the esters, vinyl phospho-nates and allylic alcohols. Metal-coordination through these groups would form more strained (4.5 coordinate) metal chelate and are thus less easy to form. Even less functionalized alkenes are compounds where the heteroatom that contain the free electron pair is moved further away in the molecule. Chelation in alkenes of this type is not beneficial due to the large ring size of
R
Chiral metal catalystH2 R*
REntry
4
6
7
2
1
3
5
29
9
ee (%)[M]/Ligand (mol%)
0.0040 Ru(L2)(OAc)2Et
Et
Et
iPr
iPr
tBu
tBu
0.0075 [Rh(cod)(L2)][Cl]
0.0040 Ru(L2)(OAc)2
0.0075 [Rh(cod)(L2)][Cl]
0.0040 Ru(L2)(OAc)2
0.0075 [Rh(cod)(L2)][Cl]
16
30
35
40
PP
DuPhosL3
0.0050 Ru(L3)Cl20.0250 tBuOK 86
41
the chelate and can also be prevented in cases where the geometrical shape of the substrate prohibits chelation. We consider non-functionalized alkenes as those not containing any functionality at all, i.e. completely aliphatic sub-stances (with the exception of the alkene function in itself). Benzene rings could of course theoretically act as coordinating groups by π-stacking with aryl-groups from the ligands, but usually alkenes containing aromatic hydro-carbons are considered non-coordinating as well. Of course, there are no clear-cut lines in most cases, and some substrates can behave as coordinating in some cases and non-coordinating in other, but a rough classification can definatly be made, as presented in Figure 9.
Figure 9. Classification of alkenes based on their ability to chelate a metal fragment.
Up until 15 years ago, the only general system that reduced non-functionalized alkenes in high selectivity were the titanocene system devel-oped by Buchwald.57 Several trisubstituted alkenes could be reduced in en-antioselectivites >90% using this system (Scheme 19). It was later developed into a catalytic system based on zirconium that efficiently hydrogenated even tetrasubstituted alkenes.58 Both of these systems however suffered from drawbacks; high catalyst loadings and long reaction times were required, and the catalyst had to be activated by nBuLi and was highly sensitivity to oxy-gen.
P(O)(OR)2
COOR
Non-functionalizedFunctionalized Weakly-functionalized
R'OOC NHCOR
COOH
OH
B
NHCOR
Cy Ph
Bu
O
O
SO2Ar
42
Figure 19. Asymmetric hydrogenation of unfunctionalized alkenes with Buchwald’s titanocene catalyst.
Iridium N,P–systems
Wilkinsons catalyst, RhCl(PPh3)3, was the first really active homogeneous hydrogenation catalyst (2, Scheme 20).37,59 It can hydrogenate mono- and disubstituted alkenes, but tri- and tetrasubstituted are less easy to reduce using this complex. Wilkinson’s catalyst, upon addition of hydrogen forms the octahedral, coordinatively saturated species RhCl(H)2(PPh3)3, which is in equilibrium with the catalytically active RhCl(H)2(PPh3)2.60 In order to im-prove catalytic efficiency, modified catalysts of the type [Rh(cod)(PPh3)2][PF6] were developed by Schrock and Osborn (3, Scheme 20).41b
Scheme 20. Wilkinson’s catalyst 2 requires of the alkene to compete with PPh3 while the system 3 is more active due to easier competition with solvent molecules.
Ti H
0.05 1130 atm H2
THF, 50 oC, 48h95% ee
1
O O
RhPh3P ClPh3P PPh3
RhPh3P PPh3Ph3P H
RhPh3PPh3P
+
PF6–2 H2 Cyclooctane
RhPh3P SPh3P S
+
PF6–
3
Cl
HH2 RhPh3P HPPh3Cl
H
PPh3
2
Ox. Add. of H2 and alkenecoordination
43
The [Rh(cod)(PPh3)2][PF6] system is more soluble in polar solvents and at least as active as Wilkinson’s, probably because the coordination of alkene does not requires dissociation of a triphenylphosphine, but they still failed at reducing bulky alkenes. They are also less air sensitive since the non-coordinating counter ion PF6 leaves a formal positive charge on the metal. In both cases, upon alkene coordination, the two PPh3 groups are both cis to the alkene (mutually trans), thus creating a quite sterically hindered olefin coor-dination site (Scheme 20).61 The reactions are typically performed in coordi-nating solvents (alcohols, acetone or THF) in order to stabilize the catalyst and prevent precipitation.
In 1977, about 10 years after Wilkinson’s discovery of RhCl(PPh3)3 as hydrogenation catalyst, Crabtree and co-workers also studied the catalytic system [(M(diene)(Phosphine)2][PF6] (M = Ir or Rh). While both metal complexes exhibited catalytic behavior in coordinating solvents, the iridium catalyst was much less active.62 This difference in activity was attributed to formation of more stable solvate-complexes of iridium. The iridium com-plexes [Ir(solv)2(PPh3)2][PF6] and the hydrogen adduct [Ir(H)2(solv)2(PPh3)2][PF6] were stable enough to be isolated. Crabtree rea-soned that by switching to a non-coordinative solvent, the activity could be increased. In benzene however, the solubility of the catalysts were low lead-ing to poor activity. Chlorinated solvents such as CH2Cl2 and CHCl3 that are both polar and non-coordinating gave significant improvement of the reac-tion rates, especially for the iridium complex.63 In a ligand screening exper-iment, the mixed-ligand complex [Ir(cod)(Py)PCy3][PF6] (4, Scheme 21) was found to be an exceptionally active hydrogenation catalyst.64 Not only was it faster than the corresponding diphosphine catalysts, but it also could reduce tri- and tetrasubstituted alkenes.19 The corresponding rhodium-catalyst did not show the same ability to reduce highly hindered alkenes, partly because the smaller rhodium atom creates a more sterically crowded active site.65 The reason for the improvement in catalytic efficiency when going from [Ir(cod)(phosphine)2][PF6] to [Ir(cod)(Py)PCy3][PF6] was mainly attributed to sterical factors, pyridine is a considerably smaller ligand, and while the ligands are trans in the diphosphine case, the low trans effect and small size of the pyridine seems to favor a cis conformation of the pyridine and PCy3 ligands.66 It is thus possible that upon addition of dihydrogen, a complex with an extremely open coordinating site could form, able to coor-dinate even tetrasubstituted alkenes (Scheme 21).
The results obtained by Crabtree strongly indicates that the catalyst rest-ing state is Ir(H)2(Py)(PCy3)(alkene)(solv) and the rate determining step in hydrogenation being the migratory insertion of the alkene into the Ir-H bond.
44
Scheme 21. A possible explanation for the high activity of Crabtree’s catalyst: 4 form an extremely open alkene coordination site compared to the similar diphos-phine complex 3. X = Solvent, alkene or H2.
In migratory insertion, the coordination number of the metal is decreased by one and the substitution of the diphosphines for the more electron rich PCy3/Py ligands might thus facilitate this step.
Interestingly the complex [Ir(cod)Py2][PF6] does not add dihydrogen oxi-datively and when comparing [Ir(cod)(PR3)2]+ with [Ir(cod)Py(PR3)]+, the former is more prone to undergo oxidative addition. This is contrary to what one would believe since pyridine is a strong σ-electron donor and a very weak π-acceptor, thus increasing the electron density on the metal. The iridi-um complexes of the type [Ir(cod)(L)2][PF6], are unusually acidic and con-trary to similar complexes more prone to coordinate halides (act as an acid) than protons (act as a base). Thus, [Ir(cod)LL’][PF6] behaves, at least to some extent, different from what is usually seen in similar complexes.19 We will later see, when examining these systems further, that the increased elec-tron-density supplied by PCy3 and pyridine can help the metal to reach high-er oxidation states, probably speeding up the catalysis.
Crabtree’s catalyst is also unique because it is exceptionally insensitive to oxidation by O2 and haloalkanes (the reactions are run in CH2Cl2). It is slowed somewhat by coordinating functional groups such as ketones and alcohols that can coordinate to the metal, and with amine bases, the catalyst is slowed down significantly by deprotonation.19 It also suffered one great disadvantage; as the high activity of the catalyst was dependent on non-coordinating solvents, the catalyst became prone to deactivation.64 While coordinating solvents can stabilize the catalyst by coordination to the metal, non-coordinating solvents does not stabilize metal complexes to any great extent. Thus, when the alkene is not a very good ligand, as in the case of tri- and especially tetra-substituted alkenes, or when the alkene has been con-
RhPh3PPh3P
+ PF6–
Cyclooctane
3
3 H2
RhH PPh3
Ph3PX
H
IrCy3PPy
+ PF6–
4
IrCy3P HPy
X
H
Cyclooctane
3 H2
+ +
Very unhinderedcoordination site
X X
45
sumed, the catalyst irreversibly form the trimeric iridium species [(Ir(H)2PyPCy3)3(µ3-H)][PF6]2.67
2.1.3 Chiral iridium N,P-ligated catalysts In 1997, based on Crabtree’s catalyst [Ir(cod)PyPCy3][PF6] (4), Pfaltz and co-workers synthesized the first chiral [Ir(cod)(N,P)][PF6] complex using a chiral phosphinooxazoline (PHOX)68 as the N,P-chelating species (Scheme 22, L4 and L5).69 While the complex only performed moderately in the hy-drogenation of imines, it was later tested in the hydrogenation of tri- and tetrasubstituted alkenes with very good results (Scheme 22).70
Scheme 22. Pfaltz’s first asymmetric hydrogenation of tri- and tetrasubstituted al-kenes using chiral [Ir(cod)(PHOX)][BArF] complexes as precatalysts.
The Ir-PHOX catalyst initially had the same problem as Crabtree’s catalyst; it deactivated over the course of the reaction and for catalyst loadings below 3 mol%, full conversion of the substrate could not be obtained.70-71 With the conclusions drawn by Crabtree in mind, that the catalyst deactivated due to poor alkene coordination, Pfaltz and co-workers reasoned that the problem might be counteracted by using a more weakly coordinating counter ion. A
IrPN
+ BArF–
*
Ph
Ph
0.001 [Ir(cod)(L4)][BArF] Ph
Ph*
PN
* =N
O
Ph
>99% conv.97% ee
R
PhR2
50 atm H2
CH2Cl2, 25 oC, 2h
R
PhR2
R1*R1
*
0.02 [Ir(cod)(L5)][BArF]
Ph*
>99% conv.81% ee
L4 L5
PPh2 N
O
(o-tol)2P
46
very weakly coordinating counter ion would be loosely attached to the metal and thus help alkene coordination. Indeed, replacing PF6 with the extremely weakly coordinated counter-ioyn BArF (BArF = tetrakis(3,5-bis(trifluoromethyl)phenyl)borate) (Scheme 23) similar anionic species showed increased catalyst turnover rate and stability.70,72 Since BArF is read-ily synthesized by quadruple alkylation of BF3 with 3,5-ditrifluoromethyl-bromomagnesium-benzene73 it has become the standard counter-ion to use with these catalytic systems.
Scheme 23. Preparation of NaBArF.
Following the successful introduction by Pfaltz, a range of chiral Crabtree mimics of the type [Ir(cod)(N,P*)][BArF] have been evaluated for the asymmetric hydrogenation of non- and weakly-functionalized alkenes.74
Mechanistic aspects of asymmetric hydrogenation using chiral mimics of Crabtree’s catalyst
The mechanism by which hydrogenation with chiral analogues of Crabtree’s catalyst operates has been proven difficult to elucidate experimentally. The majority of the indications are based on quantum chemical calculations and the mechanism still a subject of discussion, yet some experimental data on the topic has been accumulated. Pfaltz, Meuwly and co-workers studied the addition of dihydrogen to [Ir(cod)(PHOX)][BArF] in THF using NMR.75 They found that, as in the case of Crabtree’s catalyst, cyclooctadiene is quickly removed by hydrogenation, forming mainly the two diastereomers of the dihydrogen addition product with the hydrides in cis conformation and one of them trans to the oxazoline nitrogen (Scheme 24). The additional two coordination sites were occupied by THF molecules. Quantum chemical calculations suggested that the reason for the diastereomer distribution was that the N,P-ligand presented a larger steric bulk above the N-Ir-P plane, thus preffering to place the small hydride ligand there (Scheme 24). For di-chloromethane, a complex mixture of hydridic species was obtained, but DFT calculations indicated that similar structures were reasonable.
BF3 4
F3C CF3
MgBr
+
F3C CF3
B
1) Reflux, o.n.2) Na2CO3
THF 4
Na-BArF
–
Na+
47
Scheme 24. Activation of [Ir(cod)(PHOX)][BArF] by H2 in THF results in a mixture of diastereomeric octahedral dihydride complexes. (The BArF
– counter-ion has been omitted for clarity)
Rhodium-catalyzed hydrogenation of alkenes with complexes of the type [Rh(cod)(PPh3)2][PF6] indisputably goes through a Rh(I)/Rh(III) cycle (vide supra), so it is sensitive to initially assume that a similar mechanism operates in the iridium case. Indications that this is actually the case has been present-ed by Dietiker and Chen, who studied gas-phase reactions of [Ir(cod)(PHOX)][BArF].76 Using a tandem ESI-MS setup they derived the following data:
1) Upon pumping a pre-reacted solution of pre-catalyst, H2 and styrene in-to the mass spectrometer using electrospray vaporization, masses corre-sponding to [Ir(PHOX)(styrene)(H2)2]+, [Ir(PHOX)(styrene)(H2)]+ and [Ir(PHOX)(styrene)]+ were detected. 2) When instead feeding the MS with [Ir(cod)(PHOX)][BArF], H2 and ethylbenzene, the ion [Ir(PHOX)(EtPh)]+ was detected among other spe-cies. When [Ir(PHOX)(EtPh)]+ was isolated (using MS/MS) and infused into a collision cell fed with argon, [Ir(PHOX)(styrene)]+ was the major species detected. 3) When [Ir(PHOX)(styrene)]+ was collided with D2 gas, ions with mass-es corresponding to [Ir(PHOX)(styrene)(D)]+ and [Ir(PHOX)(styrene)-(D2)]+ was detected along with [Ir(PHOX)(styrene)]+ that had one or two incorporated deuterium atoms. These experiments indicates that: 1) Polyhydride intermediates with iridium in oxidation states +1, +3 and +5 can form. 2) Since the styrene must have formed by dehydrogention and loss of H2, the reaction and by microscopic reversibility, also its elementary steps are reversible in the gas phase. 3) Since only mono- and di-deuterated complexes with styrene could be detected, a mechanism involving an Ir+5 species with more than two hy-drides (one hydrogen molecule), is unlikely.
IrPN
+* IrP H
N S
+*
H
S
IrP HN S
+*
S
H
+
Major Minor
THF
3 H2 Cyclooctane
48
With these results in mind, a catalytic cycle as presented as in Scheme 25, where iridium goes through oxidation states +1 and +3, was indicated for asymmetric hydrogenation of non-functionalized alkenes (cf. Scheme 16).76 After initial formation of A, replacement of a solvent molecule by the alkene results in complex B. The metal-dihydride-alkene complex B then undergoes insertion to form the σ-alkyl complex C and subsequently reductive elimina-tion to give D in which the alkane is very loosely bound (if it is an aromatic compound it may well form a η6-π-system with the benzene ring). The al-kane is easily replaced by solvent molecules and another dihydrogen to re-form A. Since oxidative addition of dihydrogen to square-planar iridium-complexes of this kind is very feasible, the rate-determining step might be the migratory insertion to form C. This is provided that; 1) the alkene coor-dination is fast, and 2) that enough hydrogen is available in solution. The former will naturally depend on steric and electronic properties of the al-kene, and the second on temperature, hydrogen pressure and convection. Pfaltz, Blackmond and others have shown that H2 diffusion can in fact be rate limiting in cases where the alkene is reduced very rapidly.71-72
Scheme 25. Hydrogenation of alkenes by a dihydride Ir(I)/Ir(III) catalytic cycle.
Different mechanistic proposals have been put forth by the groups of An-dersson and Burgess. DFT calculations performed by both groups indicate that the resting state species is actually complex E (Scheme 26) where an additional molecule of dihydrogen has displaced a solvent molecule. The alkene would then coordinate trans to phosphorpous and vertical to the N-Ir-P plane. Calculations by Brandt and Andersson on a truncated system and with ethene as the alkene, revealed that E was transformed to F by concomi-tant migratory insertion and cleavage of the coordinated H2 into two new hydrides (Path A).77 The trihydride Ir(V) species F then quickly undergoes reductive elimination to form G which then re-enters the catalytic cycle.
IrSN P
H
H
S
IrN P
H
H
S
IrN P
H
H
S
IrHN P
HS
S, H2S
A
B
C
D
49
Scheme 26. Hydrogenation of alkenes by tetrahydride Ir(III)/Ir(V) catalytic cycle.
Burgess, Hall and co-workers suggested a very similar mechanism, but in their case, the phosphine in the ligand was replaced by a heterocyclic car-bene.78 Starting also at complex E, one of the hydrogen atoms from the di-hydrogen is transferred to the alkene by metathesis to form intermediate H. Reductive elimination then forms the same complex G as from Path A.
Clearly, these two pathways are very similar, both including the species E and G and both going through an Ir(III)/Ir(V) cycle. The rate determining step in both cases was the first hydrogen transfer to the alkene that happens at the same time as the cleavage of the dihydrogen molecule. Although no experimental evidence for these mechanisms has been presented, other re-ports have also showed that especially path A is probably a very feasible pathway when the ligand is an N,P-donor.79
A few things are worth to point out here: - It would not be unreasonable that the dihydride (I/III) path could be op-erational at low hydrogen concentrations and the tetrahydride (III/V) at higher concentrations of hydrogen.
IrSN P
H
H
S
H2
IrN PH
H
H H
IrN PH
H
H
H
IrHN P
HH
H
IrN P
HH
HH
2 SPath A Path B
A
E
F
G
H
IrN PH
H
HH
IrN P
H
H
H
H
H2
50
- In some cases, when iridium can coordinate to heterofunctions in the al-kene substrate, the dihydride mechanism should operate (Scheme 25, S replaced by heteroatom in the alkene, complexes B–D). Some indications that this is really the case has been presented.80
In order to further probe the mechanistic issue, we designed a simple exper-iment that we hoped would shed some light over which mechanism was op-erational.81 It should be noted that in the dihydride mechanism (Scheme 25) both hydrogens are transferred from the same dihydrogen molecule while in the tetrahydride variant they are coming from two different molecules (Scheme 26) and thus resembles a classical monohydride mechanism.82 Then, if a 1:1 mixture of H2 and D2 were used for the catalytic hydrogena-tion of styrene, a product isotope pattern as presented in Scheme 27 would arise.
Scheme 27. The theoretical reaction outcome of the hydrogenation of styrene under a 1:1 mixture of H2 and D2.
Naturally, kinetic isotope rate differences would perturb this distribution and more importantly, hydrogen-hydride exchange reactions taking place by the metal would form mixed H-D species if the hydrogenation reaction is not
Ph
1 : 1H2 : D2
Dihydridemechanism
Ph
1 : 1H2 : D2
Mono- or Tetra-hydride mechanism
Ph
HH
Ph
DD
50% 50%
Ph
HH Ph
HD Ph
DH
Ph
DD
25% 25% 25% 25%
51
Figure 10. Theoretical mass spectra arising from hydrogenation of styrene with a dihydride or monohydride hydrogenation catalyst.
fast enough. But if we imagine the idealized case, a mass analysis of the product distribution would look as shown in Figure 10. For the experiment we used a common N,P-ligated chiral Crabtree mimic of the type [Ir(cod)(N,P)][BArF] and for comparison, we took help from our Wilkinson friends. The hydrogenation catalyst (RhCl(PPh3)3 operates by a dihydride mechanism, and Rh(H)(CO)(PPh3)3 by a monohydride mechanism. In order to reasonably satisfy all catalysts needs, the experiment was performed in benzene. Thus, three test tubes were charged with solvent, styrene and one of the three catalysts. Since the iridium catalyst is significantly faster, 0.5 mol% was used while 1 mol% was used of the two rhodium catalysts. After purg-ing with argon the reactor was pressurized to 5 bar with a H2 : D2 mixture ~1:1. After stirring for 2 h, the pressure was released and the mixtures ana-lyzed by GC-MS and the mass spectra combined over the entire peak to ob-tain a total average spectrum. The results are presented in Figure 11; the data does not take into account the natural isotope distribution, it is taken directly from the averaged mass spectra. As can be seen, there is a general kinetic isotope effect that results in a lower abundance of the di-deuterated species. But more importantly, there is a clear similarity between the monohydride catalyst Rh(H)(CO)(PPh3)3 and the iridium catalyst [Ir(cod)(N,P)][BArF], supporting the previous DFT results and the perception that catalysts of the type [Ir(cod)(N,P)][BArF] operates through a tetrahydride Ir(III)/Ir(V) cata-lytic cycle. It has to be mentioned that, although unlikely, it is possible to obtain an isotope distribution as the one seen for the iridium catalyst even it operates by a dihydride mechanism provided that the metal-hydride-hydrogen exchange is significantly faster than the hydrogenation.
106 108m/z
Dihydride Mono- or Tetrahydride
Counts
106 108m/z
Counts
107
H,H D,D
H,H
H,D
D,D
52
Figure 11. M+ ions mass analysis of products from the hydrogenation of styrene under H2:D2 1:1.
Of course, a measurement of the iridium catalysts hydrogen exchange rate would settle the question but unfortunately there was no time for such a study.
In addition to the data presented above, I would also like to the isotope separation of 1,2-dideuterio-1-phenylethane and its non-deuterated analogue by GC. Using an Agilent DB-1 column, the isotopes were slightly separated and could probably be fully separated after method optimization (Figure 12).
53
Figure 12. Separation of 1,2-dideutero- and regular ethylbenzene by GC.
It is worth mentioning at this point, in case it was not clear from previous mechanistic discussions, that the asymmetry arising from the N-P-donor system is crucial for success in these systems. The P,P-ligands used for Rh and Ru catalysis can be divided into four major classes:
1) Backbone chirality (DIOP) 2) Phosphorous chirality (DIPAMP) 3) Axial chirality (BINAP) 4) Planar chirality (BPPFA)
In all of these cases, even thou structural and minor electronical differences may exist, there is no strong electronic stimulus for an alkene to coordinate trans to one of the phosphines instead of the other (cf. Figure 8). Thus, it is easy to see how these catalytic systems are not optimal for the hydrogenation of non-functionalized alkenes as the steric demands on a selective catalyst of this type would be much higher than in a ligand bearing two significantly different chelating species.
2.2 Catalyst development 2.2.1 Selectivity in hydrogenation with chiral N,P-ligated iridium complexes While the amount of available ligands for the [Ir(cod)(N,P*)][BArF] system have grown large, most of the published ligands have proven to give high
DD
54
enantioselectivity only for a certain narrow type of alkenes.83 It appears as if the strict sterical requirements put on the ligands to obtain high selectivity also to a large degree prevent their generality. Thus in synthesis, for reduc-tion of a specific alkene, one would have to screen a large array of different ligands to find the best one and the incitement to provide modular and easily prepared ligands is thus strong. The Andersson group has developed two major series of ligands that perform well in the asymmetric hydrogenation of an unusually large set of prochiral trisubstituted alkenes.84 The first system comprises a variable ligand backbone, containing an oxazole, thiazole or imidazole N-donor and a phosphine or phosphinite P-donor (Class 1, Figure 13). This ligand class has proven very useful as a general ligand forming catalysts that are selective for many non-functionalized alkenes.85 The se-cond system, based on a 2-aza-norbornane scaffold, consists of a six-membered iridium chelate, an oxazoline N-donor and an azaphosphinite P-donor (Class 2, Figure 13).86 This ligand backbone has proven effective for asymmetric hydrogenation of several alkene types including enamines,87 enol phosphinates88 and vinyl boronates89 and has been further modified to contain a thiazole N-donor.90 A generalized structure of the Andersson group ligands, highlighting the individual elements is shown in Figure 13.
Figure 13. Generalized ligand structure and the two major ligand classes developed by Andersson and co-workers.
Computational77-78 and to some extent also experimental75 studies indicate that for a generalized N,P-ligand (Figure 13) forming a complex cation [Ir(H)2Z(N,P*)]+ (Z = Solvent or H2 cf. Schemes 25 and 26) upon activation with hydrogen, the primary steric environment sensed by the incoming alkene derives from the group R, which points out towards the alkene which is arriv-ing trans to P. This situation is illustrated in Figure 14; a) shows the gen-
N
X
N
PAr2
RX Y
N
PAr2
R
NX
P
R
ArAr
Ir
BArF
Chiral centre
Backbone
Heterocycle
Steric modifier
Class 2Class 1
55
Figure 14. Rationalized view of the steric environment sensed by an alkene ap-proaching a [Ir(H)2Z(N,P*)]+ complex. Z = Solvent or H2.
eralized structure of the complex [Ir(H)2Z(N,P*)(alkene)]+ where R is the substituent group attached to the heterocycle (Figure 13). With ligands from classes 1 and 2, exemplified here by ligands L6 and L7, differently shaped coordination pockets are formed. This is illustrated in b) where the complex is viewed along the Ir-P-bond i.e. the way that an imaginary alkene is ap-proaching. For L6, the phenyl group on the thiazole ring will point out to-
ZIr
HN PAr2
HR
HN
Z
H
PAr2
i ii
iii iv
HN
Z
H
PAr2
R
i ii
iii iv
R
Class 1 Ligand L6 Class 2 Ligand L7
!!
Z = H2 or Solvent
S
N
PPh2
Ph
N
O
N
P(oTol)2
a)
b)
c)
Activate catalyst-alkene complex
Class 1 Class 2
56
wards the reader and down. For L7 on the other hand, due to the different backbone, the isopropyl group on the oxazoline moiety will point out to-wards the reader but up. The situation can be rationalized as presented under c) i.e. with a simple quadrant model. Again the situation is from the perspec-tive of the incoming alkene. The dark-gray quadrants (iii for class 1 and i for class 2) represent areas that are occupied by the R-groups and the light-gray quadrants areas that are somewhat encumbered by the presence of an aryl group on the phosphorous (ii for class 1 and iv for class 2). The other quadrants does not have any parts of the ligand pointing up towards the in-coming alkene and are thus considered to be completely un-hindered.
For any N,P-ligated complex of this kind, the position of the steric bulk can be determined by measuring the angle (θ) from the N-Ir-P plane up to the center of the R-group as shown in Figure 14 c).79a For ligands of class 1, the angle is negative (–36.3º for L6) and for ligands of class 2 it is positive (+77.5º for L7) indicating that the quadrant accomodating the R-group will be in the lower and upper corner respectively.
The quadrant system in Figure 14 c) can be used to predict the absolute configuration of the products derived from asymmetric hydrogenation of
Figure 15. Selectivity model for prediction of the absolute configuration.
trisubstituted alkenes using [Ir(cod)(N,P*)][BArF] catalysts. Since a trisub-stituted alkene only has one hydrogen substituent, it will be placed in the most crowded quadrant to minimize steric interactions. Since the hydrides are added from “below”, L6 and L7 gives products of opposite absolute con-figuration (Figure 15) upon alkene reduction. This selectivity model has
i ii
iii iv
[Ir(cod)L6][BArF]
R1R
R2
i ii
iii ivR1R
R2
[Ir(cod)L7][BArF]
R1R
R2 R1R
R2
R1
RR2
R1
RR2
57
proven to correctly determine the absolute configuration for almost all sub-strates studied by the Andersson group to date.
2.2.2 Pinene-derived ligands for asymmetric hydrogenation (Paper I) In 2006, Pfaltz and co-workers reported on a new type of tetrahydroquino-line-based ligands for the asymmetric hydrogenation of non- and weakly-functionalized alkenes.91 Using catalysts 4, derived from this ligand set, high activity and enantioselectivity could be obtained in the iridium-catalyzed reduction of a range of substrates (Scheme 28). The resemblance between this ligand type and the successful ligand L6 developed by Andersson is striking. Clearly this is a very privileged structure for catalysts of the type [Ir(cod)(N,P)][BArF].
Scheme 28. Asymmetric hydrogenation of trisubstituted alkenes using tetrahydro-quinoline ligands 4.
We envisioned then, that by using the naturally occurring (+)-α-pinene as a building block, a backbone could be constructed in a two-step synthesis that could allow construction of a range of different similar ligands (Scheme 29). This could potentially be beneficial since a) chiral pool synthesis of diastere-omeric ligands can simplify synthesis and eliminate the need for preparative separation of enantiomers and b) the additional chirality of the pinene struc-ture could serve to tune the catalyst further (cf. Scheme 28). As presented in Paper I,92 the initial target ligands, both diastereomers of L8 and L9 (Scheme 29) were prepared, converted into iridium complexes exo-5, endo-5, exo-6 and endo-6 and tested in the iridium-catalyzed asymmetric hydrogenation of
R1
ArR2
1 atm H2
CH2Cl2, 25 oC, 30 min
R1
ArR2 *
BArF
N RO
Ar2P Ir
4
0.02 4
>90% ee forsix substrates
58
Scheme 29. Concept of pinene-derived tetrahydroquinoline N,P-ligands.
trisubstituted alkenes. Unfortunatly, the synthesis proved to be overall very low yielding and the final iridium-catalysts were often less selective and generally very slow compared to the benchmark structure by Pfaltz.91 Since the screening results are presented and the catalyst synthesis is discussed in some detail in Paper 1, only a few comments on interesting chemistry will be considered here.
Ligand Synthesis For the synthesis of the common backbone structure 8 (Scheme 30) (+)-α-pinene was treated with singlet oxygen (generated by UV-light and tetra-phenylporphyrine, TPP, as photosensitizer) to produce the hydroperoxide 7 by oxygen-ene reaction.93 The hydroperoxide was decomposed in situ using acetic anhyddride and dimethylaminopyridine (DMAP) as catalyst as first reported by Mihelich and Eickhoff94 to give the enone (–)-pinocarvone.
Scheme 30. Synthesis of common backbone structure 8 from (+)-α-pinene.
Kröhnke annulation95 of pinocarvone with acetophenone-pyridine salt and ammonium acetate gave 8 as reported by Malkov and Kocovsky.96 On the
(+)-!-pinene
N RX PAr2
BArF–
N PhX
Ph2P Ir
+
L8 R = Ar = Ph, X = CH2L9 R = Ar = Ph, X = O
5 R = Ar = Ph, X = CH26 R = Ar = Ph, X = O
(+)-!-pinene
N Ph
O2h" O
(–)-pinocarvone
8
Ac2OO OHTPP Pyridine
DMAP7
N+Ph
OI– NH4OAc
HOAc110 oC
59
course towards the phosphinite ligands L9 (Scheme 29, X = O), 8 was affect-ed by mCPBA to give N-oxide 9 (Scheme 31). The diastereomeric mixture 10 was prepared by Boekelheide rearrangement when treating 9 with isobu-tyric anhydride.97 exo-10 could then be obtained pure by recrystallization and column chromatography of the mother liquor yielded endo-10. To obtain L9, 10 was hydrolyzed and phosphorylated as discussed in Paper 1.
Scheme 31. Progress towards L9 and L8.
For the synthesis of phosphine ligands L8 (Scheme 29, X = CH2), 8 was first alkylated in the benzylic position by means of nBuLi and DMF to yield the diastereomeric mixture of aldehydes 11 which was reduced by sodium boro-hydride to 12 (Scheme 31) as reported by Chelucci and co-workers.98 When we could not separate the diastereomeric alcohols 12 and subsequently nei-ther its tosylated derivative 13 (Scheme 32), we had hoped that the BH3-protected phosphines could be separated. Unfortunatly, substitution with BH3Ph2P– at reasonable conditions only gave small amounts of products 14 and when heating the reaction mixture, the elimination product 15 was iso-lated as the only product. We did however recognize that the exo isomer was in general converted to products in a larger extent than the endo. In fact, when studying a molecular model of endo-13, one can see that the two me-thyl groups on the bridge of the pinene moiety may restrict free rotation of the CH2OTs group, making it highly inaccessible for incoming nucleophiles Figure 16.
N Ph
8
N+ PhmCPBA
N PhO
9 10
O–
O(C(O)iPr)2
O
N Ph
O
N Ph
OH
NaBH4
nBuLi DMF
11 12
60
Scheme 32. Reactions of the diastereomeric mixture of the intermediate 6 with dif-ferent nucleophiles.
Substitution thus appeared to be a hopeless cause, but the small, spear-like azide anion was able to substitute the exo isomer and even the endo isomer to a small extent. Somewhat encouraged by this, we attempted the substitu-tion by I– and to our utter enjoyment, exo-13 reacted with a reasonable rate so that resolution was possible. Thus, by reflux in acetone with NaI for 14 h followed by column chromatography, exo-16 was isolated in 40% yield. The iodide turned out to be more reactive towards BH3Ph2P–, giving the BH3-protected ligand in 48% yield, which was considered excellent under the circumstances. The free phosphine ligand was then produced after removing the BH3 protecting group by stirring in diethylamine. HNEt2 is the reagent of choice for this type of deprotections, since its volatility allows direct remov-al under a stream of inert gas, thus limiting phosphine oxidation.
N Ph
OH
12
N Ph
OTsTsCl
Pyridine N Ph
I
NaIAcetonereflux 14 h13
H(BH3)PPh2nBuLi
-78 to 40 oC
N Ph
14 15
NaN3
N Ph
N3
N Ph
PPh2
ClPPh2Li
Low conversion
40% yield
N Ph
PPh2
Very low conversion
> 50 oCH3B
16
61
Figure 16. Nucleophilic attack on the CH2OTs group in endo-13 is unfavourable.
After the preparation and subsequent transformation of compound 16 into the final complex, the task to phosphorylate endo-12 was undertaken. First, the stronger iodination system of I2/PPh3/base was used to completely con-vert all exo-12 into the iodide, which could be removed by chromatography leaving pure endo-12. Since substitution with a nucleophilic phosphine was out of the question, we attempted a method that to our knowledge has not been used to any significant extent in the preparation of chiral phosphines. The alcohol endo-12 was brominated with SOBr2 and the resultant aliphatic bromide converted to the Grignard reagent, which was allowed to attack ClPPh2 at room temperature to form endo-L8 (Scheme 33).
Scheme 33. Preparation of L8 by Grignard addition to diphenylphosphine chloride.
Iridium catalysis Application of the complexes 5 and 6 (Scheme 29) in the asymmetric hydro-genation of a range of different trisubstituted alkenes gave mixed results as presented in Paper I. The complexes derived from endo-ligands generally gave higher conversions, indicating a more open coordination site, while the two phosphinite complexes 6 generally gave lower enantioselectivity which may be due to lower stability. For all complexes except endo-5 the enanti-oselectivity is so low that comparisons between them becomes largely irrel-evant. Surprisingly, the complex endo-5 gave significantly better selectivity for most substrates than the other three complexes with the selectivity reach-ing 97% ee. It should be noted thou that α-pinene with an optical purity of
N Ph
OTs
endo-13
Nu Nu
N Ph
OH
endo-12
N Ph
Br
SOBr2N Ph
PPh2
17 endo-L8
1) Mg2) ClPPh2
62
98% was used, limiting the enantioselectivity obtainable in the catalytic hy-drogenations. Furthermore, it is clear from the absolute configurations of the products that the chirality of in the 8-position (when considering the ligands as quinolines) is the dominating one while the chirality derived from the pinene moiety is secondary and serves as a tuning (or rather de-tuning, when comparing to the results obtained for complex 4) function. This is in agree-ment with the general view that chirality built into the six-membered metal chelate is of crucial importance.
One can also conclude that the results agree with the selectivity model previously developed in the group and discussed above.
2.2.3 Phosphite-oxazolines as extremely modular ligands (Paper II) Another type of ligand that has proven effective for the iridium-catalyzed asymmetric hydrogenation of non-functionalized olefins is 18, shown in Scheme 34.99 This ligand class has the same type of coordinating functionali-ties as discussed earlier for other ligands, i.e. a phosphinite P-donor and an oxazoline N-donor.
Scheme 34. Ligands of the type 18 performs well in the hydrogenation of 1,1-disubstituted alkenes.
In particular, they have shown high selectivity for hydrogenation of 1,1-disubstituted alkenes. The best result obtained with these ligands, 94% enan-tiomeric excess in the hydrogenation of an α-ethyl styrene is with L10 illus-trated in Scheme 34. While these ligands already have a high degree of tune-ability, we envisioned that it could be extended further. By replacing the diarylphosphinite with a phosphite moiety, one could further tune the envi-ronment around the phosphorous atom by for instance, introduction of an entity with axial chirality such as BINOL. In addition, when this project was
1 atm H2
CH2Cl2, 25 oC, 30 min
*0.01 [Ir(cod)L10][BArF]
O
NO PAr2
R2R2 RR1
L10 R = Ph, R1 = Me, R2 = Bn
O O94% ee
18
63
initiated in 2007, only one example of phosphite P-donors in chiral mimics of Crabtree’s catalyst had been presented and given moderate results,100 so we were interested in seeing if it could be a useful approach. The phosphites are, compared to phosphines, still good σ-donors but much stronger π-acceptors. Therefore, phosphite ligands gives a more electron poor iridium as compared to phosphines and phosphinites. Hence, since the rate-limiting step in hydrogenation is the insertion (or oxidative addition and insertion, vide supra), one could be concerned that phosphite ligands may reduce the reaction rate.
An advantage of phosphites in catalyst preparation is the low sensitivity to oxygen as compared to phosphines.101 On the other hand, alkyl phosphines are sensitive to hydrolysis so water and alcohols has to be avoided when handling the free ligand (Figure 17). Upon complexation to iridium the sen-sitivity to hydrolysis drops substantially.
Figure 17. The electron density on phosphorous affected by oxygen alters the prop-erties when going from phosphine to phosphite.
A modular approach to the preparation of phosphite-oxazolines from amino acid esters has previously been reported102 and we used this synthetic proto-col to prepare the catalyst library (24, Scheme 35). Starting from 19, the methyl ester of either L-serine (R1 = H) or L-threonine (R1 = Me), the corre-sponding amides 20 were prepared by alkylation with RCOCl. DAST cy-clization103 yielded oxazolines 21 and two consecutive alkylations with R2MgBr gave the alcohol 22.
Ph P Ph
Ph
Ph P O
PhPh Ph P O
OPh
Ph
O P O
OPh
Ph
Ph
phosphine phosphinite phosphonite phosphite
Sensitivity to oxidation
Sensitivity to hydrolysis
64
Scheme 35. Preparation of phosphite ligand library 24.
Phosphorylation with the appropriate phosphorochloridite (prepared in situ from PCl3 and the corresponding diol) gave phosphite ligands 23 which were purified by filtration through a short plug of Al2O3 under dry conditions. Complexation to iridium with [Ir(cod)Cl]2 and subsequent ion-exchange with NaBArF gave the iridium catalysts 24 (Scheme 35). This method gives cata-lysts with four easily tunable positions and up to three chiral centres in a five-step synthesis, which is efficient with yields in the range 75-95% for each step.
A screening of the ligand library versus a range of trisubstituted alkene substrates proved the advantage of a modular ligand and confirmed the use-fulness of the tunable phosphite (Scheme 36, see Paper II for a more com-prehensive report on screening results). While L11, carrying the (S)-BINOL on phosphorous, was the best ligand for E (trans) alkenes, L12 proved to be most successful in reduction of the more challenging Z (cis) alkenes.
23
NH2HO
COOMe
NHHOCOOMe
O R O
NR
O
O
O
NR
O
R2R2
P O
O
R1 R1 R1
R1
RCOClEt3N
DAST
2 R2MgBr
O
NR
OH
R2R2
R1
Cl P O
O
19 20 21
22
BArF–
O
NR
O
R2R2
PIr
R1
OO+
24
1) [Ir(cod)Cl]22) NaBArF
R = C6H5, 4-CH3-C6H4 or 4-CF3-C6H4R1 = H or CH3R2 = H, CH3 or C6H5
OO
R3
R3
OO
R4
R4
O
O=
R3 = tBu or TMSR4 = tBu, OMe or H
(S) or (R)
or
65
Scheme 36. Evaluation of phosphite-oxazolines as ligands in iridium-catalyzed asymmetric hydrogenation, a few representative results.
It is worth mentioning here, as presented in Paper II allylic alcohos and al-lylic acetates also are reduced selectively and without byproduct formation. It has been suggested that electron rich ligands are better catalysts for these substrates since they generally produce less acidic catalysts and thus dis-courages potential removal of a protonated allylic group. Burgess and co-workers have shown that N,C (C = N-heterocyclic carbene) ligands produce significantly less acidic hydrogenation catalysts than the corresponding phosphinites.104 N-heterocyclic carbenes are stronger σ-donors than phos-phinites, a fragment which in turn is a stronger σ-donor and a much weaker π-acceptor than the phosphites.105 In either case, it is clear that the electron withdrawing phosphite does not seem to increase substrate decomposition for these substrates. Furthermore, even thou no rate measurements were made, it is clear from the short time and mild conditions that the electron withdrawing properties of the phosphite does not have a large impact on the reaction rate.
Ligand L11 also performed very well in the hydrogenation of some 1,1-disubstituted alkenes (Scheme 36). The selectivity is slightly improved com-pared to the results obtained with the original phosphinite backbone99b and the catalyst loading could be decreased even further. These are challenging substrates in terms of selectivity, for instance, ligands L6 and L7 that have
R' 1 atm H2
CH2Cl2, 25 oC, 30 min
R'*0.002 [Ir(cod)L11][BArF]
L11 R = Ph, R1 = H, R2 = MeL12 R = Ph, R1 = H, R2 = Me
R' = Et: 95% eeR' = tBu: >99% ee
O
NR
O
R2R2
P
R1
OO
tBu
tBu
OO
tBu
tBu
O
O=
L11
O
O
O
50 atm H2
CH2Cl2, 25 oC, 2 h
0.01 [Ir(cod)L][BArF]
E (Ligand = L11): > 99% ee (R)Z (Ligand = L12): 92% ee (S)
O
*
L12
66
proven very useful for a large cohort of trisubstituted alkene, failed misera-bly in asymmetric hydrogenation of these 1,1-disubstituted alkenes giving selectivites in the range 30-70% ee.
2.3 Method development 2.3.1 Applications of chiral iridium-N,P complexes in synthesis While the absolute majority of the work with hydrogenation of alkenes using chiral mimics of Crabtree’s catalyst has concerned simple, unfunctionalized alkenes, progress has been made to broaden the scope of these catalyst.84 Lack of functionalization is, after all, not a desirable property in organic synthesis in general. It is probably fair to say, that in asymmetric iridium-catalyzed hydrogenation, there has been a wild hunt for ligands that surpass-es the magic 95% ee line, but in almost all cases, only a limited set of stand-ard alkenes were evaluated (cf. Paper I and II). While researchers have been eager to apply their specific modifications to the ligand backbones, much less focus has been directed towards what the hydrogenation system actually can achieve in synthesis. The groups of Pfaltz, Burgess and Andersson has to some extent addressed this question. Pfaltz group has focused on developing ligands for reduction of the challenging tetrasubstituted and purely alkyl substituted alkenes and applied the catalysis in some total syntheses.106 Bur-gess have focused on studying the more intimate details of the catalytic sys-tem and used it to reduce other systems such as dienes.80,107 In the group of Andersson, the potential of the catalytic system has been tested towards sev-eral types of alkenes. For instance, vinyl fluorides,108 vinyl phosphine ox-ides109 and trisubstituted 1,1-biaryl alkenes110 have been reduced in high enantioselectivity yielding enantiomerically enriched compounds that other-wise would be hard to obtain.
2.3.2 Asymmetric hydrogenation of 1,1-disubstituted alkenes (Paper III) As discussed earlier, several attempts have been made and some progress towards the hydrogenation of 1,1-disubstituted alkenes with catalysts of the type [Ir(cod)(N,P)][BArF] was made both by the group of Pfaltz and by us, but only for a limited set of standard alkenes. A few other systems that can selectively reduce this type of alkenes exist. Some minor success has been seen using Ru-DuPhos as noted in 2.2 (Table 1) but the enantioselectivity did not reach above 90% ee.56b Marks and co-workers used a samarium-catalyst 25 (the CHTMS2 group is removed by reductive elimination when the pre-catalyst is subjected to H2) to reduce 2-phenyl-1-butene with high selectivity (Scheme 37) but the reaction was very temperature dependent, –
67
80 °C was required to achieve high selectivity and rigorous exclusion of air and water was essential.111 Co and Kim have developed an interesting cata-lytic system based on chiral bidentate iminophosphoranes112 26 which they have applied in the asymmetric hydrogenation of some unfunctionalized alkenes, including one 2-phenyl-butene (Scheme 37).113
Scheme 37. Asymmetric hydrogenation of 1,1-disubstituted alkenes.
Their catalytic system is unusual since it works and gives good selectivity using both rhodium and iridium (although Ir gives higher product yields) and contains two identical coordinating groups.
One could argue that reduction of prochiral 1,1-disubstituted alkenes is of limited use since they necessarily gives a product that bears a methyl group. But on the other hand, several examples of chiral methyl-groups appears when scanning through Merck index. Some examples, taken from medicinal chemistry, are shown below in Figure 18. In fact, the Noyori hydrogenation to obtain naproxen is a good example of the use of this methodology.
Sm
1 atm H2
heptane, –80 oC, 2 h
*0.005 25
96% ee
Si
(–)-Menthyl
TMS
TMS Fe
NPPh2
PPh2
N
N
M
iPriPr
iPr
iPr
25 26
10 atm H2
CH2Cl2, 25 oC, 24 h
*0.02 26
M = Ir: 99% yield, 90% eeM = Rh: 39% yield, 97% ee
O O
+
BF4–
68
Figure 18. Examples of CH3-containing tertiary stereocentres taken from medicinal chemistry.
Non-functionalized terminal alkenes present a unique challenge in asymmet-ric hydrogenation. While trisubstituted alkenes rely on a specific alkene con-formation (i.e. E/Z) and essentially requires only a closed “quadrant” in the metal coordination site to provide selectivity, (See section 2.2.1 and Figure 14 and 15) the situation for 1,1-disubstituted alkenes is different. In a com-parison with trisubstituted alkenes as depicted in Figure 19 a), the two lowest energy transition states for the terminal alkene-catalyst complex are going to be of vertically the same energy and give the opposite stereochemical out-come. Instead, for a 1,1-disubstituted alkene, a situation such as in Figure 19 b) is required to obtain enantioselectivity. While no substituent is allowed into the dark-gray quadrant, there must be differentiation between R and R1 in the light-gray one, thus providing only one low-energy transition state.
Figure 19. Selectivity in the hydrogenation of disubstituted terminal alkenes puts larger requirements on the sterics around the active site of the catalyst.
The second problem with terminal alkenes is the risk of alkene isomerization in the cases where thermodynamically favored trisubstituted alkenes can form. Isomerization happens when reductive elimination is slow and β-hydride elimination takes place instead (Scheme 38 a)).41b
CF3
F3C NO
p-F-Ph
AprepitantSubstance P antagonist
R
N
HN
Dexdomitor!2-adreno agonist
O
OHO
NaproxenNSAID
R R1
R2
R R1
No stereodifferentiation
R R1
RR1
Same energy -
or
Only one possibility
Trisubstituted alkene Disubstituted alkenea)
b)RR1
R R1 RR1
69
Scheme 38. The two pathways leading to alkene isomerization.
Reductive elimination is favoured by sterically crowded complexes (since it releases strain) and by electron deficient metals (since it lowers the oxidation state). Another possibility for alkene isomerization arises when the metal is basic enough to absorb a proton from the coordinated alkene to form a η3-π-allyl (Scheme 38 b)).114 This usually happens when the catalyst has not yet added dihydrogen, so the process is suppressed by high hydrogen pressures.115
Scheme 39. Typical outcome when performing alkene hydrogenation with [Ir(cod)(N,X)][BArF] (X = P or C) under D2.
In the asymmetric hydrogenation of alkenes with chiral mimics of Crabtree’s catalyst [Ir(cod)(N,X)][BArF] (X = P or C), both the Andersson and the Bur-gess groups have performed deuterium studies to determine the degree of alkene isomerization.116 An adaption of these results is shown in Scheme 39. In the hydrogenation of 2-phenyl-1-butene, a significant degree of deuterium scrambling to the allylic position is seen. For the bulkier 2-phenyl-3-methyl-1-butene and the trisubstituted trans-2-phenyl-2-butene on the other hand, no isomerization could be observed. The two latter substrates would form more congested π-allyl complexes, thus preventing scrambling.
Given the above discussion, iridium complexes bearing phosphite ligands, being more electron-poor and bulkier than corresponding phosphines and
Ph Ph
HIr
Ph
Ph Ph
Ir H
Ph
Ir
Ir
H
Insertion
H abstraction
!-Elimination
(ox. add.)Red. Elim.
a)
b)
Ph Ph Ph
[Ir(cod)(N,P)][BArF]D2
Ph
D (1.0)
D (0.2–0.4)
D (0.9)Ph
D (1.0)D (1.0)
PhD (1.0)
D (1.0)
[Ir(cod)(N,P)][BArF]D2
[Ir(cod)(N,P)][BArF]D2
70
phosphinites, should have the potential to perform well in the asymmetric hydrogenation of 1,1-disubstituted alkenes. As discussed in 2.2.3 (Paper II),117 phosphite-oxazolines of the type presented does indeed give highly selective N,P-ligands for the iridium-catalyzed hydrogenation. In order to determine the true potential of this catalytic system and be able to present a general method for asymmetric hydrogenation of terminal alkenes we under-took a study using an enlarged ligand set and a large array of 1,1-disubstituted aryl alkenes.
Results Based on our previous study (Paper II) we started out by screening the benchmark alkene 2-phenyl-1-butene against a slightly modified ligand li-brary. We found that L13, where two phenyl-groups are present on the linker together with a phenyl-group on the oxazoline and a (S)-BINOL phosphite was the best ligand (Scheme 40). In a screening of a range of aryl-alkyl al-kenes, enantioselectivities above 90% ee was continuously obtained. The catalyst tolerated severe steric modifications of the alkyl group, heterocyclic aryl-groups (Ar = 2-furyl, 2-thiophenyl and 2-pyridyl) and also hetero-alkyl groups (R = CH2OH, CH2OAc and CH2TMS) gave excellent results.118
Scheme 40. Asymmetric hydrogenation of 1-aryl-1-alkyl-ethenes using L13.
In order to enantioselectively reduce disubstituted biaryl alkenes, some mod-ifications of the ligand structure and reaction conditions were required. In-creasing the catalyst loading to 1 mol% and the hydrogen pressure to 50 bar was necessary in order for the reaction to complete over two hours at room temperature. For the alkenes 1-phenyl-1-(1-naphthyl)-ethene and 1-phenyl-1-(ortho-methyl-phenyl)-ethene (Scheme 41 a) and b)) L14, carrying a 2,6-dimethylphenyl group on the oxazoline and a tetra-tert-butyl-biphenyl phos-phite was required to obtain very high selectivity. Additionally, the extreme-
RAr1 atm H2
CH2Cl2, 25 oC, 2 h RAr *
0.002 [Ir(cod)L13][BArF]
O
NPh
O
PhPh
P O
O
L13
OO
O
O=
16 examples>90% ee
71
ly crowded alkene 1-phenyl-2,6-dimethylphenyl-ethene was subjected to the same conditions but it was not reduced at all.
Scheme 41. Asymmetric hydrogenation of 1-aryl-1-aryl-ethenes.
The alkene in Scheme 41 c) is interesting since the two aryl groups of the alkene has almost identical structure but different electron-attracting power. We wanted to see if it was possible to obtain any selectivity for this type of alkene since selectivity would mean that electronic factors alone can direct the reaction. Indeed, up to 65% ee was obtained using the catalyst [Ir(cod)L15][BArF].
A few years before our study, Börner and co-workers had successfully at-tempted asymmetric hydrogenation of non-coordinating alkenes using cata-lysts of the type [Ir(cod)(N,P)][BArF] and propylene carbonate as a solvent.119 Using propylene carbonate as a solvent did not significantly im-prove the selectivity of the reaction in general, but the possibility to avoid chlorinated solvents is always beneficial. More importantly however, prod-uct alkanes could be extracted from the reaction solution with hydrocarbon
Ph50 atm H2
CH2Cl2, 25 oC, 2 h*
0.01 [Ir(cod)L14][BArF]
O
NAr
O
PhPh
P O
O
L14 Ar = 2,6-di-CH3-C6H3L15 Ar = 4-CF3-C6H4
O
O=
>90% ee
Ph50 atm H2
CH2Cl2, 25 oC, 2 h*
0.01 [Ir(cod)L14][BArF]
99% ee
50 atm H2
CH2Cl2, 25 oC, 2 h
*0.01 [Ir(cod)L15][BArF]
65% ee
Ph
Ph
F3C O F3C O
OO
tBu
tBu
tBu
tBu
a)
b)
c)
72
solvents, leaving the catalyst in the propylene carbonate ready to perform hydrogenation with a new load of alkene substrate.
Since one of the inherent drawbacks with homogeneous hydrogenation is that the catalyst is hard to re-use, we wanted to see if this would also work with the phosphite-oxazoline system without loss of selectivity and activity. Three different 1,1-disubstituted alkenes that gave >99% ee in CH2Cl2 were thus hydrogenated in propylene carbonate using 1 mol% of the appropriate catalyst and 50-100 bar H2. To our delight the selectivity was retained in all cases and up to five repeated extraction/hydrogenation cycles could be per-formed with only 3% loss of enantiomeric excess. The alkene conversion dropped between 10-20 % over the repeated reactions and successively longer reaction times was required but this is to be expected since a fraction of the catalyst will probably deactivate when the alkene is consumed (vide supra). The reason why so excellent results are obtained in the asymmetric hydro-genation of terminal alkenes using the phosphite-oxazoline systems is un-clear. But in view of the discussion about the steric demands and problems with isomerization that are present for this type of substrates one could spec-ulate that the increased bulk around the phosphite P-donor as compared to PAr2 contributes to form a more demanding environment around the active metal centre. Additionally, the weaker electron donating properties of phos-phites may help to avoid isomerization both by speeding up reductive elimi-nation and preventing hydrogen abstraction to form η3-π-allyl intermediates.
2.3.3 Catalytic asymmetric synthesis of chiral azacycles (Paper IV) Chiral nitrogen-containing heterocycles are present in many synthetic and natural compounds and much effort has been devoted to their catalytic enan-tioselective synthesis.120 Asymmetric hydrogenation has been applied; reduc-tion of cyclic imines,121 pyrroles122 and pyridines123 has been demonstrated and also a few examples of the asymmetric hydrogenation of monocyclic enamides and enamines.124
We thought that since chiral mimics of Crabtree’s catalyst are excellent for the hydrogenation of non-functionalized trisubstituted alkenes, they would be able to selectively reduce prochiral cyclic alkenes in which the alkene and the heteroatom are distant from each other (Structure D, Scheme 42). Such alkenes can be prepared with relative ease by a simple coupling reaction followed by ring-closing metathesis (Scheme 42).125 Since asym-metric hydrogenation of cyclic imines, enamines and enamides is dependent on the heteroatom functionality and by necessity puts the chiral center in α- or β-position, this would be a complementary method to build heterocyclic fragments with a distal chirality.
73
Scheme 42. Synthetic strategy to obtain optically active heterocycles.
We prepared a range of five-, six- and seven-membered cyclic alkenes D by this method. Ring-closing metathesis was performed using Grubbs 2nd gen-eration Ru-catalyst 27. This catalyst is superior for this purpose since it, con-trary to the 1:st generation catalyst 28, has high activity in the formation of trisubstituted alkenes.126 27 was thus prepared from 28 by ligand exchange, generating the N-heterocyclic carbene thermolytically (Scheme 43).127
Scheme 43. Preparation of Grubbs 2nd generation metathesis catalyst.
We initially screened our catalyst library against two representative six-membered tosyl-protected dehydropiperidines as reported in Paper IV.128 Based on this screen, the catalyst based on L16 was used to hydrogenate ali-phatic derivatives and the catalyst based on L6 was used to hydrogenate ar-omatic derivatives (Table 2). The reactions were performed with 0.5 mol% catalyst under 50 bar H2 for 15 h. Interestingly, in this case and in fact in a majority of the reports to date, including Paper III, it is found that when re-placing for instance Ph- with F3C-C6H4- on the alkene, i.e. decreasing the electron-density, the conversion and enantioselectivity drops somewhat (Ta-ble 2 entry 4 vs. entries 6 and 8. To my knowledge no rationale has been presented explaining the loss of enantioselectivity but possibly, since elec-tron-poor alkenes does not coordinate as strongly to the metal, it will be less affected by the steric environment and thus reduced with lower selectivity. We found that the more electron-rich ligand L17 gave both higher conversion and enantioselectivity in the asymmetric hydrogenation of electron-poor alkenes (Table 2 entries 7 and 9). The increased electron-density on the met-
NR1
*R
NR1
R
NR1
R
NR1
R
X H
R = Alkyl or ArylR1 = Ts or CbzX = Br, I or OMs
[Ir(cod)(N,P)][BArF]K2CO3 0.04 27
A B C D E
RuPCy3
NN
PhClCl
MesMes
27
RuPCy3
NN
PhClCl
MesMes
C6F5PCy3
Toluene, 90 oC, 3h
28
74
al imposed by this ligand probably facilitates back-bonding to the electron-poor alkenes, thus strengthening the M-alkene bond and facilitating migrato-ry insertion.
Table 2. Asymmetric hydrogenation of dehydropiperidines.
Methyl- and phenyl-substituted five- and seven-membered alkenes were also studied. While the methyl-substituted variants were reduced with moderate enantioselectivity, high enantioselectivty was obtained for the phenyl-substituted alkenes (Scheme 44). The results show that for the seven-membered alkenes, the selectivity is not really affected by the position of the alkene, probably because the substrate is flexible and can bend away from the catalyst. One could interpret this as if all three seven membered alkenes behaves as trans-2-phenyl-2-butene. For the much more rigid five-membered alkene the situation is reversed and even the nitrogen-substituent affects the selectivity. Changing from tosyl to Cbz results in an increased enantioselectivity but lower conversion.
N
O
N
P(oTol)2
N
R
Ts
!!
N
R
Ts
CH2Cl2 RT, 15 h
N
N
PPh2
PhS
N
PPh2
Ph
0.005 [Ir(cod)(L)][BArF]
Conv. % ee %
Me
Bn
CH2OH
4-MeO-C6H5
Ligand
100
100
97
97 (–)
97 (–)
92 (–)
C6H5 100 >99 (+)
100 99 (+)
92 98 (–)
74 96 (–)
50 bar H2
4-Br-C6H5
4-F3C-C6H5
L16
L6
L17
L17
R
4-Br-C6H5
4-F3C-C6H5
L16
L16
L6
L6
L6
L16 L6 L17
68 94 (+)
19 87 (+)
Entry
1
2
3
4
5
6
7
8
9
75
Scheme 44. Asymmetric hydrogenation of five- and seven-membered cyclic alkenes using L18.
2.3.4 Asymmetric hydrogenation of cyclic alkenes (Paper V) With a successful method for the asymmetric hydrogenation of N-heterocyclic alkenes as presented in Paper IV it was a natural extension to look further into other prochiral cyclic alkenes. We wanted to see 1) what other 3,4-alkenes (A and B) that could be reduced with this method and 2) if it could be used for 2,3-alkenes (C and D) as well (Figure 20).
Figure 20. Generalized structures for the substrates we wanted to explore.
The results of the study are discussed in Paper V and will not be present-ed to any significant extent here. Instead a few interesting observations will be considered more in detail. A very rough presentation of the results is
NTs
R1
conv (%) ee (%)
NTs
R1
>99
96 (-)
98 (-)
NX
Ph
NX
Ph78 99 (+)
R2 R2
R1= Ph R2= H
R1= H R2= Ph >99
X = Ts
X = Cbz
>99 85 (+)
conv (%) ee (%)
50 bar H2, CH2Cl2r.t. 15 h
NTs
conv (%) ee (%)NTs
>99 90 (+)
Ph Ph
S
NN
PPh2
Ph
L18
[Ir(cod)L18][BArF]
50 bar H2, CH2Cl2r.t. 15 h
[Ir(cod)L18][BArF]
50 bar H2, CH2Cl2r.t. 15 h
[Ir(cod)L18][BArF]
X
R
X
R
X
R
X
R
3,4-alkenes 2,3-alkenes
A B C D
76
shown in Figure 21 with the aim to display general trends.129 With the divi-sion into classes A, B, C and D (Figure 20) and further dividing C and D into electron-rich (C’ and D’) and electron-poor (C’’ and D’’) 2,3-alkenes one can see some general trends.
Figure 21. Trends in the reaction outcome of asymmetric hydrogenation of cyclic alkenes with catalysts of the type [Ir(cod)(N,P)][BArF].
First of all, it is clear that alkenes lacking adjacent functionality, A and B, in general are reduced faster. This is perhaps not surprising since chiral ana-logues of Crabtree’s catalyst works best for non-functionalized alkenes. Electron-poor (C’’ and D’’) and electron-rich (C’ and D’) alkenes are both reduced more slowly, but probably for different reasons. While electron-poor alkenes may coordinate poorly, electron rich alkenes are not as susceptible to accept the nucleophilic hydride and the adjacent free electron pair may dis-rupt alkene coordination. The Ir-catalyzed hydrogenation of enamines and enols frequently gives low enantioselectivities87 and here similar trends are observed since C’ alkenes are reduced in 40-60% ee. Cyclic ketones and lactones are reduced with high enantioselectivity albeit slowly. In general, five-membered cyclic alkenes are reduced with lower selectivity, possibly due to the size and additional strain. It should be pointed out that the entropic gain from reduction of cyclic alkenes is lower than for acyclic ones, and this is probably one of the reasons for the comparably low turnover frequencies.
N
R
N
R
N
R
O
R
O
R
R
R
O
R
O
R
R
R R
E EO
OR
O
R
OO
NR
ee
Conv.
Ts Ts
Ts
TsA
B
C'
D'
C''D''
77
Asymmetric hydrogenation of cyclic ketones
For the asymmetric hydrogenation of cyclic ketones, i.e. quite electron defi-cient alkenes, the electron-rich ligand L19 gave highest conversion and enan-tioselectivity. This again shows the same trend that electron-rich ligands performs best for electron-poor alkenes as discussed in paper IV. While six-membered cyclic ketones could be reduced completely and in high enanti-oselectivity provided that the reaction was run for 24 h, five-membered vari-ants were not reduced to any significant degree under these conditions (Scheme 45). The same trend was observed for lactones although not as prominent (Paper V).
Scheme 45. Asymmetric hydrogenation of a five- and six-membered cyclic ketone using [Ir(cod)L19][BArF] as catalyst.
As mentioned in the beginning of this chapter, metal hydrides can be a source of significant acidity, and this was indicated during these investiga-tions. We found that when performing the hydrogenation of 3-phenylcyclohex-2-enone in dichloromethane, significant amounts of the carbonyl-reduction product was also formed (Scheme 46 a)). This problem was solved by exchanging the dichloromethane for 2,2,2-trifluoroethanol, a weakly coordinating alcoholic solvent that could serve as a proton-acceptor, giving the pure alkene hydrogenation product in 94% ee (Scheme 46 b)). For the butyl and methyl derivatives, this effect was essentially not seen during the alkene hydrogenation in CH2Cl2, but the reaction conditions had to be controlled carefully since carbonyl reduction was seen if leaving the reaction running for too long after the alkene had been fully consumed. This implies that the catalyst is directly involved in the carbonyl-reduction, as expected,
N
N
P(o-Tol)2
Ph
L19
50 bar H2, CH2Cl2r.t. 24 h
[Ir(cod)L19][BArF]
O O
*
O
No reaction
50 bar H2, CH2Cl2r.t. 24 h
[Ir(cod)L19][BArF]>99% conv. 92% ee
78
and that the reason for the competition in the case of the phenyl derivative is probably the unfeasible coordination/reduction of the conjugated phenyl-substituted alkene.
Scheme 46. Reactions of 3-phenylcyclohex-2-enone with [Ir(cod)L19][BArF] under H2 and different conditions.
Curious to see if the secondary reaction could be suppressed in dichloro-methane, hydrogenation was performed in the presence of 1.5 eq. of ethanol or trifluoroethanol. Interestingly, a ~50:50 mixture of the corresponding ether diastereomers was obtained (Scheme 46 c)). This reaction outcome, a reductive alkylation, was unexpected and to further probe this phenomenon, the corresponding racemic ketone and allylic alcohol was prepared and sub-jected to the reaction conditions (Scheme 47). Surprisingly, when the satu-rated ketone was used as the substrate, only the cyclohexanol product was obtained but when the allylic alcohol was used, a mixture of alcohol and ether was obtained. The reductive alkylation of ketones have been performed over heterogeneous Pd catalysts by the action of H2 and alcohols,130 but to my knowledge not by homogeneous metal catalysts and not coupled with asymmetric alkene hydrogenation.
O
Ph
O
Ph*100 bar H2, CH2Cl2
r.t. 24 h
0.01 [Ir(cod)L19][BArF]+
OH
Ph*
*
O
Ph100 bar H2, CH2Cl2
r.t. 24 h
0.01 [Ir(cod)L19][BArF]OR
Ph*
*
1.5 ROH
75% 25%
cis:trans approx 1:1R = Et or CH2CF3
a)
b)
O
Ph
O
Ph*100 bar H2, F3CCH2OH
r.t. 24 h
0.01 [Ir(cod)L19][BArF]
c)
Only product94% ee
79
Scheme 47. Reaction of pre-reduced compounds with [Ir(cod)L19][BArF] under H2.
It is hard to draw conclusions from these few experiments. Probably, the amount of water in the reaction is pivotal for the outcome and since it was not performed under anhydrous conditions, the water content may fluctuate. Two similar mechanistic suggestions are presented in Scheme 48, both are acid-catalyzed and starts with hydrogenation of the double bond and release of the ketone. The fact that the diastereoselectivity is low indicates that the substrate is released from the catalyst before the reduction of the C=O func-tion. The virtually non-existing cis:trans selectivity also points towards an SN1 process, perhaps the secondary carbocation formed from dehydration of cyclohexanol is attacked by ROH (Scheme 48, violet). Another option is that the oxonium ion is reduced by an iridium-hydride (blue).
O
Ph100 bar H2, CH2Cl2
r.t. 24 h
0.01 [Ir(cod)L19][BArF]OH
Ph
a)
b)
OH
Ph
OH
Phr.t. 24 h
0.01 [Ir(cod)L19][BArF]1.5 EtOH
1.5 EtOH
100 bar H2, CH2Cl2
1:1 cis:transrac.
rac.
OEt
Ph
+
50%1:1 cis:trans
50%1:1 cis:trans
80
Scheme 48. Possible mechanisms for the iridium-catalyzed reductive alkylation.
Even thou this reductive alkylation reaction is interesting, the low diastere-oselectivity limits its usefulness and the reaction was not studied further at this point.
Synthesis of (S),(S)-morphan The ability of [Ir(cod)L19][BArF] to enantioselectively hydrogenate α,β-unsaturated ketones and the high tolerance of the catalytic system towards functional groups inspired us to use this methodology in the preparation of elusive synthetic building blocks. We envisioned that the bicyclic alkaloid morphan,131 the heart in the opiate structure, could be prepared using the corresponding aminoketone and its prochiral α,β-unsaturated counterpart as key intermediates (Scheme 49).
O
Ph
O
Ph
A
ROH
O
Ph
A H
C+
Ph
H OR
Ph
ROH
Ph
ORO
Ph
O+RA
H
[Ir], H–
1) [Ir], H2
R = H or EtA = Lewis ([Ir]) or Brönstedt (H+) acid
2) A
–HOA
–HOA
[Ir], H2[Ir], H2
Hydrogenolysis
81
Scheme 49. Retrosynthetic analysis for employment of asymmetric hydrogenation in the preparation of (S),(S)-morphan.
Optically active morphan has been prepared previously by a synthesis apply-ing chiral axiliaries, the preparation was however lengthy and low yielding.132 Since Crabtree-type catalysts are retarded by the presence of amines,19 we needed to prepare the α,β-unsaturated ketone with a protected ethylamine substituent. This was achieved in a two-step synthesis starting from commercially available 3-methoxy-phenethylamine 29 (Scheme 50). Birch reduction, N-protection and enol hydrolysis afforded 30 in 77% yield. Trifluoroacetate was chosen as amine protecting group since it is stable to dilute acid and catalytic hydrogenation but easily removed by basic hydroly-sis.
Scheme 50. Synthesis of the N-protected α,β-unsaturated ketone substrate 30.
Asymmetric hydrogenation of 30 using 1 mol% [Ir(cod)L19][BArF] and 100 bar H2 for 24 hours gave the protected aminoketone 31 in quantitative yield and 90% ee (Scheme 51). Deprotection of the amine with base followed by in situ imine reduction with sodium cyanoborohydride gave (S),(S)-morphan which is volatile and therefore directly protected as the N-Boc amide.132 Unfortunately, the yield going from 31 to 33 was low and attempts to scale up the reaction gave even poorer yields, possibly due to oligomerizations at the unprotected ketoamine stage. Increasing the dilution only slightly im-proved the situation.
HNO
H2N
O
H2N
(S),(S)-morphan
OMeO
H2N HNCOCF3
MeO
H2N1) TFAA2) H+/H2O
Na/NH3EtOHNH3 (l)
29 30
82
Scheme 51. Preparation of N-Boc morphan in 90% ee from 30 using asymmetric hydrogenation.
Experimental data for compounds 30 and 31 Compound 30:
Enol ether (1.5 g, 9.8 mmol, 1.0 eq.), obtained from 29 by birch reduction according to previously reported procedures,133 was dissolved in methanol (10 ml) and triethyl amine (2.8 ml, 19.6 mmol, 2.0 eq.) was added. The solu-tion was cooled to 0 °C and ethyl trifluoroacetate (1.51 ml, 12.7 mmol, 1.3 eq.) was added. The reaction mixture was allowed to reach room temperature over night where after the reaction was complete according to TLC. After cooling to 0 °C, the solution acidity was adjusted by addition of 1M aq. HCl until pH = 1 and the reaction stirred for an hour at room temperature. Water and Et2O was added and the phases separated. The aqueous phase was ex-tracted twice with Et2O and the combined organic phases washed with water and brine and dried over Na2SO4. Evaporation of the solvent yielded crude material which was purified by column chromatography (pentane:EtOAc 1:1 Rf = 0.26) to give pure 30 as a white solid in 77% yield. White solid, mp = 63.9-64.8 °C. 1H NMR - (CDCl3, 500 MHz): δ 6.99 (br s, 1H), 5.87-5.84 (m, 1H), 3.61-3.56 (m, 2H), 2.54-2.48 (m, 2H), 2.39-2.31 (m, 4H), 2.04-1.97 (m, 2H) 13C NMR - (CDCl3, 100 MHz): δ 199.3, 161.4, 157.6 (q, J= 37.5 Hz), 127.5, 115.9 (q, J= 287.9 Hz), 37.4, 37.34, 37.29, 29.6, 22.7 19F NMR - (CDCl3, 376 MHz): δ –76.4 IR - (neat, cm-1): 3300, 1700, 1673, 1557, 1176, 1149 HRMS - (CI, methane) m/z = 236.0902, calcd. For C10H13F3NO2 [M + H+]: 236.0893
0.01 [Ir(cod)L19]+[BArF]-
100 bar H2 O
HNHN
(S),(S)-morphan90% ee
COCF3
1) K2CO3, H2O2) Na(CN)BH3
N-Boc-morphan
CH2Cl2RT, 24 h
Boc2O31 32
33
O
HNCOCF3
30
83
Compound 31: Racemic 31 was prepared by catalytic hydrogenation of 30 using [Ir(COD)PyPCy3]+[BArF]–. See Supporting Information for Paper V for de-tailed hydrogenation procedure (reduction method B). The enantiomers were separated by chiral HPLC equipped with a diode array detector. (Chiralcel AS-H, Hexane:iPrOH 90:10 0.5 ml/min) The crude material was purified by column chromatography (pentane:EtOAc 1:1, Rf = 0.41) to give 31 as a colourless oil (quantitative yield). 1H NMR - (CDCl3, 500 MHz): δ 7.29-7.18 (bs, 1H, NH), 3.42-3.30 (m, 2H), 2.41-2.20 (m, 3H), 2.09-1.98 (m, 2H), 1.96-1.91 (m, 1H), 1.89-1.79 (m, 1H), 1.69-1.51 (m, 3H), 1.40-1.32 (m, 1H) 13C NMR - (CDCl3, 100 MHz): δ 211.6, 157.5 (q, J= 37.6 Hz), 115.8 (q, J= 288.2 Hz), 47.6, 41.3, 37.5, 36.4, 35.3, 30.8, 25.0 19F NMR - (CDCl3, 376 MHz): δ –76.3 IR - (neat, cm-1): 3311, 2939, 1699, 1555, 1448, 1153 HRMS - (CI, methane) m/z = 237.0973, calcd. For C10H14F3NO2 [M+]: 237.0977
84
2.4 Conclusion
Two new chiral N,P ligand classes were developed for use in the iridium-catalyzed asymmetric hydrogenation. The ligands were bound to iridium to form complexes of the type [Ir(cod)(N,P)][BArF], which were tested in the enantioselective reduction of weakly functionalized alkenes.
The first ligand type contained a tetrahydroquinoline backbone and was based on a pinene-moiety. The resulting catalysts displayed moderate activi-ty and selectivity, which was partly attributed to the crowded ligand struc-ture.
The second ligand type were oxazolines carrying different biaryl-phosphite P-donors. These ligands were prepared in good yield from amino acids and their iridium-complexes were highly active hydrogenation cata-lysts that, notably, reduced 1,1-disubstituted in excellent enantioselectivity. The success of these ligands showed the advantages of highly modular lig-and scaffolds, and that phosphites are viable P-donors in asymmetric hydro-genation using chiral mimics of Crabtree’s catalyst.
Further modifications of the phosphite-oxazoline ligands allowed im-proved selectivity in the hydrogenation of terminal alkenes and, again, demonstrated the worth of a flexible ligand synthesis. A wide range of aryl-alkyl and aryl-aryl substituted alkenes were enantioselectively reduced using this catalytic system thus providing an efficient route to chiral 1,1-disubstituted ethanes.
Iridium catalysis was applied to the asymmetric hydrogenation of cyclic alkenes. By combining ring-closing metathesis with asymmetric hydrogena-tion, a range of chiral pyrrolidines, piperidines and azepanes could be syn-thesized via a short catalytic synthesis route and in high enantioselectivity. The method was extended to include other hetero- and carbocyclic substrates and high enantioselectivities were obtained for six-membered alkenes. In general, cyclic alkenes devoid of adjacent heteroatoms were reduced faster and in higher enentioselectivities, illustrating the preference of Crabtree-type catalysts for non-functionalized substrates.
85
3 Renewable resources
3.1 Metal catalysis for renewable feedstocks As discussed in Section 1.5, based on its abundance and low production costs, polysaccharides are the most promising sources of renewable carbon materials. Lignocellulose, which is a structural material composed mainly of lignin, cellulose and hemicellulose, is the major component of most plant materials and the most abundant type of biomass on Earth. Cellulose and hemicellulose, the polysaccharide parts of the material, make up approxi-mately 70–85% of lignocellulose. Lignin is a phenolic sheet-like polymer that serves as a water-proof support structure to the polysaccharide core in plant materials.134 It is chemically very different from the polysaccharides, and may serve as a future renewable source of aromatic compounds. Hemi-cellulose is a branched polymer made up of several different monosaccha-rides and its exact content and structure varies between biomass types (Fig-ure 22 b)). Cellulose, on the other hand is a linear homopolymer of D-glucose dimers connected via β-(1→4) glycoside linkages (Figure 22 a)).
Before any of these three polymeric materials can be used as sources of small building blocks for organic synthesis, they have to be decomposed into smaller molecules. Depolymerization of polysaccharides by hydrolysis for example results in sugars that can be transformed into many different useful compounds.135 In order for these sugars to serve as bulk resources for chemi-cal manufacturing; however, they must be drastically modified. Sugars have an O/C ratio close to 1.0, whereas the fossil fuels we use today have almost no oxygen content at all. Furthermore, the hydroxyl groups in sugars makes them very hydrophilic, while the hydrocarbons of oil are very hydrophobic. Thus, to supply the chemical industry with starting materials, the sugars have to be deoxygenated and transformed into compounds that are more easily handled using the available methodology and infrastructure.
86
Figure 22. Structure and momomer composition of a) cellulose and b) hemicellu-lose.
3.1.1 Hydrolysis of polysaccharides An enormous amount of effort has been put into converting polysaccharides into useful compounds.136 These processes were researched intensely during the first half of the 20th century, and especially during the two World Wars, when nations strove for self-sufficiency in terms of fuels, commodity chemi-cals and animal fodder.137 The most obvious approach to obtain small organ-ics from polysaccharides is the depolymerization to monomeric sugar units. Hemicellulose, due to its branched and amorphous nature, is relatively easy to degrade and can be removed from the more robust cellulose by mild acid treatments.138 Cellulose, which constitutes the majority of the material, on the other hand forms a largely crystalline structure held together by a net-work of hydrogen bonds, and is extraordinarily robust.139
Traditional chemical methods for cellulose hydrolysis Traditionally, a few different processes, all of them catalytic, have been used to degrade cellulose, usually with the ultimate aim of fermenting the result-
OHOHO
OHOH
OH
Cellulose
OOHO OH
OH
OOHO
OH
OH
n
!-(1 4)-glycoside linkage
OAc
Hemicellulose
OHOHO
OHOH
OHO
OH
HOOH
OHOHO
HOOH
OH
OHO
HOHO
OH
OHOHO
HOOH
CO2H
OH
Arabinose
Galactose Glucoronic acid
Mannose
Xylose
OHOHO
OHOH
OH
Glucose
Glucose
a)
b)
87
ing glucose into ethanol. It should be noted that the hydrolysis of the β-(1→4)-glycosidic bond in itself is not challenging; rather the problem is to swell, solubilize and break up the cellulose crystallites. The methods dis-cussed below uses a variety of strategies to address this problem, including dissolution in highly ionic media, surface peeling reactions, and mechanical degradation of the inert crystalline material.
- Concentrated acid hydrolysis: the first cellulose-hydrolysis process, dis-covered by Braconnot140 in 1819, utilizes concentrated Brønstedt acids to solubilize cellulose. The formation of sulfate esters (when sulfuric acid is used) and the extreme ion concentration break up the hydrogen bonds of cellulose and dissolve the material.141 Diluting the solution with water and then heating then completes the hydrolysis and typically gives high yields of glucose (Scheme 52).142 This process has been performed with sulfuric-,143 hydrofluoric-,144 hydrochloric-,145 trifluoroacetic-146 and formic147 acids and probably several others, with various degree of success. In general, this should be regarded as a mild hydrolysis method because it is not necessary to heat the solution above 50 ºC, but the generation of enormous amounts of acidic waste, complicated acid recovery, and highly corrosive nature of these acids has prevented the large-scale applications of concentrated acid hydrol-ysis.
Scheme 52. The acid-catalyzed hydrolysis of cellulose.
- Dilute acid hydrolysis: This process, which dates back to 1898, uses ∼0.5 wt% of a strong aqueous acid to hydrolyze polysaccharides.148 Dilute aque-ous acid alone is insufficient to disrupt the cellulose crystallites, so the sys-tem must be heated to 150–250 ºC to obtain significant glucose yields from
OOHO OH
OH
OOHO OH
OHH+
H2O
n
-H+
2n
Cellulose
Glucose
H+
2n H2O
OOHO OH
OH
OOHO OH
OH
n
H
O+OHO OH
OH
OOHO OH
OH
n
H
OOHO OH
OH
OHOHO OH
OH
nOH
OHOHO OH
OH
OH
+
88
cellulosic material. However, the yield of glucose from dilute acid hydroly-sis in batch reactors rarely exceeds around 50% because it is quickly degrad-ed at these temperatures.149
However, using flow reactors and stepwise hydrolysis, which minimizes the contact time between the desired product and the hot acids, the glucose yield from crystalline cellulose can be pushed up to 60–70%.150 Despite the problems with glucose degradation, dilute acid hydrolysis is the cellulose hydrolysis method that has seen most commercialization, with several plants operational during the 1950s. Today, however, this process is not used commercially since other, cheaper, methods are available for the production of glucose and ethanol.
Under acidic conditions and heat, glucose typically degrades by dehydra-tion to 5-hydroxymethylfurfural (HMF) and its hydrolysis products levulinic acid and formic acid (Scheme 53).151 Much interest has recently been fo-cused on the selective formation of HMF or levulinic acid from polysaccha-rides.152
Scheme 53. Byproduct formation during the acid-catalyzed hydrolysis of cellulose. All steps are acid-catalyzed.
- Alkaline hydrolysis: This is another old process for cellulose degradation.153 Unlike acidic hydrolysis, which involves random scissions of the solubilized polymer, basic conditions usually results in endwise peeling of the cellulose chain (Scheme 54).154 Under these conditions (T = 100–250 ºC depending on [OH–]), the product is not glucose but rather organic acids that are formed from successive eliminations and rearrangements.155 As in the case of dilute acid hydrolysis, however, the selectivity is poor and the extraction of products from the reaction medium is problematic. The cellu-
OHO
HO OHOH
OH
Glucose
OOHO
Cellulose
OligosaccharidesHumins
and other degradation
products
HMF
O
OH
Levulinic acid Formic acid
2 H2O
- 3 H2O
+
OOH
O
89
lose conversion also tends to be limited because isomerizations of the chain ends render the polymer chains inert to further peeling.
Scheme 54. Base-catalyzed hydrolysis of cellulose (by peeling from the reducing end of the polymer chain).
- Hydrothermal hydrolysis: This reaction relies on the increased self-ionization of pure water (i.e. the dissociation into H+ and OH–) at high tem-peratures (>200 ºC) and is thus both acidic and alkaline in nature.156 Under these conditions, a broad spectrum of compounds are formed from glucose decomposition; these include organic acids, and the pH of the resulting solu-tion drops during the reaction to ∼3–4 over the course of the reaction.157 Products typical of both acid- and base-catalyzed hydrolysis are observed, even when starting from pH-biased solutions,158 and the process thus suffers from similar problems of glucose instability as acid hydrolysis.
More recent technologies Both the dilute acid and hydrothermal hydrolysis processes give complex product mixtures due to degradation of the formed monosaccharides at the reaction temperatures. However, these high temperatures are required to assist the swelling and dissolution of the recalcitrant crystalline cellulose. As discussed above, concentrated acids can completely solubilize cellulose at low temperature, but their use is associated with problems with waste, corro-sion and the difficulty of reusing the solvent. In addition to concentrated acids, some molecular and ionic solvents can dissolve cellulose. N-methylmorpholine N-oxide is a useful molecular solvent for cellulose,159
CH2OH
O
OR
HOH
HO
HO
Reducing end of cellulose
R = cellulose polymer chain
(anhydro- glucose)n
CH2OH
OR
OH
OHO
HO
CH2OH
CH2OH-OOC
HO
CH2OH
OHHO
HO
HO
OH-
O
OH-
CH2OH
OH
OO
HO
+
H
OR'
OH-
Reducing end of cellulose
R' = cellulose polymer chain
(anhydro- glucose)n-1
90
mixtures of dimethylacetamide and lithium chloride are effective solvents160 and aqueous solutions of metal complexes have shown similar properties.161 In recent years, ionic liquids have emerged as a powerful (but expensive) way of solubilizing cellulose,162 and cellulose dissolution in ionic liquids can be coupled with acid-catalyzed hydrolysis to produce glucose.163 Whether in water, acid or ionic liquids, a common problem remains; it is virtually im-possible to extract monomeric sugars from dilute reaction solutions in these solvents, at least on a preparative scale. In situ conversion of the sugars to a more hydrophobic compounds allows their extraction with organic solvents, as has been elegantly demonstrated by several groups for the glucose dehy-dration product, HMF (see Scheme 53, above).164
The heterogeneous acid hydrolysis of cellulose has been studied inten-sively over the past 10 years, especially in combination with dissolution in ionic liquids.165 Naturally, the possibility of easily recovering the acid is attractive, and with the developments of zeolites, ionic resins and other solid acids, this is an interesting option. In many cases, the solid-acids used for these reactions are not truly heterogeneous; rather, protons leached from the solid into solution are responsible for the catalysis.166
3.1.2 Formation of polyols from polysaccharides Monosaccharides can be converted into sugar alcohols by metal-catalyzed hydrogenation.167 For instance, D-sorbitol is produced industrially by hetero-geneous nickel-catalyzed hydrogenation of aqueous D-glucose at 100 bar H2 at 150 ºC (Scheme 55).168 The platinum group metals typically show higher catalytic activity in this reaction and are also more resistant to deactivation and leaching.169
Scheme 55. Preparation of D-sorbitol from D-glucose by nickel-catalyzed hydrogena-tion.
The idea that sugars derived from polysaccharide hydrolysis could be re-duced in situ to form polyols was first presented by Balandin during the 1950s.170 This reaction is useful since polyols are more stable at high tem-peratures and under acidic conditions than sugars are. This approach has been especially successful for xylose and other hemicellulose hydrolysis
HO
HOOH
HOOH
D-glucose
HOOH
HOOH
D-sorbitolHO HO
OHRaney-Nickel
H2O, 150 oC
100 atm H2
91
products that are released early during lignocellulose hydrolysis; trapping them as xylitol and related compounds prevents their degradation.171 Polyols such as sorbitol and xylitol are frequently identified as future platform chem-icals since they can be produced from biomass and can be converted into several other useful starting materials for chemical industry.172 Degradation of cellulose to sorbitol under acidic and reducing conditions is by no means selective thou. In addition to sorbitol formed from direct hydrogenation of glucose, acid-catalyzed epimerization (Lobry de Bruyn-Alberda van Eken-stein transformation)173 of glucose followed by hydrogenation gives manni-tol (Scheme 56).
Scheme 56. In situ reduction of glucose under high-temperature and acidic condi-tions yields sorbitol, but also mannitol due to epimerization.
The polyols formed are not inert to decomposition, either. Under acidic, high-temperature conditions, the dehydration products sorbitan and iso-sorbide are usually formed (Scheme 57).174
CelluloseOHO
HOOH
OH
OH
Dilute acidHeat
HO
HOOH
HOOH
D-glucose
HO
HOOH
HOOH
D-mannose
OHO
OHHO
OH
HOOH
HOOH
HOOH
HOOH
D-sorbitol D-mannitol
H+ H+
H2, Metal cat.Heat
glucose
Major Minor
Heat Heat
HO HO HO
HOHO
OH OH
D-fructose
92
Scheme 57. The acid-catalyzed dehydration of sorbitol gives mixtures of sorbitan and isosorbide.
In 2006, the hydrolysis-hydrogenation of cellulose was reinvented by Fuku-oka and Dhepe who, instead of using homogeneous acid, used transition metals deposited on solid supports as catalyst together with hydrogen gas and high temperatures to degrade cellulose.175 In presence of H2, the glucose formed during hydrolysis was reduced to sorbitol and, because no homoge-neous acid was used, the formation of sorbitol degradation products was limited. Platinum and ruthenium were the best metals and the highest hexitol yield (31 mol%), was obtained with platinum on γ-alumina (2.5 wt% Pt) (Scheme 58).
Scheme 58. Production of hexitols from cellulose under hydrogen catalyzed by Pt/γ-Al2O3.
In addition to γ-alumina, several other support materials such as a protic Y zeolites (HUSY) and SiO2 were tested and it was demonstrated that the sup-port material has a strong influence on the degree of cellulose hydrolysis. The authors speculated that the metal and support material cooperated to form protons that catalyzed cellulose hydrolysis. When the reaction was performed without metal (i.e. with only support material), no sorbitol and only small amounts of glucose (1–3 mol%) was detected. Hattori and co-workers have presented a mechanism by which dihydrogen could produce protons in this heterolytic fashion (Figure 23).176
OH
OH
OH
OH
sorbitolOH
OH
O
OH
OH
OHHO
sorbitan isosorbideO
OHO
OH
H+Heat
H+Heat
-H2O -H2O
HOOH
HOOH
D-sorbitolHO
OHPt/!-Al2O3
H2O, 190 oC, 24 h
50 atm H2MicrocrystallineCellulose
HOOH
HOOH
HO
OH
D-mannitol
25% yield 6% yield
93
Figure 23. The proposed mechanism for the formation of protons by action of sup-ported metals on dihydrogen. Adapted from Hattori et al.176
Studies on the hydrolysis-hydrogenation of cellulose to polyols over supported metal catalysts A closer examination of the hydrolysis/hydrogenation of cellulose by Pt/γ-Al2O3 (Scheme 58) by Rataboul, Essayem and co-workers highlighted some important aspects of the reaction.177 For instance, by simply stirring the mi-crocrystalline cellulose in water under 50 bar H2 at 190 °C for 24 h, 45 wt% of the cellulose was solubilized, signaling the contribution of hot water treatment alone. With γ-alumina present in the reaction, the cellulose conver-sion increased to 57 wt%, and when 2.5 wt% platinum was supported on the γ-Al2O3, conversion increased further to 71 wt%. When the reaction was performed under helium instead of hydrogen, the cellulose conversion was not affected but naturally, no sorbitol was formed. It is thus clear that both the Al2O3 support itself and the metal has significant impact on the cellulose degradation, but the hydrogen gas does not seem to affect the hydrolysis at all but only serves to hydrogenate the formed monosaccharides. This obvi-ously controverts the mechanistic suggestion invoking hydrogen spillover (Figure 23) as a catalytic function.
Alumina exists in several allotropic forms of which γ-Al2O3 is the most common to use in heterogeneous catalysis due to its small particle size and large surface area.178 It is also more expensive and less stable than most oth-er phases. α-Al2O3 is the most stable alumina phase and together with amor-phous Al2O3 is the most common form found in nature.
We wanted to explore the reaction shown in Scheme 58 in detail and hopefully better understand the role of the support material by comparing cellulose conversion and hexitol yield produced using catalysts on different types of alumina supports. We thus prepared 2.5 wt% Pt on α- and γ-Al2O3 and on acidic, neutral and basic alumina samples according to the impregna-tion procedure reported by Fukuoka and Dhepe.175 The cellulose degradation reactions were also set up according to the previously reported procedures (see below for detailed procedure) and the results are presented in Table 3.
Metal
Support material Lewis acid site
H2
H
He-
H+
Formed proton
H
94
The yields of the two hexitols from cellulose degradation over Pt/γ-Al2O3
(entry 1) were almost identical to the ones obtained by Fukuoka and Dhepe using the same catalyst and conditions.175 With Pt/α-Al2O3 (entry 2), both the conversion and the yields dropped significantly, probably as a result of the lower surface area of the catalyst support. The conversion and hexitol yields from hydrolysis/hydrogenation over Pt on acidic, neutral and basic alumina (with the same particle size) differed. The cellulose conversion with acidic Al2O3 (entry 4) was 67% and thus similar to what obtained with γ-phase while the conversion over neutral alumina was only 35% (entry 5). The highest cellulose conversion (75%) was obtained using basic alumina (entry 6) and this catalyst also gave the highest total hexitol yield (36%). The yield with acidic alumina was similar, while neutral alumina produced only 13% total hexitols due to the low cellulose conversion. We had expected basic alumina to give the to give least cellulose hydrolysis, neutral to be intermediate and acidic alumina to be the best based on the number of acidic sites in the material, but the situation is clearly not as simple as that. It is possible that the presence of acidic or basic sites does not affect the reaction at all and that other more elusive differences were responsible for the differ-ences in hydrolysis rate. A more careful study of the catalyst structures would be necessary in order to extend this discussion.
Table 3. Platinum-catalyzed degradation of cellulose over different support materi-als.
Entry Support materiala Conversion (wt%)b Yield (mol%)c
Sorbitol Mannitol Total AA1 γ- Al2O3AA 60 25 6 32 AA2 α- Al2O3AA 44 12 2 14 AA3d γ- Al2O3AA
67 29 6 35 AA4 Acidic Al2O3AA 67 29 6 35 AA5 Neutral Al2O3AA 35 11 2 13 AA6 Basic Al2O3AA 75 28 8 36 AA7 Dowex 50W (H+)AA 45 1 0 1 AA8 No catalystAA 35 0 0 0 aAll supports had been impregnated with 2.5 wt% Pt. b Defined as the amount of solubilized cellulose (see detailed experimental below). c Determined by HPLC. d Impregnated with H2PtCl6 instead of Pt(0).
In order to further study acidity effects, γ-Al2O3 that had been modified with H2PtCl6, was tested as a catalyst (entry 3). This material should release six equivalents of HCl when subjected to H2 and heat and, indeed, a small in-crease in cellulose conversion and hexitol yield was observed. This can be attributed to increased acidity due to release of protons from HCl. Finally, the protic ion-exchange resin Dowex 50W was tested as a support material (entry 7). This material should also release protons into solution but does not provide any significant surface area and bears no acidic or basic sites, as in the case of alumina. Surprisingly, this material gave similarly low conver-
95
sion as α-Al2O3 and almost no sugar alcohols. Although the low hexitol yields can be explained simply by a very small catalytic area available for reduction, the low conversion indicates that Brønsted acidity of the support is only of minor importance for cellulose conversion under these reaction conditions. Taken together, these results highlight the significant effect of Pt/Al2O3 in the degradation of cellulose.
The work by Fukuoka and Dhepe sparked interest in the metal-catalyzed degradation of cellulose not only in Uppsala, but all around the world. Since then, several systems have been developed to produce polyols from cellu-lose. Modifying the catalyst together with optimization of reaction time and temperature, selectivity for different polyols can be achieved.179 For in-stance, smaller polyols such as ethylene glycol can be produced by perform-ing the reaction at higher temperatures (220–240 °C) over suitable metal catalysts.180
While some progress towards metal-catalyzed transformation of cellulose under high-temperature conditions have been made, enormous challenges still exist. In Table 3, entry 6, the cellulose conversion is 75% but the yield hexitols is only 36%. The metal-catalyzed hydrogenolysis of hexitols into shorter polyols, as noted above, can of course be one decomposition path-way. This reaction has been used for the direct preparation of ethylene gly-col, glycerol and other small alcohols from sugars and sorbitol.181 Dehydra-tion/hydrogenation reactions of polysaccharides can give rise to other by-products such as aliphatic alcohols and saturated hydrocarbons.182 Dumesic and co-workers have used heterogeneous metal catalysts to reform sorbitol to a variety of gaseous compounds including CO, CO2, CH4, H2 and others.183 Scheme 59 shows the possible reactions of aqueous sorbitol over metal catalysts in presence of dihydrogen at elevated temperatures (the pro-cesses temperatures span over the range 150–400 °C) and serves to illustrate both the challenge of selectivity and the opportunities offered by metal cata-lysts for the transformation of oxygen-rich biomass materials.
96
Scheme 59. Possible products from the metal-catalyzed transformations of aqueous sorbitol under hydrogen and at elevated temperatures.
Procedure for the platinum catalyzed formation of hexitols from cellulose In a glass-lined high pressure reactor, microcrystalline cellulose (Avicel PH101) (240 mg) and Pt/Al2O3 (102 mg) were suspended in H2O (25 ml). The reaction subjected to mechanical stirring (400rpm), purged three times with argon and pressurized to 50 bar with hydrogen gas. The reactor was then heated to 190 °C for 24h. After this time, the reactor was allowed to cool and then vented. The reaction suspension was then filtered and the solid residues dried and their weight was recorded. The filtrate was diluted up to 30.0 ml with H2O and then analyzed by HPLC equipped with a refractive index detector.
The HPLC separation was performed with a Rezex RPM Pb2+ column, 100*7.8 mm, and water as eluent 0.5ml/min at 60 °C. The hexitol yields were determined using calibration curves.
The cellulose conversion was calculated as follows: [(Initial mass of cata-lyst and cellulose) – (Mass of solid residues)] / (Initial mass of catalyst and cellulose)
OH1,2-propane diol
ethylene glycol
OHOHHO
glycerol
OHmethanol
HO
HO OH
isosorbideO
OHO
OH
sorbitolReforming
DehydrationHydrogenationHydrogenolysis
CleavagesRearrangementsDehydrationsDehydrogenations
Hydrogenolysis
HO OH
CO2 H2
CH4
OH
O
OH
Fischer-Tropsh
Alkanes
O
CO
acroleinlactic acid
hexane
1,2-propanediol
methane
97
Alumina supports: α-Al2O3 (corundum) (~100 mesh) was purchased from Sigma-Aldrich. γ-Al2O3 (~185m2/g) was purchased from Strem chemicals. Acidic, neutral and basic Al2O3
(Brockman I, 150 mesh, ~155m2/g) were purchased from Sigma-Aldrich.
3.2 Dehydrogenative Decarbonylation of alcohols 3.2.1 The transformation RCH2OH to RCHO + H2 Acceptorless dehydrogenation of alcohols (i.e. with the release of H2(g)) by metal-catalysis have been explored thoroughly.184 The main motivation has been 1) the dehydrogenation of alcoholic waste to form valuable H2, and 2) oxidation of alcohol without risk of over-oxidation of aldehydes. In 1977, Dobson and Robinson presented the metal-catalyzed dehydrogenation of alcohols with the release of H2 using Ru(OCOCF3)2(CO)(PPh3)2 34.185 Both primary and secondary alcohols could be oxidized, liberating aldehydes or ketones (Scheme 60).
Scheme 60. The acceptorless dehydrogenation of alcohols.
As in the case of alkenes, carbonyl hydrogenation is thermodynamically favorable186 so the reverse process dehydrogenation has to be driven forward by removal of hydrogen from the reaction mixture. Dobson and Robinson accomplished this by performing the reaction neat in refluxing alcohols, thus liberating gaseous H2. By measuring the amounts of evolved gas, the reac-tion rates could be monitored and they found that the reaction was promoted by additional trifluoroacetic acid. The authors also noted that when exchang-ing trifluoroacetate for acetate, the catalyst lost almost all activity. This ef-fect, combined with other observations led to the conclusion that the rate-determining step was the alcohol coordination and that –OOCCF3, being the best leaving group was more prone to dissociate and leave room for the in-coming alcohol.
Jung and Garrou found that a modified complex in which tri-phenylphosphine had been replaced by bidentate bis(diphenylphosphino) propane (DPPP) was more active in the dehydrogenation of certain alco-hols.187 When dehydrogenating n-heptanol using 34 as catalyst, conversion was very poor but interestingly, small amounts of hexane was detected in the reaction solution. Upon a closer examination, Ru(OCOCF3)2(CO)2(PPh3)2 35 was detected as a major component in the mixture (Scheme 61). Since 35
R R/H
OH
R R/H
O[M], !- H2
98
was completely inactive in the dehydrogenation reaction, decarbonylation of aldehydes to form metal-dicarbonyls was concluded to be an important cata-lyst deactivating reaction.
Scheme 61. The dehydrogenation of alcohols is hampered by decarbonylation of the product aldehyde and the formation of carbonylated catalyst.
Base-promoted dehydrogenation of alcohols was reported in 1987 by Morton and Cole-Hamilton.188 Using Ru(H)2(N2)(PPh3)3 as pre-catalyst together with NaOH (10 eq. with respect to catalyst) several primary alcohols could be dehydrogenated. Again the reactions were performed in neat alcohols and in a closed system (T = 150 °C).189 The authors did not attempt to probe the mechanistic details of the transition but claim to have isolated the heptacoor-dinate complex Ru(H)4(PPH3)3 by precipitation at lower temperature and thus suggested that liberation of dihydrogen could be the rate-determining step in the reaction.
In addition to using acid and base as co-catalysts, photocatalytic dehydro-genation reactions have been presented. Under photolytic conditions (125W Hg lamp, λ ≤ 300 nm), Wilkinson’s catalyst and derivatives thereof affects dehydrogenation of secondary alcohols at high rates, even at room tempera-ture (Scheme 62).190
Scheme 62. Dehydrogenation of isopropanol catalyzed by Wilkinsons catalyst and UV light.
Wilkinsons catalyst will not catalyze the reaction without irradiation, but addition of Et3N and PPh3 to the solution allows formation of RhH(PPh3)4 which is an active dehydrogenation catalyst in refluxing 2-propanol.191
Much progress has been made in this field since the above-discussed find-ings were reported, but two problems have essentially remained unsolved throughout the years. First, the dehydrogenation of primary alcohols has been problematic, partly due to the potential poisoning of the catalysts by decarbonylation (vide supra). Second, performing the reactions at low (<100 ºC) without the use of additives (usually base) or irradiation has proven dif-
RuO
PPh3
Ph3POC
O
CF3
OO
CF3
HO HexRuCO
PPh3
Ph3POC
OOCCF3OOCCF3
H2 + n-Hexane +!
34 35
OHRhR3P Cl
R3P PR3
h!
25 oC, neat in 2-propanol
O+ H2
99
ficult. Recently however, Beller and co-workers presented a Ru-P-N-P cata-lytic system that could dehydrogenate both primary and secondary small alcohols at reflux temperatures without the addition of base (Scheme 63).192
Scheme 63. Catalytic dehydrogenation of small alcohols using a Ru-P-N-P system.
The catalyst, 36, is formed in situ from Ru(H)2(CO)(PPh3)3 and the P-N-P ligand and the turnover frequencies after two hours were 8382/h for isopro-panol and 1483/h for ethanol, using 4.0 and 3.1 ppm of catalyst respectively. A catalytic cycle, based on and outer-sphere dehydrogenation step, was briefly proposed (Scheme 64). 36 liberates hydrogen to form a pentacoordi-nate amide complex which in turn accepts the substrate and affects the dehy-drogenation.
Scheme 64. Mechanistic proposal by Beller and co-workers, for the dehydrogenation of alcohols catalyzed by 36.
In conclusion, dehydrogenation is a challenging reaction for several reasons. The formation of aldehydes is thermodynamically uphill, and buildup of aldehydes that coordinates to the metal suppresses the reaction. It is probably worth noticing that all the examples discussed, and in fact most, if not all,
OH
R
RuPh3P HPh3P CO
reflux, neat
O
R+ H2
H
PPh3
NHPiPr2
PiPr2
RuN PP CO
H
H
+ - 3 PPh3H
For R = CH3: 4.0 ppm 36For R = H: 3.1 ppm 36
36
CH3: 8382 h-1
H: 1483 h-1TOFs after 2h
RuN PP CO
H
H
H
RuN PP CO
H
H2
OH
R
O
R
36
100
successful catalysts for this reaction, does not operate by simply doing a oxidative addition into the H-O bond to form H-M-OR. Instead, there is a function in the ligand, in the solvent or additive that abstracts the proton.
3.2.2 The transformation RCHO to RH + CO Unlike dehydrogenation of alcohols, decarbonylation of aldehydes is exer-gonic since it both increases entropy and forms a stable CH-bond. Decar-bonylation can, just like hydrogenation and dehydrogenation, be done using RhCl(PPh3)3.193 The stoichiometric reaction with aldehydes (RCHO) takes place at room temperature to form Rh(CO)Cl(PPh3)2 and the corresponding alkane (RH).194 When aldehydes that carry β-hydrogens are used, small amounts of the alkene product is also formed along with H2 (Scheme 65). The product Rh(CO)Cl(PPh3)2 is active as a catalyst >200 ºC albeit ineffec-tively. The high temperature required is unpractical for several reasons, al-dehyde stability is one problem and the high temperatures makes β-elimination even more significant, thus lowering selectivity.
Scheme 65. Stoichiometric decarbonylation of aliphatic aldehydes with RhCl(PPh3)3 forms alkanes and alkenes.
When the chelating diphosphine bis(diphenylphosphino)propane (DPPP) is used instead of two monodentate phosphines as in trans-Rh(CO)Cl(PPh3)3, the reaction proceeds at significantly lower temperatures.195 The complex Rh(DPPP)2Cl catalyze the decarbonylation of heptanal and benzaldehyde at temperatures >100 ºC (Scheme 66). The decarbonylation of heptanal is much more selective under these conditions with hexane being the only product.
RhPh3P ClPh3P PPh3 25 oC, CH2Cl2
24hH
O
HexRhPh3P Cl
OC PPh3+ Hexane
-PPh3+
90% yield 86% yield
1-Hexene+14% yield
+H2
101
Scheme 66. Catalytic decarbonylation of aldehydes using rhodium bearing diphos-phine ligands.
Madsen and co-workers studied the mechanism of this reaction in detail and found that the reaction goes through an oxidative addition, migratory extru-sion, and reductive elimination cycle, with the migratory extrusion being the rate-determining step.196 The precatalyst looses one of its P,P-ligands to form a very reactive tricoordinated Rh(I) species. This species then enters the catalytic cycle and coordinates at least one CO to form Rh(CO)(DPPP)L (probably after one catalytic cycle). It was not deduced if L = CO or Cl but computational data indicated that the energy difference between the two cases L = CO and L = Cl was small.
Tsuji and co-workers, who had done many of the initial discoveries more than 30 years ago, recently presented an improved system for aldehyde de-carbonylation. They found that a simple catalytic system, based on [Ir(cod)Cl]2 and PPh3 1:2 was effective for aldehyde decarbonyation in re-fluxing dioxane.197 Using 2.5 mol% catalyst, both aromatic and aliphatic aldehydes were decarbonylated over 24 h, corresponding to TOFs of ~2 /h (Scheme 67). Again, when aliphatic aldehydes carrying β-hydrogens were decarbonylated, <10% alkenes were also obtained from competing β-elimination.
Scheme 67. Iridium-catalyzed decarbonylation of aldehydes.
A neat synthetic application of decarbonylation, relating back to the hydro-genation of 1,1-disubstituted alkenes (Paper III), has been presented. Car-
H
O
HexHexane
TOF = 0.3 h-1
H
O
Ph
Rh(DPPP)2Cl
toluene, reflux(115 oC)
m-xylene, reflux(140 oC)
Rh(DPPP)2Cl CO +
CO + Benzene
TOF = 0.5 h-1
H
O
H
O
1,4-dioxane, reflux(101 oC)
0.025 [Ir(cod)Cl]2 CO +
CO +Ph
PhPh
Ph92% yield
72% yield(6% styrene)
0.05 PPh3
1,4-dioxane, reflux(101 oC)
0.025 [Ir(cod)Cl]20.05 PPh3
102
reira and co-workers performed rhodium-catalyzed asymmetric Michael addition on α,β-unsaturated aldehydes yielding chiral 3,3-diarylpropanals, which were subsequently decarbonylated to give 1,1-diarylethanes in high enantioselectivity (Scheme 68). The sequence could be performed conven-iently by simply evaporating the solvents of the addition reaction and replac-ing it with p-cymene and decarbonylation catalyst ([Rh(cod)Cl]2 and DPPP 1:4). Decarbonylation under a stream of N2 (to remove CO) then gave prod-uct without loss of optical purity. This method is especially suitable to pre-pare chiral diarlyethanes with two very similar aryl groups that would not be distinguishable by a hydrogenation catalyst.
Scheme 68. Asymmetric Michael addition followed by rhodium-catalyzed decar-bonylation affords 1,1-diaryl substituted ethanes in high enantioselectivity.
3.2.3. Dehydrogenative Decarbonylation As mentioned, the dehydrogenation of alcohols is thermodynamically unfa-vorable and not easily performed. This prompted Morton and Cole-Hamilton to couple it with decarbonylation and water gas shift (Scheme 69).188 The water gas shift reaction is highly exothermic, well making up for the uphill dehydrogenation reaction. Additionally, since the goal was production of fuels from alcoholic waste material, the additional hydrogen was considered a useful product. The overall reaction is presented in Scheme 69.
Ph
O
H(HO)2B
O
0.03 [Rh(C2H4)2Cl]20.5 KOH
Ph
O
H
0.08 DPPP0.02 [Rh(cod)Cl]2
O
Ph
O
p-cymene 140 oC, 44 hN2 stream
+
MeO
iBuPh
L20
0.033 L20
MeOH / H2O50 oC, 75 min
71% yield (two steps)93% ee
103
Scheme 69. Coupling together dehydrogenation, decarbonylation and water-gas shift, results in a highly exergonic reaction.
The reactions were performed at 120 ºC in a sealed system and neat in etha-nol with 5 vol% H2O but the results were poor, probably because the formed CO was not released but instead poisoned the catalysts. In order to obtain noteworthy gas formation, NaOH had to be added to the system.198 Base catalyze both the dehydrogenation, by deprotonating the alcohol to form alkoxide that coordinates easier to the metal catalyst, and the water-gas shift by nucleophilic attack on coordinated CO to liberate CO2 (M-CO + –OH → CO2 + M-H) The best results were obtained using [Rh(bipy)2]Cl (bipy = 2,2’-bipyridine) as catalyst and 1M NaOH, giving TOFs >100 h-1. Since ethanol was used, methane was formed in addition to H2 and the molar ratio between H2, CO and CH4 was 9:1:1. This carbon-loss was attributed mainly to aldol condensations between ethanal molecules under basic conditions.
As can be deduced from Scheme 69, subjecting alcohols (RCH2OH) to dehydrogenation followed by decarbonylation (dehydrogenative-decarbonylation) yields H2, CO and RH. Aside from the water-gas shift cou-pling strategy, two other successful approaches to accomplish this reaction have been presented.
A photocatalytic approach, as briefly discussed for dehydrogenation above, was taken by Sadow and co-workers.199 Irradiation can assist removal of both H2 and CO from a metal, and together with 10 mol% of a rhodium catalyst 37, primary alcohols could be converted over 24–72 hours (Scheme 70). The reaction was performed at room temperature under irradiation from a 450W Hg-lamp and gave high yields of both aliphatic and aromatic hydro-carbons. Since no aldehydes could be detected during the reaction, it was assumed that dehydrogenation is the slowest step and that competition be-tween alcohol and CO for open coordination sites limited the turnover fre-quency.
Dehydrogenation
Decarbonylation
Water-Gas shift
RCH2OH RCHO H2
RCHO CORH
CO H2O CO2 H2
!G~ +40 kJ/mol
!G~ –30 kJ/mol
!G~ –40 kJ/mol
RCH2OH RH 2H2 CO2
+
+
+
+ +
+
!G~ –30 kJ/mol
104
Scheme 70. Dehydrogenative decarbonylation catalyzed by complex 37 and photo-catalysis.
Very recently, Olsen and Madsen reported the iridium-catalyzed thermal dehydrogenative decarbonylation of alcohols.200 Combinations of iridium precursors and mono- and bidentate phosphines were tested and [Ir(cod)Cl]2 and BINAP (1:2) was shown to give highest yields of naphthalene when subjecting 2-naphthylmethanol to the reaction conditions. Reflux the alco-hols in mesitylene (bp = 165 ºC) with 5 mol% catalyst for 16–24 hours gave high product yields. The reaction was surprisingly tolerant to functional groups, aliphatic and aromatic alcohols containing sulfides, aromatic halo-gens, ethers and amides could be decomposed. This is the first example of effective and selective dehydrogenative decarbonylation of alcohols using solely metal catalyst and heat.
3.2.4 Transfer of CHOH from polyols to alkenes (Paper VI) During our effort to study the mechanistic details of the dehydrogenative decarbonylation of alcohols using Ir-P,P systems (not included in this the-sis)201 we became interested in performing the reaction with polyol sub-strates and in the possibility to study it by monitoring a secondary reaction where CO and H2 is consumed.
Degradation of polyols We initially subjected the diol 38 to dehydrogenative decarbonylation in refluxing mesitylene (bp = 165 °C) using the catalytic system developed by Olsen and Madsen200 and found that, in a comparison with the mono-alcohol 2-(2-naphthyl)ethanol, 38 was completely converted to 2-methyl-naphthalene in less than twice the time when the same molar ratio of catalyst per OH groups used. In addition to this finding, analysis of the reaction mix-ture revealed that 39 was a major intermediate in the reaction (Scheme 71).
RhN HN CO
benzeneRT, 24 h
+ H2
N
CO
0.1 37h!
37
R OH RH + CO
BPh O
NBPh
N
O
N ONN
NBPh=
R = Cy - 94% yieldR = Ph - 92% yieldR = Bn - 99% yield
105
Scheme 71. Dehydrogenative decarbonylation of the diol 38 builds up 39 during the formation of 2-methylnaphthalene.
Dehydrogenation of primary alcohols is usually slower than dehydrogena-tion of secondary alcohols,192 but no rationale for this phenomenon has been presented and it is possible that aldehyde build-up accompanied with decar-bonylation and catalyst deactivation is the reason for the lower apparent reactivity of primary alcohols (see part 3.2.1). When we compared 2-(2-naphthyl)ethanol 40 with the secondary alcohol 1-phenyl-2-propanol using the [Ir(cod)Cl]2:BINAP catalytic system refluxing in mesitylene, the former is dehydrogenated significantly faster. This may be due to that continuous removal of the formed aldehydes (in the case of 40) allows dehydrogenation to proceed unabated. It is therefore likely that 39 is formed by dehydrogena-tion of the primary alcohol of 38 followed by tautomerization. Tautomeriza-tion of α-hydroxy aldehydes to α-hydroxy ketones is frequently observed in solutions containing lewis-acidic metal complexes.202 In addition to diols, longer polyols could also be degraded, albeit requir-ing longer reaction times.
Tandem dehydrogenative decarbonylation - hydroformylation Since polyols could be decomposed to liberate H2 and CO, we decided to attempt coupling the dehydrogenative decarbonylation with a reaction that consumes the gases, namely hydroformylation.203 For this purpose a dual-reactor setup as shown in Figure 24 was conceived. The reactor consisted of two schlenk tubes, joined together with a tube to allow passage of gases from reactor A where the dehydrogenative decarbonylation was performed (liberation of H2 and CO) to reactor B where hydroformylation was per-formed (consumption of H2 and CO).
OHOH
OOH
H2
38 39
- 2 CO- H2
0.025 [Ir(cod)Cl]20.05 BINAP
mesitylene, reflux
106
Figure 24. Dual-reactor setup for tandem dehydrogenative decarbonylation – hydro-formylation.
The reactor was also equipped with a cold-finger condenser and a 0–2000 mbar pressure reader. For the hydroformylation, styrene was chosen as sub-strate and as catalyst we chose Wilkinson’s hydroformylation catalyst Rh(H)(CO)(PPh3)3 which was used in 1 mol% loading with respect to sty-rene.204 To our knowledge, the hydroformylation of alkenes has not been performed with less than 1 bar total partial pressure of CO and H2, although many hydroformylation catalysts have been developed throughout the years, most operate under 10–100 bar pressure. Rh(H)(CO)(PPh3)3 is one of the few catalysts that has been thoroughly studied and shown to operate effi-ciently at atmospheric pressure.204-205
Initially, we performed the dehydrogenative decarbonylation of mono-alcohol 2-(2-naphthyl)ethanol 40 in refluxing mesitylene with 5 mol% cata-lyst ([Ir(cod)Cl]2:(S)-BINAP 1:2). We found that with 1.5 equivalents of 40, the styrene was consumed after 44 hours, leaving a mixture of the linear and the branched aldehyde product in reactor B (Scheme 72). Analysis of the residues from the dehydrogenative decarbonylation in reactor A revealed that 93% of the alcohol had been consumed giving β-methylnaphthale in 85% yield, thus showing that 1.5 equivalents of CHOH functions per styrene is enough to allow complete conversion of the alkene in our system.
Cold fingercondenser
Pressure reader
Reactor ADehydrogenative Decarbonylation
!
Reactor BHydroformylation
Gas transfer tube
107
Scheme 72. The iridium-catalyzed formation of CO and H2 coupled with hydro-formylation of styrene yields a mixture of aldehydes.
The poor linear-branched aldehyde selectivity is not surprising since Rh(H)(CO)(PPh3)3 exhibits modest selectivity in the hydrogenation of sty-rene at one atmosphere of H2:CO 1:1. The total partial pressure of gases, corrected for the vapour pressure of the solvents, was ∼130 mbar throughout the reaction. In fact, the selectivity towards one of the aldehydes never ex-ceeded a 3:2 ratio during our study, and the regioselective hydroformylation of styrenes under these very low-pressure conditions will have to be a sub-ject of future studies. With these encouraging results in mind, we went on to optimize our system for the use of longer polyols.
Figure 25. Alcohols used in our initial screening.
Figure 25 depicts the alcohols we used in our initial investigation and Table 4 illustrates the reaction outcome. The temperatures given in Table 4 are the temperatures in the oil bath heating reactor A, since the pressure in the system increases during the reaction, temperatures were set well above the boiling points of the solvents. The results using mono-alcohol 40 (entry 1) has already been discussed and knowing that full conversion of styrene was possible, we moved on and replaced the mono-alcohol with diol 38 as substrate for the dehydrogenative decarbonylation. With this substrate, still using 5 mol% catalyst per CHOH–
0.075 [Ir(cod)Cl]20.15 BINAP
mesitylene, reflux, 44h
H2 + CO
PhPh
Ph
O
OH
O
0.01 Rh(H)(CO)(PPh3)3benzene, RT, 44h
1
1.5
Reactor A, Dehydrogenative Decarbonylation
Resctor B, Hydroformylation
85% yield
40% yield58% yield
40
OH OHOH
OH
OH
OH
OH
O
3840 41
108
group, high conversion of styrene was maintained (Table 4, entry 2). The formation of aldehydes in 85% yield was accompanied by 11% yield of the hydrogenated product ethylbenzene. The formation of ethylbenzene is a function of an excess of hydrogen and it is likely that the rapid formation of ketoalcohol 39 (Scheme 71) is responsible for an initial formation of H2.
Table 4. Dehydrogenative decarbonylation of alcohols coupled with the hydro-formylation of styrenes.
Reactor A Reactor B Entry Substratea [Ir]b Solvent Temp. Yield (mol%)c
(mol%) (ºC) Aldehydes EtPh Styrene 1 40 5.0 Mes 185 98 0 <1 2 38 5.0 Mes 185 85 11 <1 3 41 5.0 Mes 185 26 10 57 4 38 2.5 DGDE 210 84 13 <1 5 41 2.5 DGDE 210 64 23 10
a 1.5 molar equivalents of CHOH moieties relative to the amount of styrene. b Mol% with respect to the number of CHOH moieties in reactor A. c Determined by 1H NMR analysis of the crude reaction solution using 1,3,5-trimethoxybenzene as internal standard. Tetraol 41 was also tested as a substrate but it had poor solubility in me-sitylene and gave only 36% conversion of styrene (Table 4, entry 3). In order to improve the solubility of polyols diethylene glycol diethylether (DGDE) was used as solvent (bp = 185 °C). Using these conditions the catalyst load-ing could be dropped to 2.5 mol% per CHOH-group while maintaining good conversion of styrene. 84 and 64% yield of aldehydes was obtained using diol 38 and tetraol 41 as sources of CO and H2 (entries 4 and 5). Notably, the amount of ethylbenzene increased with the number of CHOH-groups in the substrate. This effect is probably due to fast dehydrogenations and tautomer-izations, analogous to the effect seen for diol 38. As shown in Scheme 73 a polyol A would dehydrogenate and form isomers containing internal carbon-yls B. For a longer polyol several dehydration-isomerization could take place, magnifying the fast H2 release, thus producing more hydrogenation product.
Scheme 73. Successive dehydrogenations and isomerizations of a polyol forms ke-tones and releases H2.
R
OH
OH
n
A
R
O
OH
OHm
n-m
B
+ mH2[Ir]
109
Next we wanted to use the hexitol sorbitol as a substrate for the CHOH-transfer. The selective formation of sorbitol and other polyols is possible both from the sugar monomers and sugar alcohols206 and, as discussed in 3.1.2, they can also be obtained directly from cellulose by hydrolytic-reductive techniques. Thus, we saw this as an opportunity to use the hydro-formylation reaction with renewable CO and H2. Table 5 lists the results from CHOH-transfer from sorbitol to styrene under a variety of conditions using DGDE as solvent throughout. In all cases, except when mentioned explicitly, [Ir(cod)Cl]2 and (S)-BINAP (1:2) were used as precatalysts for the dehydrogenative decarbonylation in reactor A and Rh(H)(CO)(PPh3)3 was used for the hydroformylation, 1 mol% relative to styrene in reactor B. Initially we used 1.5 equivalents of CHOH-groups relative to the amount of styrene, giving 53% yield of aldehydes and incomplete conversion of styrene (Table 5, entry 1). Lowering the catalyst ratio to 1 mol% per CHOH-group (i.e. 6 mol% with respect to sorbitol) (entry 2) gave only slightly inferior results, showing again that polyols are a favourable substrate for dehydro-genative decarbonylation. Increasing the reaction time only improved the aldehyde yield slightly, possibly because the dehydrogenative decarbonyla-tion is limited by sorbitol decomposition such as the formation of cycliza-tion-products. Increasing the amount of CHOH-functions to 2 (i.e. 0.33 equivalents of sorbitol relative to styrene) afforded better conversion of sty-rene and 66% yield of aldehydes (entry 3). This experiment was repeated three times to ensure reproducibility and gave product yields identical within a few percent between runs. Next, we wanted to see if the chirality of the catalyst affected the reaction, (R)- and rac-BINAP was therefore tested as ligands to iridium (entries 4 and 5). While (R)-BINAP gave virtually identical results as the S-isomer, race-mic BINAP gave significantly lower conversion and yield. The reason for the lower activity can be attributed to solubility problems since precipitation of solid materials was observed when the racemic ligand was used. An attempt to lower the reaction time (entry 6) to 22 hours gave almost a 1:1:1 ratio of aldehydes, ethylbenzene and styrene indicating that hydrogena-tion takes place during the earlier part of the reaction. Lowering the heating bath temperature (entry 7) to 195 °C also decreased the yield of aldehydes. To ensure reproducible results, the system was kept strictly air and water-free. DGDE (miscible with H2O) was dried by heating with LiAlH4 for sev-eral days followed by vacuum-distillation into a schlenk flask containing 4Å molecular sieves. Presence of water in the reaction lowered the styrene con-version and aldehyde yields, presumably by coordinating to the catalyst. Additionally, together with the presumably lewis-acidic catalyst promoted hydrolysis of the ethereal solvent, which was unacceptable since the alcohol-ic hydrolysis products underwent dehydrogenative decarbonylation, giving hydroformylation of styrene even in the absence of polyol substrate! Under
110
the dry experiment conditions, a blank reaction (Table 5, entry 8) without sorbitol still gave 2% yield of aldehydes.
Table 5. Dehydrogenative decarbonylation of sorbitol coupled with hydroformyla-tion of styrene.
Reactor A Reactor B Entry Substratea [Ir]b Temp.c Time Yield (mol%)d
(CHOH eq) Mol% (ºC) (h) Aldehydes EtPh Styrene 1 1.5 2.5 210 44 53 21 22 2 1.5 1 210 44 46 20 24 3 2.0 1 210 44 66 28 5 4e 2.0 1 210 44 65 30 3 5f 2.0 1 210 44 50 30 15 6 2.0 1 210 22 32 31 33 7 2.0 1 195 44 25 27 41 8 0 1 210 44 2 0 96 9 2.0 0 210 44 0 0 99
a Molar equivalents of sorbitol CHOH moieties relative to the amount of styrene in reactor B. b Mol% with respect to the number of CHOH moieties in reactor A. c Temperature in the heating bath of reactor A. d Determined by 1H NMR analysis of the crude reaction solution using 1,3,5-trimethoxybenzene as internal standard. e Using (R)-BINAP as ligand. f Using rac-BINAP as ligand. In addition to sorbitol, the polyols mannitol (C6), xylitol (C5), erythritol (C4) and glycerol (C3) were tested as substrates for the tandem dehydro-genative decarbonylation-hydroformylation (See Paper VI for details). Out of these, glycerol afforded the highest yields and least hydrogenation prod-uct, showing again that shorter polyols are preferred from a selectivity point of view. Glycerol is not only obtainable from polysaccharides, it also exists in a large surplus as a byproduct from production of fatty acid methyl esters (FAMEs) that are used as biodiesels and much attention has been given to finding applications for this byproduct.207 Reduction of trigycerides, derived from vegetable oils, gives long aliphatic alcohols (saturated fatty alcohols), glycerol and other C3-alcohols,208 so we decided to use the CHOH-transfer system on a mixture of this kind. Thus, a 3:1 mixture of 1-hexadecanol and glycerol was subjected to dehydrogenative decarbonylation and coupled with hydroformylation of styrene (Scheme 74). A 1.5:1 ratio of CHOH-units to styrene was used and since 1-hexadecanol is a mono-alcohol and thus reacts more slowly, longer (66 h) reaction time was used. From the dehydrogenative decarbonylation reaction (Reactor A), pen-tadecane was isolated in 89% yield by diluting the reaction with water and extracting with pentane. The hydroformylation reaction (reactor B) gave a mixture of the two aldehydes in 98% combined yield.
111
Scheme 74. Conversion of fatty alcohols and glycerol into hydrocarbons and alde-hydes by metal-catalyzed processes.
3.3 Conclusion In the first part of this chapter, the role of the support-material for the plati-num-catalyzed hydrolysis-hydrogenation of cellulose to obtain polyols was studied. The solubilization of cellulose and the yield of hexitols did not show any direct correlation to the acidity of the platinum support material. Various polyols were subjected to dehydrogenative decarbonylation and coupled with hydroformylation. It was shown that selective formation of CO and H2 from polyols is possible and that the gases generated could be used for alkene hydroformylation. Thus, using this system, it was possible to transfer CHOH functions directly from renewable polyols to synthetic inter-mediates.
HO
HO
HO
CO + H2
3 C15H32
0.04 Rh(H)(CO)(PPh3)34
0.03 [Ir(cod)Cl]20.06 (S)-BINAP
3 C16H33OH 1
4
+
PhPh
O Ph
O
98% total yield
DGDE, reflux, 66h
+
89% yield
Benzene, RT, 66h
Reactor A, Dehydrogenative Decarbonylation
Reactor B, Hydroformylation
112
4 Summary in Swedish
Metaller och organiska kolföreningar som t.ex. socker och oljor har mycket olika egenskaper, även på molekylnivå. Därför kan metaller åstadkomma kemiska reaktioner hos organiska molekyler som annars inte skulle kunna ske. Exempelvis så används finfördelad platinametall i bilars avgasrenare för att ta bort kolmonoxid och andra giftiga gaser. Gaserna absorberas först på metallen tillsammans med syrgas. På grund av interaktioner med metallen har gaserna helt andra egenskaper när de är absorberade på metallytan och reagerar därför med varandra. Från giftig kolmonoxid och syrgas bildas därför vanlig koldioxid. Bilars avgasrenare kallas för katalysatorer eftersom metallen katalyserar (snabbar upp) reaktionen mellan avgasmolekylerna och syrgas utan att själv förbrukas. Ett katalysatorsystem i en bil håller vanligtvis hela bilens livslängd och kan katalysera reningen av tusentals liter avgaser.
Metallkatalysatorer behöver inte vara fasta material utan kan även vara upplösta i ett lösningsmedel. Sådana upplösta metaller kan kopplas ihop med olika molekyler (ligander) som förändrar metallens egenskaper och gör att metallen kan katalysera helt nya reaktioner som annars hade varit omöjliga.
Vi har använt speciellt designade metall-ligand-föreningar för att göra kemiska reaktioner snabbt och rent, reaktioner som utan metallen skulle vara omöjliga att genomföra och/eller generera mycket avfall.
Kolmonoxid (CO)
Andra giftiga avgaserMetallyta
Syrgas (O2)
COO2CO2
Mindre giftiga gaserCO2
MLigand 1
Ligand 2
Metallatom M med modifierare (ligander)
Substrat
Produkt
113
I den första delen av avhandlingen har vi designat nya metallkatalysatorer
och använt dem för att sätta in så kallade kirala centra selektivt i molekyler för t. ex. läkemedelsutvekling. Molekyler med kirala centra bildar vanligtvis kemiskt identiska spegelbilder av samma substans (enantiomerer), men ge-nom att använda katalysatorer som själva innehåller kirala ligander kan en-dast en enantiomer produceras selektivt.
Ett exempel på en kiral substans är Karvon, vars olika enantiomerer alltså är kemiskt identiska, men luktar olika och alltså har olika effekter i kroppen.
Läkemedel som bara innehåller en enantiomer får unika egenskaper i kroppen, minskar oönskade bieffekter och är ofta mer potenta än blandningar av de olika enantiomerna, vilket också minskar biverkningar.
I den andra delen av avhandlingen har vi använt metallkatalysatorer för
att bryta ned förnyelsebara naturprodukter som cellulosa, till små polyalko-holer (polyoler). Naturprodukter som socker och även polyoler är mycket olika de vanliga, råoljebaserade kemiska byggstenar som vi använder för att göra plaster, läkemedel, färger etc. För att kunna använda förnyelsebara kemikalier istället för kemikalier som kommer från råolja krävs alltså att deras kemiska egenskaper förändras.
Spegel
Enantiomer A Enantiomer B (R)-Karvon
O
(S)-Karvon
O
Luktar mynta Luktar kummin
Spegel
M
Metallkatalysator medkiral ligand
Startmaterial
Reaktion
Ren Enantiomer A
114
Med hjälp av metallkatalysatorer har vi successivt överfört atomer från polyoler till andra substanser. På det viset blir förnyelsebara råvaror använd-bara för att producera läkemedel och industrikemikalier.
MetallkatalysatorMetallkatalysator
StartmaterialPolyol
ProduktPolyol
Cellulosa
Metallkatalysator
115
5 Acknowledgements
Listed below are contributors to this thesis in one way or the other. Thank you all. My supervisor Pher G. Andersson for accepting me as a PhD student in he’s group. I sincerely appreciate the freedom you have given me to pursue ideas and manage my research. My brilliant friend Tamara L. Church for all inspiration and support. Its good to have a friend who enjoys both discussing d-orbital symmetries and crawling in the mud in the search for the western tarsier. My collaborators In Tarragona, Spain, Oscar Pamies, Montserrat Dieguez and Javier Mazuela In Copenhagen, Denmark, Robert Madsen and Esben P. K. Olsen In Trondheim, Norway, Odd R. Gautun In Durban, South Africa, Thavendran Govender In Rostock, Germany, Armin Börner, Mercedes Coll and Benjamin Schäffner In the PGA-group Alexander Paptchikhine who has been with me all the time discussing the issues of man and chemist. I know you will keep it cooking. Byron Peters, the superman of chemistry. Just do not let those vile machines get you! Mr. Taigang Zhou, the zen master, who always cheers up in the lab. It has been good to have you at the hood beside me. JiaQi Li, always super produc-tive and super surprised. I wish I had half the energy that you have! Mr. Xu! (I wish I had time to think out more nicknames for you) Take it easy and good luck with your special project. An also of course; Alban, who helped me proof-reading the thesis, Dr’s. Mattias, Päivi and Pradeep, Ian, Jarle and Klas who taught me many things, some of which I did not want to know, and also Tobias, Simone, Vijay and Oleg. My diploma workers Micke Nordlund, Kristof Jess, Jonas Sandell, Thomas Wartmann, Alma Numnell, Lina Löfgren, Hermann Kriegel and Karin Sjöstedt.
116
At the department of Chemistry A special thanks goes out to the after-work gang and all the others who helped making my PhD time enjoyable by sharing drinks and food; Christian D who served as a wailing wall when things and people did not behave, Henke J my fellow sailorman, Sara N the pirate who inspired me to start with photography, Rikard E, Classe A, Supaporn S, Anna L, Maxim G, Magnus, Andreas, Laura, Jia-Fei, Ola, Adam, Matt, Julius, Johan L, Miran-da, Alvi, and Puneet. All the shiny happy biochemists Cissi, Åsa, Tony, Angelica, Sofia, Nisse, Anna T, Johan W, Ann, Diana, Gunnar, Marcus and Malin My students, for making me laugh, both intentionally and not. The seniors at the department, especially; Adolf G, Joseph S, Peter D, Thomas N, Henrik O and Lars B for useful discussions and advice. All the TA-personnel; especially Gunnar, Bosse and Tomas, for helping me out and making things work at the department. Lars B and Micke W for employing such a lovely administrator. I also want to thank my friends, especially; Lina, Jesper, Klas and Tamara for all the great times and for always being there. And also; Mattias N, Jessika, Per, Gustaf, Michael, Erik, Johanna E, Anuja and Antti. Hoppetossans crew: Adde and André Malin, Anna, Måns and Kerstin for good times in Corvara. The Canadians; Allison, Ulises, Ted, Dora, Holly and Katy. My old classmates, Alle, Robert, Jolla, Eric and Danne Kaggen, Irja, Liselotte, Anders, Erik and of course my cameraman, John. The NBC-cycle the globe team: Elof and Per My fuzzy friends: Alfons, Bill, Kåre, Sally, Selma, Fiffi and Calle. My family and relatives; Especially aunt Evi for all support, fun and showing me the meaning of atti-tude! And all the mafiosos and mafiosas in the Italian Mafia.
My sweet Johanna and, the coolest man, Hugo, for (almost) making me grow up and putting up with me during the writing of this thesis. My wonderful family; Mamma, Pappa and Syster-Yster And all others who should have been mentioned, but were not.
117
6 References
(1) Nógrádi, M. Stereoselective Synthesis; VCH Verlagsgesellschaft: Weinheim, 1995.
(2) Kelvin In Baltimore Lectures on Molecular Dynamics and the Wave Theory of Light, Appendix H; C. J. Clay & Sons: London, 1904.
(3) Leitereg, T. J.; Guadagni, D. G.; Harris, J.; Mon, T. R.; Teranishi, R. J. Agric. Food Chem. 1971, 19, 785.
(4) Flack, H. D. Acta Crystallogr., Sect. A 2009, 65, 371. (5) Smith, M. B.; March, J. March's Advanced Organic Chemistry; John Wiley
& Sons: Hoboken, 2007. (6) Guijarro, A.; Yus, M. The Origin of Chirality in the Molecules of Life; The
Royal Society of Chemistry: Cambridge, 2009. (7) Carey, F. A.; Sundberg, R. J. Advanced Organic Chemistry, Part A;
Springer: New York, 2007. (8) (a) Job, A.; Janeck, C. F.; Bettray, W.; Peters, R.; Enders, D. Tetrahedron
2002, 58, 2253 (b) Enders, D.; Eichenauer, H. Angew. Chem. Int. Ed. 1979, 18, 397.
(9) Rothenberg, G. Catalysis: Concepts and Green Applications; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, 2008.
(10) Deloux, L.; Srebnik, M. Chem. Rev. 1993, 93, 763. (11) Berzelius, J. J. Årsberättelse om framstegen i Fysik och Kemi; P. A.
Nordstedt: Stockholm, 1835. (12) van Leeuwen, P. W. N. M. Homogeneous Catalysis; Kluwer Academic
Publishers: Dordrecht, 2004. (13) IUPAC Compendium of Chemical Terminology, 2nd ed. ; Blackwell
Scientific Publishing: Oxford, 1997. (14) Wang, X.; Andrews, L.; Riedel, S.; Kaupp, M. Angew. Chem. Int. Ed. 2007,
46, 8371. (15) Crabtree, R. H. In The Organometallic Chemistry of the Transition Metals;
John Wiley & Sons, Inc., 2005. (16) Vaska, L. Acc. Chem. Res. 1968, 1, 335. (17) Greenwood, A.; Earnshaw, A. Chemistry of the Elements; Butterworth-
Heinemann: London, 1997. (18) Gomes Carneiro, T. M.; Matt, C.; Braunstein, P. Coord. Chem. Rev. 1989,
96, 49. (19) Crabtree, R. Acc. Chem. Res. 1979, 12, 331. (20) Aresta, M.; Rossi, M.; Sacco, A. Inorg. Chim. Acta 1969, 3, 227. (21) Fujita, K.-i.; Yoshida, T.; Imori, Y.; Yamaguchi, R. Org. Lett. 2011, 13,
2278.
118
(22) Clark, J. H.; Deswarte, F. E. I. In Introduction to Chemicals from Biomass; John Wiley & Sons, Ltd, 2008.
(23) van Santen, R. A. In Catalysis for Renewables; Wiley-VCH Verlag GmbH & Co. KGaA, 2007.
(24) Lange, J.-P. In Catalysis for Renewables; Centi, G., van Santen, R. A., Eds.; Wiley-VCH: Weinheim, 2007.
(25) Sticklen, M. B. Nat. Rev. Genet. 2008, 9, 433. (26) Rylander, P. N. Hydrogenation Methods; Academic Press: London, 1985. (27) Crabtree, R. H. In The Organometallic Chemistry of the Transition Metals;
John Wiley & Sons, Inc., 2005. (28) Chinn, M. S.; Heinekey, D. M.; Payne, N. G.; Sofield, C. D.
Organometallics 1989, 8, 1824. (29) Chinn, M. S.; Heinekey, D. M. J. Am. Chem. Soc. 1990, 112, 5166. (30) Heinekey, D. M.; Oldham, W. J. Chem. Rev. 1993, 93, 913. (31) Kubas, G. J. Acc. Chem. Res. 1988, 21, 120. (32) Bautista, M.; Earl, K. A.; Morris, R. H.; Sella, A. J. Am. Chem. Soc. 1987,
109, 3780. (33) Bianchini, C.; Mealli, C.; Meli, A.; Peruzzini, M.; Zanobini, F. J. Am.
Chem. Soc. 1988, 110, 8725. (34) Bersuker, I. B. In Electronic Structure and Properties of Transition Metal
Compounds; John Wiley & Sons, Inc., 2010. (35) Albright, T. A.; Hoffmann, R.; Thibeault, J. C.; Thorn, D. L. J. Am. Chem.
Soc. 1979, 101, 3801. (36) Halpern, J. Inorg. Chim. Acta 1985, 100, 41. (37) Osborn, J. A.; Jardine, F. H.; Young, J. F.; Wilkinson, G. J. Chem. Soc. A
1966, 1711. (38) Knowles, W. S.; Sabacky, M. J. Chem. Commun. (London) 1968, 1445. (39) Horner, L.; Siegel, H.; Büthe, H. Angew. Chem. Int. Ed. 1968, 7, 942. (40) Dang, T. P.; Kagan, H. B. J. Chem. Soc. D 1971, 481. (41) (a) Osborn, J. A.; Schrock, R. R. J. Am. Chem. Soc. 1971, 93, 3089 (b)
Schrock, R. R.; Osborn, J. A. J. Am. Chem. Soc. 1976, 98, 2134. (42) (a) Knowles, W. S.; Sabacky, M. J.; Vineyard, B. D.; Weinkauff, D. J. J.
Am. Chem. Soc. 1975, 97, 2567 (b) Vineyard, B. D.; Knowles, W. S.; Sabacky, M. J.; Bachman, G. L.; Weinkauff, D. J. J. Am. Chem. Soc. 1977, 99, 5946.
(43) Knowles, W. S. J. Chem. Educ. 1986, 63, 222. (44) (a) Halpern, J. Science 1982, 217, 401 (b) Brown, J. M. Chem. Soc. Rev.
1993, 22, 25. (45) Chi, Y.; Tang, W.; Zhang, X. In Modern rhodium-catalyzed organic
reactions; Evans, P. A., Ed.; Wiley-VCH: Weinheim, 2005. (46) Miyashita, A.; Yasuda, A.; Takaya, H.; Toriumi, K.; Ito, T.; Souchi, T.;
Noyori, R. J. Am. Chem. Soc. 1980, 102, 7932. (47) Takaya, H.; Ohta, T.; Sayo, N.; Kumobayashi, H.; Akutagawa, S.; Inoue,
S.; Kasahara, I.; Noyori, R. J. Am. Chem. Soc. 1987, 109, 1596. (48) Ohta, T.; Takaya, H.; Kitamura, M.; Nagai, K.; Noyori, R. J. Org. Chem.
1987, 52, 3174.
119
(49) (a) Kitamura, M.; Tsukamoto, M.; Bessho, Y.; Yoshimura, M.; Kobs, U.; Widhalm, M.; Noyori, R. J. Am. Chem. Soc. 2002, 124, 6649 (b) Ashby, M. T.; Halpern, J. J. Am. Chem. Soc. 1991, 113, 589.
(50) Burk, M. J.; Feaster, J. E.; Harlow, R. L. Organometallics 1990, 9, 2653. (51) Burk, M. J. J. Am. Chem. Soc. 1991, 113, 8518. (52) Burk, M. J.; Wang, Y. M.; Lee, J. R. J. Am. Chem. Soc. 1996, 118, 5142. (53) Noyori, R.; Ohta, M.; Hsiao, Y.; Kitamura, M.; Ohta, T.; Takaya, H. J. Am.
Chem. Soc. 1986, 108, 7117. (54) Burk, M. J.; Stammers, T. A.; Straub, J. A. Org. Lett. 1999, 1, 387. (55) Zhu, G.; Chen, Z.; Zhang, X. J. Org. Chem. 1999, 64, 6907. (56) (a) Ohta, T.; Ikegami, H.; Miyake, T.; Takaya, H. J. Organomet. Chem.
1995, 502, 169 (b) Forman, G. S.; Ohkuma, T.; Hems, W. P.; Noyori, R. Tetrahedron Lett. 2000, 41, 9471 (c) Inagaki, K.; Ohta, T.; Nozaki, K.; Takaya, H. J. Organomet. Chem. 1997, 531, 159.
(57) Broene, R. D.; Buchwald, S. L. J. Am. Chem. Soc. 1993, 115, 12569. (58) Troutman, M. V.; Appella, D. H.; Buchwald, S. L. J. Am. Chem. Soc. 1999,
121, 4916. (59) Young, J. F.; Osborn, J. A.; Jardine, F. H.; Wilkinson, G. Chem. Commun.
(London) 1965, 131. (60) (a) Halpern, J.; Wong, C. S. J. Chem. Soc. Chem. Commun. 1973, 629 (b)
Meakin, P.; Jesson, J. P.; Tolman, C. A. J. Am. Chem. Soc. 1972, 94, 3240. (61) Halpern, J.; Okamoto, T.; Zakhariev, A. J. Mol. Catal. 1977, 2, 65. (62) Shapley, J. R.; Schrock, R. R.; Osborn, J. A. J. Am. Chem. Soc. 1969, 91,
2816. (63) Crabtree, R. H.; Gautier, A.; Giordano, G.; Khan, T. J. Organomet. Chem.
1977, 141, 113. (64) Crabtree, R. H.; Felkin, H.; Morris, G. E. J. Organomet. Chem. 1977, 141,
205. (65) Crabtree, R. H. J. Chem. Soc. Chem. Commun. 1975, 647. (66) Crabtree, R. H.; Felkin, H.; Fillebeen-Khan, T.; Morris, G. E. J.
Organomet. Chem. 1979, 168, 183. (67) Chodosh, D. F.; Crabtree, R. H.; Felkin, H.; Morris, G. E. J. Organomet.
Chem. 1978, 161, C67. (68) Helmchen, G.; Pfaltz, A. Acc. Chem. Res. 2000, 33, 336. (69) Schnider, P.; Koch, G.; Prétôt, R.; Wang, G.; Bohnen, F. M.; Krüger, C.;
Pfaltz, A. Chem. Eur. J. 1997, 3, 887. (70) Lightfoot, A.; Schnider, P.; Pfaltz, A. Angew. Chem. Int. Ed. 1998, 37,
2897. (71) Blackmond, D. G.; Lightfoot, A.; Pfaltz, A.; Rosner, T.; Schnider, P.;
Zimmermann, N. Chirality 2000, 12, 442. (72) Smidt, S. P.; Zimmermann, N.; Studer, M.; Pfaltz, A. Chem. Eur. J. 2004,
10, 4685. (73) Nishida, H.; Takada, N.; Yoshimura, M.; Sonoda, T.; Kobayashi, H. Bull.
Chem. Soc. Jpn. 1984, 57, 2600. (74) Woodmansee, D.; Pfaltz, A. In Iridium Catalysis; Andersson, P. G., Ed.;
Springer: Berlin, 2011; Vol. 34.
120
(75) Mazet, C.; Smidt, S. P.; Meuwly, M.; Pfaltz, A. J. Am. Chem. Soc. 2004, 126, 14176.
(76) Dietiker, R.; Chen, P. Angew. Chem. Int. Ed. 2004, 43, 5513. (77) Brandt, P.; Hedberg, C.; Andersson, P. G. Chem. Eur. J. 2003, 9, 339. (78) Fan, Y.; Cui, X.; Burgess, K.; Hall, M. B. J. Am. Chem. Soc. 2004, 126,
16688. (79) (a) Church, T. L.; Rasmussen, T.; Andersson, P. G. Organometallics 2010,
29, 6769 (b) Hopmann, K. H.; Bayer, A. Organometallics 2011, 30, 2483. (80) Zhou, J.; Ogle, J. W.; Fan, Y.; Banphavichit, V.; Zhu, Y.; Burgess, K.
Chem. Eur. J. 2007, 13, 7162. (81) Verendel, J. J.; Andersson, P. G. Unpublished Work, 2011. (82) Hegedus, L. S.; Söderberg, B. C. G. In Transition Metals in the Synthesis of
Complex Organic Molecules; University Science Books, 2010. (83) Diesen, J. S.; Andersson, P. G. In Modern reduction methods; Andersson,
P. G., Munslow, I. J., Eds.; Wiley-VCH: Weinheim, 2008. (84) Church, T. L.; Andersson, P. G. Coord. Chem. Rev. 2008, 252, 513. (85) (a) Källström, K.; Hedberg, C.; Brandt, P.; Bayer, A.; Andersson, P. G. J.
Am. Chem. Soc. 2004, 126, 14308 (b) Hedberg, C.; Källström, K.; Brandt, P.; Hansen, L. K.; Andersson, P. G. J. Am. Chem. Soc. 2006, 128, 2995 (c) Kaukoranta, P.; Engman, M.; Hedberg, C.; Bergquist, J.; Andersson, P. G. Adv. Synth. Catal. 2008, 350, 1168.
(86) Trifonova, A.; Diesen, J. S.; Chapman, C. J.; Andersson, P. G. Org. Lett. 2004, 6, 3825.
(87) Cheruku, P.; Church, T. L.; Trifonova, A.; Wartmann, T.; Andersson, P. G. Tetrahedron Lett. 2008, 49, 7290.
(88) Cheruku, P.; Diesen, J.; Andersson, P. G. J. Am. Chem. Soc. 2008, 130, 5595.
(89) Paptchikhine, A.; Cheruku, P.; Engman, M.; Andersson, P. G. Chem. Commun. 2009, 5996.
(90) Li, J.-Q.; Paptchikhine, A.; Govender, T.; Andersson, P. G. Tetrahedron: Asymmetry 2010, 21, 1328.
(91) Kaiser, S.; Smidt, S. P.; Pfaltz, A. Angew. Chem. Int. Ed. 2006, 45, 5194. (92) Verendel, J. J.; Andersson, P. G. Dalton Trans. 2007, 5603. (93) Alberti, M. N.; Orfanopoulos, M. Chem. Eur. J. 2010, 16, 9414. (94) Mihelich, E. D.; Eickhoff, D. J. J. Org. Chem. 1983, 48, 4135. (95) Kröhnke, F. Synthesis 1976, 1976, 1. (96) Malkov, A. V.; Bell, M.; Castelluzzo, F.; Kočovský, P. Org. Lett. 2005, 7,
3219. (97) Collomb, P.; von Zelewsky, A. Tetrahedron: Asymmetry 1995, 6, 2903. (98) Chelucci, G.; Berta, D.; Fabbri, D.; Pinna, G. A.; Saba, A.; Ulgheri, F.
Tetrahedron: Asymmetry 1998, 9, 1933. (99) (a) Blankenstein, J.; Pfaltz, A. Angew. Chem. Int. Ed. 2001, 40, 4445 (b)
McIntyre, S.; Hörmann, E.; Menges, F.; Smidt, S. P.; Pfaltz, A. Adv. Synth. Catal. 2005, 347, 282.
(100) Hilgraf, R.; Pfaltz, A. Adv. Synth. Catal. 2005, 347, 61. (101) van Leeuwen, P. W. N. M. Appl. Catal., A 2001, 212, 61. (102) Diéguez, M.; Pàmies, O. Chem. Eur. J. 2008, 14, 3653.
121
(103) Phillips, A. J.; Uto, Y.; Wipf, P.; Reno, M. J.; Williams, D. R. Org. Lett. 2000, 2, 1165.
(104) Zhu, Y.; Fan, Y.; Burgess, K. J. Am. Chem. Soc. 2010, 132, 6249. (105) Chianese, A. R.; Li, X.; Janzen, M. C.; Faller, J. W.; Crabtree, R. H.
Organometallics 2003, 22, 1663. (106) (a) Schrems, M. G.; Neumann, E.; Pfaltz, A. Angew. Chem. Int. Ed. 2007,
46, 8274 (b) Bell, S.; Wüstenberg, B.; Kaiser, S.; Menges, F.; Netscher, T.; Pfaltz, A. Science 2006, 311, 642.
(107) Cui, X.; Fan, Y.; Hall, M. B.; Burgess, K. Chem. Eur. J. 2005, 11, 6859. (108) Engman, M.; Diesen, J. S.; Paptchikhine, A.; Andersson, P. G. J. Am.
Chem. Soc. 2007, 129, 4536. (109) Cheruku, P.; Paptchikhine, A.; Church, T. L.; Andersson, P. G. J. Am.
Chem. Soc. 2009, 131, 8285. (110) Tolstoy, P.; Engman, M.; Paptchikhine, A.; Bergquist, J.; Church, T. L.;
Leung, A. W. M.; Andersson, P. G. J. Am. Chem. Soc. 2009, 131, 8855. (111) Giardello, M. A.; Conticello, V. P.; Brard, L.; Gagne, M. R.; Marks, T. J. J.
Am. Chem. Soc. 1994, 116, 10241. (112) Co, T. T.; Shim, S. C.; Cho, C. S.; Kim, T.-J.; Kang, S. O.; Han, W.-S.; Ko,
J.; Kim, C.-K. Organometallics 2005, 24, 4824. (113) Co, T. T.; Kim, T.-J. Chem. Commun. 2006, 3537. (114) Faller, J. W.; Smart, C. J. Organometallics 1989, 8, 602. (115) Buriak, J. M.; Klein, J. C.; Herrington, D. G.; Osborn, J. A. Chem. Eur. J.
2000, 6, 139. (116) (a) Perry, M. C.; Cui, X.; Powell, M. T.; Hou, D.-R.; Reibenspies, J. H.;
Burgess, K. J. Am. Chem. Soc. 2002, 125, 113 (b) Kaukoranta, P.; Church, T. L.; Verendel, J. J.; Andersson, P. G. Unpublished Work, 2009.
(117) Dieguez, M.; Mazuela, J.; Pamies, O.; Verendel, J. J.; Andersson, P. G. Chem. Commun. 2008, 3888.
(118) Mazuela, J.; Verendel, J. J.; Coll, M.; Schäffner, B.; Börner, A.; Andersson, P. G.; Pamies, O.; Dieguez, M. J. Am. Chem. Soc. 2009, 131, 12344.
(119) Bayardon, J.; Holz, J.; Schäffner, B.; Andrushko, V.; Verevkin, S.; Preetz, A.; Börner, A. Angew. Chem. Int. Ed. 2007, 46, 5971.
(120) (a) Huang, P.-Q. In Asymmetric Synthesis of Nitrogen Heterocycles; Wiley-VCH Verlag GmbH & Co. KGaA, 2009 (b) Toyooka, N. In Asymmetric Synthesis of Nitrogen Heterocycles; Wiley-VCH Verlag GmbH & Co. KGaA, 2009 (c) Troin, Y.; Sinibaldi, M.-E. In Asymmetric Synthesis of Nitrogen Heterocycles; Wiley-VCH Verlag GmbH & Co. KGaA, 2009.
(121) Brunel, J. M. Recent Research and development in Organic chemistry 2003, 7, 155.
(122) Kuwano, R. Heterocycles 2008, 76, 909. (123) (a) Glorius, F. Org. Biomol. Chem. 2005, 3, 4171 (b) Zhou, Y.-G. Acc.
Chem. Res. 2007, 40, 1357. (124) (a) Church, T. L.; Andersson, P. G. In Chiral amine synthesis: Methods,
Developments and Applications; Nugent, T. C., Ed.; Wiley-VCH: Weinheim, 2010 (b) Lim, S. H.; Ma, S.; Beak, P. J. Org. Chem. 2001, 66, 9056.
(125) Deiters, A.; Martin, S. F. Chem. Rev. 2004, 104, 2199.
122
(126) Vougioukalakis, G. C.; Grubbs, R. H. Chem. Rev. 2009, 110, 1746. (127) Blum, A. P.; Ritter, T.; Grubbs, R. H. Organometallics 2007, 26, 2122. (128) Verendel, J. J.; Zhou, T.; Li, J.-Q.; Paptchikhine, A.; Lebedev, O.;
Andersson, P. G. J. Am. Chem. Soc. 2010, 132, 8880. (129) Verendel, J. J.; Li, J.-Q.; Quan, X.; Peters, B.; Zhou, T.; Gautun, O. R.;
Govender, T.; Andersson, P. G. Chem. Eur. J. 2012, 18, 6507. (130) (a) Fujii, Y.; Furugaki, H.; Tamura, E.; Yano, S.; Kita, K. Bull. Chem. Soc.
Jpn. 2005, 78, 456 (b) Bethmont, V.; Fache, F.; Lemaire, M. Tetrahedron Lett. 1995, 36, 4235.
(131) Ginsburg, D. J. Org. Chem. 1950, 15, 1003. (132) Amat, M.; Pérez, M.; Minaglia, A. T.; Bosch, J. J. Org. Chem. 2008, 73,
6920. (133) Parker, K. A.; Fokas, D. J. Org. Chem. 2005, 71, 449. (134) Zakzeski, J.; Bruijnincx, P. C. A.; Jongerius, A. L.; Weckhuysen, B. M.
Chem. Rev. 2010, 110, 3552. (135) Lichtenthaler, F. W.; Peters, S. C. R. Chim. 2004, 7, 65. (136) Harris, E. E. Adv. Carbohydr. Chem. 1949, 4, 153. (137) Abatzoglou, N.; Chornet, E. In Polysaccharides: Structural Diversity and
Functional Versatility; Dumitriu, S., Ed.; CRC Press: New York, 1998. (138) Wyman, C. E.; Decker, S. R.; Himmel, M. E.; Brady, J. W.; Skopec, C. E.;
Viikari, L. In Polysaccharides: Structural Diversity and Functional Versatility: Second Editon; Dumitriu, S., Ed.; CRC Press: New York, 2005.
(139) Krässig, H. A. Cellulose: Structure, Accessibility and Reactivity; Gordon and Breach Publ.: Singapore, 1993.
(140) Braconnot, H. Annalen der physik 1819, 63, 347. (141) (a) Xiang, Q.; Lee, Y. Y.; Pettersson, P. O.; Torget, R. W. Appl. Biochem.
Biotechnol. 2003, 107, 505 (b) Roman, M.; Winter, W. T. Biomacromolecules 2004, 5, 1671.
(142) Garves, K. In Hydrolysis of cellulose: Mechanisms of enzymatic and acid catalysis; Brown, R. D., Jr., Jurasek, L., Eds.; American Chemical Society: Washington D.C., 1979.
(143) Camacho, F.; González-Tello, P.; Jurado, E.; Robles, A. J. Chem. Technol. Biotechnol. 1996, 67, 350.
(144) Selke, S. M.; Hawley, M. C.; Hardt, H.; Lamport, D. T. A.; Smith, G.; Smith, J. Ind. Eng. Chem. Prod. Res. Dev. 1982, 21, 11.
(145) Higgins, F. J.; Ho, G. E. Agricultural Wastes 1982, 4, 97. (146) Fengel, V. D.; Wegener, G.; Heizmann, A.; Przyklenk, M. Holzforschung
1977, 40, 65. (147) Sun, Y.; Lin, L.; Pang, C.; Deng, H.; Peng, H.; Li, J.; He, B.; Liu, S.
Energy & Fuels 2007, 21, 2386. (148) Simonsen, E. Angew. Chem. 1898, 11, 195. (149) (a) Saeman, J. F. Ind. Eng. Chem. 1945, 37, 43 (b) Mok, W. S.; Antal, M.
J., Jr.; Varhegyi, G. Ind. Eng. Chem. Res. 1992, 31, 94. (150) Harris, J. F.; Baker, A. J.; Connor, A. H.; Jeffries, T. W.; Minor, J. L.;
Pettersen, R. C.; Scott, R. C.; Springer, E. L.; Wegner, T. H.; Zerbe, J. L. “Two-stage, dilute sulfuric acid hydrolysis of wood: An investigation of the
123
fundamentals,” U.S. Department of Agriculture, Forest Products Laboratory, 1985.
(151) Antal, M. J.; Mok, W. S. L.; Richards, G. N. Carbohydr. Res. 1990, 199, 91.
(152) (a) Van de Vyver, S.; Geboers, J.; Jacobs, P. A.; Sels, B. F. ChemCatChem 2011, 3, 82 (b) Moreau, C.; Belgacem, M. N.; Gandini, A. Top. Catal. 2004, 27, 11.
(153) Colbran, R. L.; Davidson, G. F. J. Text. Inst. 1961, 52, T73. (154) Nevell, T. P. In Cellulose Chemistry and its Applications; Nevell, T. P.,
Zeronian, S. H., Eds.; Ellis Horwood: Chichester, 1985. (155) Alfredsson, B.; Samuelson, O. Svensk Papperstidning 1968, 71, 679. (156) (a) Bobleter, O. In Polysaccharides: Structural Diversity and Functional
Versatility: Second Editon; Dumitriu, S., Ed.; CRC press: New York, 2005 (b) Akiya, N.; Savage, P. E. Chem. Rev. 2002, 102, 2725.
(157) Bobleter, O. Prog. Polym. Sci. 1994, 19, 797. (158) Jin, F.; Zhou, Z.; Enomoto, H.; Moriya, T.; Higashijima, H. Chem. Lett.
2004, 33, 126. (159) Kuo, C.-H.; Lee, C.-K. Biores. Technol. 2009, 100, 866. (160) Potthast, A.; Rosenau, T.; Sartori, J.; Sixta, H.; Kosma, P. Polymer 2003,
44, 7. (161) Saalwächter, K.; Burchard, W.; Klüfers, P.; Kettenbach, G.; Mayer, P.;
Klemm, D.; Dugarmaa, S. Macromolecules 2000, 33, 4094. (162) (a) Swatloski, R. P.; Spear, S. K.; Holbrey, J. D.; Rogers, R. D. J. Am.
Chem. Soc. 2002, 124, 4974 (b) Pinkert, A.; Marsh, K. N.; Pang, S.; Staiger, M. P. Chem. Rev. 2009, 109, 6712.
(163) Li, C.; Zhao, Z. K. Adv. Synth. Catal. 2007, 349, 1847. (164) (a) Binder, J. B.; Raines, R. T. J. Am. Chem. Soc. 2009, 131, 1979 (b) Su,
Y.; Brown, H. M.; Huang, X.; Zhou, X.-D.; Amonette, J. E.; Zhang, Z. C. Appl. Catal., A 2009, 361, 117 (c) Chheda, J. N.; Roman-Leshkov, Y.; Dumesic, J. A. Green Chem. 2007, 9, 342.
(165) (a) Rinaldi, R.; Schüth, F. ChemSusChem 2009, 2, 1096 (b) Rinaldi, R.; Schuth, F. Energy Environ. Sci. 2009, 2, 610.
(166) Rinaldi, R.; Meine, N.; Stein, J. v.; Palkovits, R.; Schüth, F. ChemSusChem 2010, 3, 266.
(167) Mäki-Arvela, P.; Holmbom, B.; Salmi, T.; Murzin, D. Y. Catal. Rev. 2007, 49, 197
(168) Oster, B.; Fechtel, U. In Ullmann's Encyclopedia of Industrial Chemistry; Wiley-VCH Verlag GmbH & Co. KGaA, 2000.
(169) (a) Makkee, M.; Kieboom, A. P. G.; van Bekkum, H. Carbohydr. Res. 1985, 138, 225 (b) Broekhuis, R. R.; Budhlall, B. M.; Nordquist, A. F. Ind. Eng. Chem. Res. 2004, 43, 5146.
(170) (a) Balandin, A. A.; Vasyunina, N. A.; Barysheva, G. S.; Chepigo, S. V. Bull. Acad. Sci. USSR 1957, 6, 403 (b) Balandin, A. A.; Vasyunina, N. A.; Chepigo, S. V.; Barysheva, G. S. Proc. Acad. Sci. USSR 1959, 128, 839.
(171) (a) Baudel, H. M.; de Abreu, C. A. M.; Zaror, C. Z. J. Chem. Technol. Biotechnol. 2005, 80, 230 (b) Robinson, J. M.; Burgess, C. E.; Bently, M. A.; Brasher, C. D.; Horne, B. O.; Lillard, D. M.; Macias, J. M.; Mandal, H.
124
D.; Mills, S. C.; O'Hara, K. D.; Pon, J. T.; Raigoza, A. F.; Sanchez, E. H.; Villarreal, J. S. Biomass & Bioenergy 2004, 26, 473.
(172) (a) Werpy, T.; Petersen, G. “Top value added chemicals from biomass.,” U.S. Department of Energy, Pacific Northwest National Laboratory, 2004 (b) Bozell, J. J.; Petersen, G. R. Green Chem. 2010, 12, 539.
(173) Speck, J. C., Jr. Adv. Carbohydr. Chem. 1958, 13, 63. (174) (a) Geboers, J.; Van de Vyver, S.; Carpentier, K.; Blochouse, K. d.; Jacobs,
P.; Sels, B. Chem. Commun. 2010, 46, 3577 (b) Palkovits, R.; Tajvidi, K.; Procelewska, J.; Rinaldi, R.; Ruppert, A. Green Chem. 2010, 12, 972.
(175) Fukuoka, A.; Dhepe, P. L. Angew. Chem. Int. Ed. 2006, 45, 5161. (176) Hattori, H.; Yamada, T.; Shishido, T. Res. Chem. Intermed. 1998, 24, 439. (177) Jollet, V.; Chambon, F.; Rataboul, F.; Cabiac, A.; Pinel, C.; Guillon, E.;
Essayem, N. Green Chem. 2009, 11, 2052. (178) Rashkeev, S. N.; Sohlberg, K.; Glazoff, M. V.; Novak, J.; Pennycook, S. J.;
Pantelides, S. T. Physical Review B 2003, 67, 115414. (179) Verendel, J. J.; Church, T. L.; Andersson, P. G. Synthesis 2011, 1649. (180) (a) Zhang, Y.; Wang, A.; Zhang, T. Chem. Commun. 2010, 46, 862 (b)
Luo, C.; Wang, S.; Liu, H. Angew. Chem. Int. Ed. 2007, 46, 7636. (181) (a) Zhou, J. H.; Zhang, M. G.; Zhao, L.; Li, P.; Zhou, X. G.; Yuan, W. K.
Catalysis Today 2009, 147, Supplement, S225 (b) Blanc, B.; Bourrel, A.; Gallezot, P.; Haas, T.; Taylor, P. Green Chem. 2000, 2, 89 (c) Montassier, C.; Ménézo, J. C.; Hoang, L. C.; Renaud, C.; Barbier, J. J. Mol. Catal. 1991, 70, 99 (d) van Ling, G.; Ruijterman, C.; Vlugter, J. C. Carbohydr. Res. 1967, 4, 380.
(182) (a) Huber, G. W.; Cortright, R. D.; Dumesic, J. A. Angew. Chem. Int. Ed. 2004, 43, 1549 (b) Serrano-Ruiz, J. C.; Luque, R.; Sepulveda-Escribano, A. Chem. Soc. Rev. 2011, 40, 5266.
(183) (a) Roman-Leshkov, Y.; Barrett, C. J.; Liu, Z. Y.; Dumesic, J. A. Nature 2007, 447, 982 (b) Huber, G. W.; Dumesic, J. A. Catalysis Today 2006, 111, 119.
(184) (a) Dobereiner, G. E.; Crabtree, R. H. Chem. Rev. 2009, 110, 681 (b) Friedrich, A.; Schneider, S. ChemCatChem 2009, 1, 72 (c) Johnson, T. C.; Morris, D. J.; Wills, M. Chem. Soc. Rev. 2010, 39, 81.
(185) Dobson, A.; Robinson, S. D. Inorg. Chem. 1977, 16, 137. (186) Wiberg, K. B.; Crocker, L. S.; Morgan, K. M. J. Am. Chem. Soc. 1991, 113,
3447. (187) Jung, C. W.; Garrou, P. E. Organometallics 1982, 1, 658. (188) Morton, D.; Cole-Hamilton, D. J. J. Chem. Soc. Chem. Commun. 1987,
248. (189) Morton, D.; Cole-Hamilton, D. J. J. Chem. Soc. Chem. Commun. 1988,
1154. (190) Griggs, C. G.; Smith, D. J. H. J. Organomet. Chem. 1984, 273, 105. (191) Matsubara, T.; Saito, Y. J. Mol. Catal. 1994, 92, 1. (192) Nielsen, M.; Kammer, A.; Cozzula, D.; Junge, H.; Gladiali, S.; Beller, M.
Angew. Chem. Int. Ed. 2011, 50, 9593. (193) Tsuji, J.; Ohno, K. Tetrahedron Lett. 1965, 6, 3969. (194) Ohno, K.; Tsuji, J. J. Am. Chem. Soc. 1968, 90, 99.
125
(195) Doughty, D. H.; Pignolet, L. H. J. Am. Chem. Soc. 1978, 100, 7083. (196) Fristrup, P.; Kreis, M.; Palmelund, A.; Norrby, P.-O.; Madsen, R. J. Am.
Chem. Soc. 2008, 130, 5206. (197) Iwai, T.; Fujihara, T.; Tsuji, Y. Chem. Commun. 2008, 6215. (198) Morton, D.; Cole-Hamilton, D. J.; Utuk, I. D.; Paneque-Sosa, M.; Lopez-
Poveda, M. J. Chem. Soc. Dalton Trans. 1989, 489. (199) Ho, H.-A.; Manna, K.; Sadow, A. D. Angew. Chem. Int. Ed. 2012, 51,
8607. (200) Olsen, E. P. K.; Madsen, R. Chem. Eur. J. 2012, Accepted. (201) Olsen, E. P. K.; Verendel, J. J.; Andersson, P. G.; Madsen, R. Manuscript
in preparation. (202) (a) Nagorski, R. W.; Richard, J. P. J. Am. Chem. Soc. 2001, 123, 794 (b)
Choudhary, V.; Pinar, A. B.; Sandler, S. I.; Vlachos, D. G.; Lobo, R. F. ACS Cat. 2011, 1, 1724 (c) Román-Leshkov, Y.; Moliner, M.; Labinger, J. A.; Davis, M. E. Angew. Chem. Int. Ed. 2010, 49, 8954.
(203) Breit, B. In Topics in Current Chemistry; Krische, M., Ed.; Springer: Berlin, 2007; Vol. 279.
(204) Evans, D.; Osborn, J. A.; Wilkinson, G. J. Chem. Soc. A 1968, 3133. (205) Brown, C. K.; Wilkinson, G. J. Chem. Soc. A 1970, 2753. (206) Ruppert, A. M.; Weinberg, K.; Palkovits, R. Angew. Chem. Int. Ed. 2012,
51, 2564. (207) (a) Koutinas, A. A.; Du, C.; Wang, R. H.; Webb, C. In Introduction to
Chemicals from Biomass; Clark, J. H., Deswarte, F. E. I., Eds.; John Wiley & Sons: Chichester, 2008 (b) Sels, B.; D'Hondt, E.; Jacobs, P. In Catalysis for Renewables; Centi, G., van Santen, R. A., Eds.; Wiley-VCH: Weinheim, 2007.
(208) (a) Turek, T.; Trimm, D. L.; Cant, N. W. Catal. Rev. 1994, 36, 645 (b) Noweck, K.; Grafahrend, W. In Ullmann's Encyclopedia of Industrial Chemistry; Wiley-VCH: Weinheim, 2006.
Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 984
Editor: The Dean of the Faculty of Science and Technology
A doctoral dissertation from the Faculty of Science andTechnology, Uppsala University, is usually a summary of anumber of papers. A few copies of the complete dissertationare kept at major Swedish research libraries, while thesummary alone is distributed internationally throughthe series Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology.
Distribution: publications.uu.seurn:nbn:se:uu:diva-182900
ACTAUNIVERSITATIS
UPSALIENSISUPPSALA
2012
