New Catalysts and New Synthetic Applications for Hydroformylation

DGMK/SCI-Conference „Synthesis Gas Chemistry", October 4-6, 2006, Dresden, Germany

New Catalysts and New Synthetic Applications for Hydroformylation Bernhard Breit Albert-Ludwigs-Universitat Freiburg, Institut fur Organische Chemie & Biochemie, Germany

Abstract In the course of this lecture most recent advances in rhodium catalyzed hydroformylation and its use in organic synthesis are presented. Particular emphasis is given to regioselective hydroformylation of terminal alkenes and its application to fine chemical synthesis as well as latest results and applications of asymmetric hydroformylation. Furthermore, a new concept for simultaneous control of regio- and stereochemistry employing catalyst-directing groups will be discussed in detail. Finally, a new concept for catalyst library generation based on ligand-self-assembly through complementary hydrogen bonding and its application to regioselective hydroformylation as well as asymmetric hydrogenation is presented.

Introduction The hydroformylation of olefins, which was discovered as early as 1938 by Otto Roelen, is today one of the largest industrially applied processes which relies on homogeneous catalysis (Scheme 1).1 More than 8 million tons of oxo products are produced per year of which the majority stems from propene hydroformylation with butanal and butanol, respectively being the major products.2 These commodities end up primarily as butyl acrylate which is an important comonomer for the production of polymer dispersions.

catalyst CHO CO/H2 R^ Me R^ .CHO R +

linear branched Scheme 1

In addition to this industrial aspect, the hydroformylation of olefins is a very attractive transformation for organic synthesis. Thus, addition of CO and H2 to an alkene function provides a new carbon carbon and a new carbon hydrogen bond. Both bond forming events could eventually lead to the formation of a new stereocenter. Simultaneously, the reaction introduces the synthetically useful aldehyde function, which elaborates the product for additional carbon skeleton expanding operations. The reaction requires only catalytic amounts of a late transition metal catalyst, with rhodium(I) complexes being the most active and selective catalysts for this reaction, and all atoms of the starting materials remain incorporated into the product. Hence, this reaction is a prototype of an atom economic transformation with all its associated economic and environmental advantages. 3 From a synthetic aspect it is interesting to note, that a benefit of the alkene function is its inertness to a large set of reagents and conditions which allows this functionality to be carried through a number of steps in a synthetic sequence, until the final one carbon chain elongation via hydroformylation is desired. However, despite these advantages and contrary to its industrial importance the hydroformylation has not been of frequent use in organic synthesis yet. This discrepancy is primarily due to the difficulty to control selectivity issues such as regioselectivity,

DGMK-Tagungsbericht 2006-4, ISBN 3-936418-57-8 33 diastereoselectivity and enantioselectivity throughout the course of the hydroformylation reaction. 4

Regioselective hydroformylation of terminal alkenes Much progress has been made on regioselective hydroformylation of terminal alkenes in favor of the linear product. In particular bidentate phosphine or phosphite ligands which have a natural bite angle of about 110 degrees will favor the linear product. The most successful ligand types are BISBI,5 BIPHEPHOS6 and XANTPHOS7 systems (Scheme 2).

MeO OMe

Ph 2P PPh 2 0 PPh 2 PPh 2

BISBI BIPHEPHOS XANTPHOS l:b = 66 : 1 for 1 -hexene l:b > 40 : 1 for a wide range of alkenes l:b = 53 : 1 for 1-octene

R'2^22Rv

0 = Natural Bite Angle Scheme 2

In particular the BIPHEPHOS system has found application in organic synthesis. Particularly attractive are Tandem type transformations in which the aldehyde as the hydroformylation product undergoes in situ further reactions with appropriate nucleophiles. Recent examples are the synthesis of conformationally restricted dipeptide surrogates employing Tandem hydroformylation intramolecular aminal formation (Scheme 3).8

[Rh(CO) 2acac] CO/H2 4 bar, C6H6, 60 C

BIPHEPHOS BocHN O COOMe O COOMe S:Rh:L (50:1:2)

(90%) Scheme 3

Another example that includes an additional skeleton extension is the Tandem regioselective hydroformylation/Knoevenagel condensation/hydrogenation reaction. Here, the rhodium catalyst employed in this reaction plays a dual role. First, it functions as a hydroformylation catalyst and second, after Knoevenagel condensation, as a hydrogenation catalyst for the electron poor bis-acceptor substituted alkene (Scheme 4).9

[Rh(CO) 2acac] CO/H2 20 bar, Me Toluene, 60 C Me Piperidine/AcOH (0.2 equiv.) OH TBSO O BIPHEPHOS EtO ^O EtO S:Rh:L (100:1:4) rs (>99:<1) (64%) Scheme 4

34 Asymmetric hydroformylation Many chiral diphosphine ligands have been evaluated with regard to induce enantioselectivity in the course of the hydroformylation reaction. 4 However, a real breakthrough occured not earlier than 1993 with the discovery of the BINAPHOS ligand by Takaya and Nozaki. 10 This was the first efficient and rather general catalyst for the enantioselective hydroformylation of several classes of alkenes such as arylalkenes, 1- heteroatom-functionalized alkenes and substituted 1,3-dienes and is still a benchmark in this area. But still a major problem in this field is the simultaneous control of enantio- and regioselectivity which limits the structural variety of suitable alkenes for enantioselective hydroformylation significantly (Scheme 5). 4 Rh(I)/L* Me CO/H2 fgr"'^ FG R FGR/Sl L* = O (R,S)-BINAPHOS linear branched

12 88 (94% ee)

79 21 (83% ee)

14 86 (92% ee)

13 87 (96% ee)

Second generation BINAPHOS type ligands have been developed recently.11 A new structural feature is the replacement of the phosphite donor by a phosphoramidite system. Improved enantioselectivities were noted albeit the problem of regioselectivity persists (Scheme 6).

[Rh(CO) 2acac] CO/H2 20 bar, C6H6, 60 C Me Ar^CHO ArJ (R, S)-P-Amidite-BINAPHOS

S:Rh:L (1000:1:4) rs 89 : 11 99% ee

(R,S)-P-Amidite-BINAPHOS

Scheme 6

Recently, a structurally new and effective chiral ligand system has been discovered which is based on a bisdiazaphospholane system (Scheme 7).12 However, so far enantioselectivities achieved with this system do not exceed those achieved with the BINAPHOS class of ligands.

35 [Rh(CO) 2acac] CO/H2 10 bar, toluene, 80 C Me A. AcOJ Bis-3,4-diazaphospholane AcO * CHO

S:Rh:L (5000:1:1.2) rs 37 : 1 96% ee

O O

O Me R = V^^Ph

Bis-3,4-diazaphospholane

Scheme 7

The major problem remains control of regioselectivity in favor of the branched regioisomer. While aryl alkenes as well as heteroatom-substituted alkenes favor the chiral branched isomer, for aliphatic alkenes such an intrinsic element of regiocontrol is not available. As a matter of fact branched selective and asymmetric hydroformylation of aliphatic alkenes stands as an unsolved problem. In this respect regio- and enantioselective hydroformylation of allyl cyanide employing a chiral bisphosphite(Kelliphite)/rhodium catalyst is a remarkable result (Scheme 8).13 After successive reduction of aldehyde and nitrile function the resulting amino alcohol could serve as a potential intermediate for the construction of TAK-637, a compound in development of Takeda Chemical Industries for urinary continence.

[Rh(CO) 2acac] 1) Pt/C, H2, MeOH OH CO/H2 10 bar, 2) Pt/C, H2, D-tartaric acid CHO 3) recrystallisation ,^CN CN toluene, 45 C — Me Me NH3 Tartrate (R,R)-Kelliphite rs 23:1 S:Rh:L (3000:1:1.1) 81% ee

Scheme 8

The asymmetric hydroformylation of a 1,3-diene has been recently used in the course of a total synthesis of the antifungal natural product Ambruticin. The retrosynthesis as well as the hydroformylation key step are depicted in Scheme 9.14

36 Ambruticin (antifungal)

Julia-Kocienski-Olefination

[Rh(CO) 2acac] CO/H2 20 bar, 1 4

5 0HC-v/^-0^Me C6 H6 , 30-35 C

^ Me (S.R)-BINAPHOS Me Me

S:Rh:L (200:1:4) iso.n = 91:9 (90%) dr= 96:4

Scheme 9

Selectivity control employing catalyst-directing groups Although significant progress has been made in the field of asymmetric hydroformylation, it is limited to a narrow substrate scope. 4 An alternative approach to a stereoselective hydroformylation might employ substrate control of a chiral alkenic starting material. Of particular use have proved directed hydroformylation variants making use of attractive substrate-catalyst interactions by way of substrate-bound catalyst-directing groups (Scheme 10) ,15 '16

= o-DPPB

Scheme 10

Thus, transformation of an allylic or homoallylic alcohol into a corresponding ortho- diphenylphosphanylbenzoate ester (o-DPPB) allows the regio- and stereoselective hydroformylation of a wide range of substrates in good to excellent levels of diastereoselectivity (Scheme 11).17

O(oDPPB) (o-DPPB)O Me O

R' O Hydroformylation S 99 : 1 96 : 4

O(oDPPB) (o-DPPB)O Me O

Met O D 90 : 10 Het = SiR'3, OR1, NPhth S95 : 5

(o-DPPB)O O PGO O(oDPPB) PGO O(oDPPB)

Me Me O Me Me O 95 : 5 96 : 4 dr S 98:2; rs S 98:2 Scheme 11

37 When chiral catalyst-directing groups such as the o/?/70-diphenylphosphanylferrocene- carboxylate (o-DPPF) was employed, a desymmetrizing directed hydroformylation of bis- alkenyl- and bis-allyl carbinols could be achieved with excellent chemo-, regio- and stereoselectivity (Scheme 12).18

(CDG*) (CDG*) R O R Rh(l) R O R O CO/H2

n = 0, 1

dr up to 99:1 ee > 99%

(Sp)-oDPPF and (Rp) Scheme 12

Self-Assembly of monodentate to bidentate ligands - a combinatorial approach to new catalysts Bidentate ligands are important for selectivity control in homogeneous metal complex catalysis. However, the quest for the ultimate ligand which gives a catalyst with optimal activity and selectivity is difficult. Since rational design still does not allow the ligand of choice for a given reaction and substrate to be predicted, the combinatorial synthesis of ligand libraries and their subsequent use has become an additional strategy. However, the rate determining step in catalyst development is in most cases the time-consuming ligand synthesis required to generate the library. An alternative way for the generation of a bidentate ligand makes use of a self-assembly process of monodentate to bidentate ligands employing hydrogen-bonding (Scheme 13).19 Thus, mixing two equivalents of 6-diphenylphosphinopyridone 1 with a transition metal salt lead to the formation of complex 2. The bidentate nature of 1 in these complexes has been proven in solution (NMR) as well as in crystalline state (X-ray). Interestingly, rhodium complexes derived from 1 displayed excellent regioselectivity and actitivity upon hydroformylation of terminal alkenes. 20 These catalysts allowed the first room temperatur ambient pressure regioselective hydroformylation which is of particular use to synthetic organic chemistry. 21

D: PPh 2 M: Pt(ll), Rh(l) Scheme 13

Since both tautomers of 1 - the hydroxypyridine 1A and the pyridone 1B - are enrgetically almost equivalent and show rapid equilibration, mixing of two different ligands 1 would furnish a mixture of the two homodimeric and the heterodimeric ligand complex. In order to generate selectively the unsymmetrical heterodimeric ligand we developed a new self- assembly platform based on the Watson-Crick base pairing of A and T in DMA. As an A-T base pair analogue, the aminopyridine (3)/isoquinolone (4) platform was selected. Thus, mixing of two monodentate ligands based on this platform in the presence of a transition

38 metal salt lead to the selective formation of the heterodimeric complex featuring a bidentate coordination mode (Scheme 14).22

(n) 4 (m ■ n) 5 Scheme 14

A library of 4x4 ligands has been screened for regioselective rhodium-catalyzed hydroformylation. From this catalyst library a system with outstanding activity and regioselectivity was identified (Scheme 15). 22

(4 (4x4)

Dx, Dy = PAr2 Scheme 15

The concept is of course very general and not restricted to hydroformylation. In principle this strategy should allow the rapid generation of complex bidentate ligand libraries which could be screened against any selectivity issue of interest in the course of a metal complex catalyzed reaction. Of particular interest is of course the control of enantioselectivity and hence, the extension of this concept towards asymmetric catalysis. Thus, a new library of chiral ligands based on the A-T model platform 3 and 4 was synthesized with chirality information being attachted to the P-donor function. From this library of 80 catalysts (8x10 ligands) a cationic rhodium catalyst that gave an enantiomeric excess of 99% ee upon hydrogenation of acrylic acid derivatives at a catalyst loading of 0.1% could be identified (Scheme 16). 23

(8) (70x8)

Dx, Dy = PAnArz, P(OR)2

Scheme 16 Thus, application of this concept to asymmetric catalysis is possible and ongoing in these laboratories.

39 Acknowledgements Financial support by DFG, the Fonds der Chemischen Industrie, BASF AG as well as the Krupp foundation is acknowledged. The author is indebted to his coworkers for their individual contributions.

References

1) O. Roelen (Chemische Verwertungsgesellschaft, mbH Oberhausen) German Patent DE 849,548, 1938/1952; Roelen, O. (Chemische Verwertungsgesellschaft, mbH Oberhausen) U.S. Patent 2,317,066 (1943); Chem. Abstr. 1944, 38, 550. 2) K. Weissermel, H.-J. Arpe, Industrial Organic Chemistry, Wiley-VCH, Weinheim, 2003; Chapter 6. 3) B. M. Trost, Science 1991, 254, 1471-1477; idem, Angew. Chem. Int. Ed. Engl. 1995, 34, 259-281. 4) B. Breit, W. Seiche, Synthesis 2001, 1-36. 5) C. P. Casey, G. T. Whiteker, Isr. J. Chem. 1990, 30, 299-304. 6) E. Billig, A. G. Abatjoglou, D. R. Bryant (UCC) U.S. Patent 4,769,498, 1988; Chem. Abstr. 1989, 111, 117 287; G. D. Cunny, S. L. Buchwald, J. Am. Chem. Soc. 1993, 115, 2066-2068. 7) M. Kranenburg, Y. E. M. van der Burgt, P. C. J. Kamer, P. W. N. M. van Leeuwen, K. Goubitz, J. Fraanje, Organometallics 1995, 14, 3081-3089; C. J. P. Kamer, P. W. N. M. van Leeuwen, J. N. H. Reek, Acc. Chem. Res. 2001, 34, 895-904. 8) I. Ojima, Org. Lett. 2002, 4575-4578. 9) B. Breit, S. K. Zahn, Angew. Chem. Int. Ed. 2001, 40, 1910-1913. 10) N. Sakai, S. Mano, K. Nozaki, H. Takaya, J. Am. Chem. Soc. 1993, 115, 7033; K. Nozaki, I. Ojima, Catalytic Asymmetric Synthesis; Ojima, I., Ed.; Wiley-VCH: New York, 2000; Chapter 7. 11) Y. Yan, X. Zhang, J. Am. Chem. Soc. 2006, 128, 7198-7202. 12) T. P. Clark, C. R. Landis, S. L. Freed, J. Klosin, K. A. Abboud, J. Am. Chem. Soc. 2005, 127, 5040-5042. 13) C. J. Cobley, K. Gardner, J. Klosin, C. Praquin, C. Hill, G. T. Whiteker, A. Zanotti- Gerosa, J. L. Petersen, K. A. Abboud, J. Org. Chem. 2004, 69, 4031-4040. 14) P. Liu, E. N. Jacobsen, J. Am. Chem. Soc. 2001, 123, 10773-10774. 15) B. Breit, Chem. Eur. J. 2000, 6, 1519-1524. 16) B. Breit, Acc. Chem. Res. 2003, 36, 264-275. 17) B. Breit, P. Demel, A. Gebert, Chemical Communications 2004, 114-115; B. Breit, G. Heckmann, S. K. Zahn, Chem. Eur. J. 2003, 9, 425-434; B. Breit, S. K. Zahn, J. Org. Chem. 2001, 66, 4870-4877; B. Breit, M. Dauber, K. Harms, Chem. Eur. J. 1999, 5, 2819-2827; B. Breit, S. K. Zahn, Angew. Chem. Int. Ed. 1999, 38, 969-971; B. Breit, Eur. J. Org. Chem. 1998, 1123-1134; B. Breit, Liebigs Ann./Recueil 1997, 1841-1851; B. Breit, Angew. Chem. Int. Ed. Engl. 1996, 35, 2835-2837. 18) B. Breit, D. Breuninger, J. Am. Chem. Soc. 2004, 126, 10244-10245; B. Breit, D. Breuninger, Eur. J. Org. Chem. 2005, 3916-3929; B. Breit, D. Breuninger, Eur. J. Org. Chem. 2005, 3930-3941. 19) B. Breit, Angew. Chem. Int. Ed. 2005, 44, 6816-6825. 20) B. Breit, W. Seiche, J. Am. Chem. Soc. 2003, 125, 6608-6609. 21) W. Seiche, A. Schuschkowski, B. Breit, Adv. Synth. Cat. 2005, 1488-1494. 22) B. Breit, W. Seiche, Angew. Chem. Int. Ed. 2005, 44, 1640-1643. 23) M. Weis, C. Waloch, W. Seiche, B. Breit, J. Am. Chem. Soc. 2006, 128, 4188-4189.