Advances in the Carbonylation of Aryl Halides Using Palladium Catalysts

FIRST SOLVIAS SCIENCE DAY 684 CHIMIA 2001,55, No.9

Chimia 55 (2001) 684--687 © Schweizerische Chemische Gesellschaft ISSN 0009-4293

Matthias Bellera* and Adriano F. Indoleseb

Abstract: The palladium-catalyzed carbonylation of aryl halides is shown to be a versatile tool for the synthesis of various benzoic and heteroaromatic acid derivatives. Recent developments from our laboratories in this area are presented. Keywords: Benzoic acid derivatives· Carbonylation . Homogeneous catalysis· Palladium

Introduction From a general point of view the leav- Elegant work by researchers from ing group of the aryl-X derivative is for- Hoffmann-La Roche also demonstrated Palladium-catalyzed carbonylation reac- mally replaced by a nucleophile with in- the industrial applicability of such a reactions of aryl-X compounds leading to corporation of one or two molecules of tion in the commercial process for carboxylic acid derivatives were estab- CO (Scheme I). In addition to aryl-, hete- Lazabemide, a monamine oxidase B in- lished in the mid-seventies by the pio- ro-aryl-, vinyl-, allyl- und benzyl-X com- hibitor. In this process the aminocar- neering work of Heck and co-workers pounds can also serve as starting materi- bonylation of the commercially available [1]. Since that time these reactions have als in these carbonylations [2]. Two ex- 2,5-dichloropyridine with ethylenedi- found a number of applications in organ- amples for carbonylation reactions of ben- amine is performed with a Pd/dppp cata- ic synthesis, and even some industrial zyl-X applied on an industrial scale for lyst with comparably high catalyst pro- processes (see below) have been realized. fine chemical production are the synthe- ductivity (Scheme 3). However, compared to other palladium- sis of Ibuprofen (>3000 to/a) developed Due to the enhanced reactivity of the catalyzed coupling reactions such as the by Hoechst-Celanese (Scheme 2) [3] and 2-position compared to the 3-position of Heck and Suzuki reaction or the Buch- the carbonylation of 1,2-dixylyldicWoride the substituted pyridines, a selective wald-Hartwig ami nation reaction, palla- to give isochromanone by Clariant AG [4]. monocarbonylation is possible [5]. dium-catalyzed carbonylation still seems somewhat underdeveloped. This review gi ves an impression of the possibilities of o this catalytic multi-component coupling Pd, Co, Ni cat. ~X ~NU reaction. In addition some of the recent + CO + NuH orNuM + HXor MX achievements from our groups in this ~~ base ~~ R R area will be presented.

Nu = OH, OR, NR2, F, Cl, CN, H, alky~ ary~ alkinyl, RC02 'Correspondence: Prof. Dr. M. Beller alnstitut fUr Organische Katalyseforschung (IfOK) Scheme 1. Catalytic carbonylation of aryl-X compounds. an der Universitat Rostock e.V. Buchbinderstr. 5--6 0-18055 Rostock Tel.: + 49 381 466930 Fax: + 49 381 466 9324 OR E-Mail: [email protected] bDr. A.F. Indolese PdCl2/PPh3 Solvias AG WRO-l055.622 Klybeckstr. 191 ~ HCI, CO CH-4002 Basel Tel.: + 41 61 68666120 Fax. + 41 61 686 66100 E-Mail: [email protected] Scheme 2. Synthesis of Ibuprofen. FIRST SOLVIAS SCIENCE DAY 685 CHIMIA 2001.55, No, 9

leads to a significant increase of the cata- 0.1 mo1% PdCI2(MeCNh ° lyst productivity. Best turnover numbers 2equiv.dppp, 1.1 equiV.NaOMed' .••.•.•••••.•••NH CI NCI -.;;: N 3 (TON 6500) were achieved at 130°C in I '-': + = ------.•••- I .-9 H j).&' the presence of Pd(PPh3kcatalyst using Cl 10 bar CO, 110 °C Cl 75% 1.2 equiv. of NEt) as the base. 2. HCl, MeOH

Scheme 3. Carbonylation of 2,5-dichloropyridine leading to lazabemide. Carbonylation of Aryl Chlorides

Regarding starting materials, chloro- Depending on the nucleophile many It should also be mentioned that in- arenes are certainly the most attractive useful transformations like hydroxy-, tramolecular carbonylations (cyclocarbo- class of aryl halides for carbonylation re- alkoxy- and aminocarbonylations are nylation) reactions leading to various het- actions on an industrial scale, because possible. By far the most frequently em- erocycles are possible if the nucleophile they are inexpensive and readily availa- ployed transition metal catalysts are pal- is attached via a suitable spacer to the ble in bulk quantities with different sub- ladium compounds, but also much cheap- aryl-X substrate. This enables the direct stitution patterns. However, the high sta- er cobalt complexes (like CO2(CO)8) [6] synthesis of aryl-Iactones, -Iactams, -ox- bility of the aromatic carbon-chlorine and - in rare cases - nickel complexes azoles, -thiazoles, -imidazoles, etc. An- bond (bond dissociation energy for PhCI (like Ni(CNh) [7] have been used. Most other variant of the discussed carbonyla- at 298 K: 402 kllmol) remains a major of the work on carbonylation of aryl-X tion reactions is the introduction of two obstacle for their widespread use in this derivatives during the last 25 years con- CO molecules in a single step leading to reaction. Initial studies towards the car- centrated on the corresponding halides as derivatives of a-keto carboxylic acids. bonylation of aromatic chlorides em- starting materials, especially aryl bro- The reaction only takes place under ployed nickel and cobalt catalysts. mides and aryl iodides. Less frequently, specific conditions and is always in com- Hence, Ni(CO)4 was used to convert l- substrates with different leaving groups petition with monocarbonylation. As an and 2-chloronaphthalene into the corre- like diazonium salts, tnflates, alkyl- or example a variety of aryl and alkenyl bro- sponding naphthoic acids in the presence arylsulfonates, f1uorosulfonates, arylio- mides and iodides were converted into of Ca(OHh in dipolar aprotic solvents at donium salts, iodoxyarenes, or sulfonyl their corresponding a-keto amides with nO-l20 °C and atmospheric CO pres- chlorides have been successfully trans- good chemoselectivity by use of second- sure [12]. Another means of activating formed into the corresponding carbonyl ary amines and alkylphosphines or biden- aryl chlorides utilized a Ni/Pd system compounds. tate phosphines like dppb [10]. containing NaI. In this case a small The variety of nucleophiles that can Despite the synthetic usefulness of amount of the aryl chloride is converted be employed makes the carbonylation of the carbonylation reaction studies, the into the iodide by nickel catalysis and aryl-X derivatives especially valuable for development of more efficient catalysts subsequently reacted with palladium organic synthesis. Hence, a plethora of has been largely ignored. In this regard [13]. When using palladium catalysts, the different compounds can be synthesized we studied the alkoxycarbonylation of 4- oxidative addition of Pd(O) to the C-C1 from the same starting material. Water, bromoacetophenone as a model reaction bond is assumed to be the rate-determin- alcohols, primary and secondary amines, [11]. It was discovered that an excess of ing reaction step in the catalytic cycle carboxylates, even fluorides lead to car- phosphine ligand (P/Pd = 14; P = PPh3) (Scheme 4) [14]. boxylic acids, esters, amides, anhydrides, and acid fluorides, respectively. Addi- "Pd(II)" "Pd(O)" tionally, aldehydes, ketones, aroyl cya- [HBaselX ~ + 2L 2L/ nides, aroyl alkenes and aroyl acetylenes ie \ Red~ are accessible with suitable nucleophiles. +/ Even anionic metal complexes like Bas~ I L2Pd(O)1 ArX [CO(CO)4]- can serve as nucleophiles, leading to metal acyl complexes as prod- L ucts [8]. Recently, we were able to pre- I l~~ L pare amides with a free NH group by H-Pd-X I Pdx(CO)yLzI I 2 I Ar-Pd-X aminocarbonylation reaction of aryl hali- I des in high yields (70-90 %) using for- L mamide as the amine source. The reac- "'-rN"~L tions require a palladium catalyst in com- co bination with a nucleophilic Lewis base which acts as an acylating catalyst such o N\-H-,,\ as e.g. imidazole or 4-dimethylaminopy- L X ridine. Aryl, heteroaryl and vinyl bro- \ ,,/ mides and chlorides are converted to the \~ Pd Ar Ar-C/ "L L II" I primary amides under mild conditions (5 II "" Pd-CO bar, 120 0c) using I mol % of a palladi- o L"""""- I um phosphine complex. Best results were X obtained in dioxane using triphenylphos- phine as the ligand and DMAP as the Scheme 4. Proposed mechanism of the palladium-catalyzed alkoxycarbonylation of aryl-X base [9]. derivatives. FIRST SOLVIAS SCIENCE DAY 686

CHIMIA 2001. 55, No. 9

Two strategies have been pursued to Table 1. Butoxycarbonylation of various aryl chloridesa. increase the rate of this step: a) reduction of the 1t-electron density of the aromatic 0.5 mol% PdCI2(PhCN)2 chloride and b) enhancement of the nu- 3 equiv. Na2C03. molecular sieves cleophilicity of the palladium(O) catalyst ,(YCI by use of strongly basic ligands. The first ~~ + CO + nBuOH approach was successfully employed by 1 bar CO, 145°C, 16 h Basset et at. by converting the chloro- arenes stoichiometrically in their respec- R = 2-F, 4-CH3, 2-0Me, 3-OMe 2mol% tive tricarbonylchromium complexes. 4-CH2C02Et The various1l6-RC6H4C1 chromium 3-chloropyridine complexes (R = H, 4-Me, 4-0Me, 4-CF3) were carbony lated at 130 °C/25 bar to the corresponding esters or aldehydes in the Entry Aryl Chloride Conversion Product Yield Chemoselectivity [%] [%]b [%]C presence of 2 mol% PdCI2(PPh3)2/5PPh3 in moderate to good yields, the hydrocar- (Xco2nBu 100 bonylation being less effective [15]. (XCI 95 100 (76) The second approach [16] was em- """ F """ F ployed most successfully by Milstein and co-workers, who used a palladium cata- CI Dco2nBu 98 2 92 lyst based on the highly electron-rich and 100 (80) bulky chelating ligand 1,3-bis(diiso- D propylphosphino)propane (dippp). Thus, carboxylic acids, esters and amides were (Xco2nBu 95 3 (XCI 100 100 synthesized in excellent yields [17]. The (91) same catalyst also turned out to be active """ Ofv1e """ Ofv1e in the formylation of aryl chlorides to al- fv1e°Dco2nBu fv1e°DCI 95 dehydes using sodium formate as the re- 84 4 lOa (68) ducing agent. The drawbacks of this cata- """I """I lyst system, however, are the difficult synthesis and the high sensitivity of this ~C02nBu ~CI 93 pyrophoric phosphine along with the 5 >99 86d nBu02C """ I (68) comparatively low turnover numbers of Et02C """ I the catalyst (1 mol% of palladium need- • 7 mmol aryl chloride, 14 ml n-butanol. b Yield (GC) of all carbonylation products (ester + acid), GC-Yield of ed). The aminocarbonylation of activated product was determined by using diethyleneglycol di-n-butylether as internal standard. Isolated yield of product aryl chlorides proceeds with high yields in brackets. C Chemoselectivity = GC-yield of product/yield of carbonylation products. d Mixture of 80% in the presence of a palladium catalyst (4-carboxyphenyl)acetic acid and 20% (4-butoxycarbonylphenyl)acetic acid. based on 1,2-bis(diphenylphosphino) ethane, when sodium iodide is used to observed catalyst productivity is almost the appropriate ligand in the right con- activate the catalyst [18]. one order of magnitude higher than pre- centration at low catalyst loadings Very recently, we examined the palla- viously reported results with non-activat- (0.005-0.5 mol% Pd). dium-catalyzed alkoxycarbonylation re- ed chloroaromatics, reducing the other- Here catalyst turnover numbers of up action of activated and deactivated aryl wise high catalyst cost significantly. The to 13,000 were obtained for the carbon- chlorides with regard to the influence of applicability of the carbonylation proto- ylation of 2- or 4-substituted chloropyrid- critical reaction parameters, product se- col to a variety of aryl chloride substrates ines. The carbonylation of heteroaryl hal- lectivity, and the performance of various as well as other nucleophiles such as ides is of special interest to fine chemical catalyst ligands. Our investigations re- amines or water was demonstrated (Ta- synthesis, since there are many examples sulted in the development of a new effi- ble I). of heteroaromatic substructures in agro- cient catalyst system based on cyclo- It is interesting to note that chIoro- chemical and pharmaceutical agents. hexyl-substituted ferrocenylphosphine arenes activated by electron-withdrawing ligands for the carbonylation of aryl chlo- groups can be carbonylated (at reason- rides [19]. A considerable advantage is able temperatures and Pd concentrations) Conclusion that these ligands are air stable and com- in the presence of Pd catalyst systems mercially available. With the PdCl2 based on the PCY3 and similar ligands. It is obvious that organic halides in (PhCNh/4 catalyst (0.5 mol% Pd) in the Also more reactive heteroaryl chlorides comparison to toluenes are less attractive presence of sodium carbonate as a base, might be activated efficiently in the pres- starting materials for the industrial syn- chlorobenzene was converted to n-butyl ence of well-known 'standard ligands' thesis of bulk benzoic acid derivatives. benzoate in an essentially quantitative such as dppf (1,I'-bis(diphenylphos- Nevertheless, the application of the car- yield within 16 h (130-145 0C). The CO phino)ferrocene) or dppb (I ,4-bis (diphe- bonylation of aryl halides and similar pressure can be as low as I bar. After op- nylphosphino)butane) (Table 2) [20]. For starting materials is of importance for the timization, a catalyst turnover number of example good to excellent yields of synthesis of higher value fine chemicals almost 1600 was observed at a Pd loading a number of N-heterocyclic carboxylic and organic building blocks. In general it of only 0.05 mol%, underlining the high acid esters were realized by carbonyl- is nowadays possible by careful optimi- productivity of the catalyst system. The ation of N-heteroaryl chlorides applying zation of the catalyst system and the reac- FIRST SOLVIAS SCIENCE DAY 687

CHI MIA 2001, 55, No.9

a Table 2. Butoxycarbonylation of N-heteroaryl chlorides . [I] a) A. Schoenberg, 1. Bartoletti, R.F. Hcck, J. Org. Chern. 1974, 39, 3318; b) A. Schoenberg, R.F. Heck, J. Org. Chem. 0.1 mol% PdCI2(PhCNh 0.6 mol% dppf, 1.2 equiv. NEt3 1974, 39, 3327; c) A. Schoenberg, R.F. X.NyCI".;;: Heck, J. Am. Chem. Soc. 1974, 96, 7761. U/ ~ +CO+nBuOH ~ [2] Reviews: a) M. Beller, in 'Applied Homo- R'X 25 bar CO, 130°C, 15 h geneous Catalysis with Organometallic Compounds', Eds. B. Comils, W.A. Herr- X=CHorN mann, VCH, Weinheim, 1996, Vol. I, p. R = CF3, OMe, C], CH3 148; b) 1. Tsuji, 'Palladium Reagents and Catalysts: Innovations in Organic Synthe- sis', Wiley, Chichester, 1995; c) H.M. Colquhoun, OJ. Thompson, M.V. Twigg, Entry Substrate Product Yield[%]b 'Carbonylation, Direct Synthesis of Car- bony] Compounds', Plenum Press, New York, 1991. [3] a) M. Beller, B. Cornils, C.D. Frohning, (yCI (yca,"""~ I 95 C.W. Kohlpaintner, J. Mol. Cata/. 1995, ~ 104, 17; b) E. J. Jang, K. H. Lee, J. S. Lee, Y. G. Kim, J. Mol. Catal. 1999, 138, 25. 2 [4] a) H. Geissler, R. Pfirmann (Clariant), DE ~CI~ I ~ca,"""~ I 94 Patent 198]5323,1998; b) H. Geissler, R. Pfirmann (Clariant), EP Patent 1086949, F3C ~ F3C ~ 2000. [5] M. Scalone, P. Vogt, (Hoffmann La Roche) 3 EP 385210 A2, 1990; Chem. Abstr. 1991, ,,"a'(yci """'(yca,,,"" 83 114, 143153k. ~I ~I [6] M. Foa, F. Francalanci, E. Bencini, A. Gardano, J. Organornet. Chern. 1985,285, 4 293. [7] I. Amer, H. Alper, J. Org. Chern. 1988,53, (Xca,,,"" 73 ((' 5147. ~ CI ~ CI [8] Y. Misumi, Y. Ishii, M. Hidai, Chem. Lett. 1994,695. 5 [9] A. Schnyder, M. Beller, G. Mehltretter, T. (rCa,""" 82 Nsenda, M. Studer, A. F. Indolese, J. Org. (:( Chern. 2001, 66, 4311. N [10] a) F. Ozawa, H. Soyama, T. Yamamoto, A. Yamamoto, Tetrahedron Lett. 1982, 6 I 23, 3383; b) T. Kobayashi, M. Tanaka, J. eyca,,,"" 74 Organornet. Chern. 1982, 233, C64; c) F. eyC~ ~ ~ ~ Ozawa, T. Sugimoto, Y. Yuasa, M. San- tra, T. Yamamoto, A. Yamamoto, Orga- nometa/lics 1984, 3, 683. 7 [11] W. Magerlein, M. Beller, A.F.lndolese, J. ~ ...... N Mol. Catal. 2000, 156, 213. 83 [12] L. Cassar, M. Foa, J. Organol1let. Chel1l. ~ CI ~ I C0 nBu u u 2 1973, 51, 381. [13] J.J. Bozell, C.E. Vogt, J. Arn. Chern. Soc. '7.0 mmol N-heteroaryl chloride, 14 ml n-butanol, 1.2 equiv. NEt], 0.1 mol% PdCh(PhCNh, 0.6 mol% dppf, 1988,110,2655. 130 °C, 25 bar CO, 15 h. bisolated yield. []4] R.F. Heck, Adv. Catal. 1977,26,323. [15] a) R. Mulin, C. Lucas, 1. Thivolle-Cazat, V. Dufaud, F. Dany, J.-M. Basset, J. tion conditions to perform such reactions Acknowledgements Chern. Soc., Chem. Comrnun. 1988, 896; in an economically sensible way. Hence, We thank Dr. Wolfgang Magerlein (If OK), b) V. Dufaud, J. Thivolle-Cazat, J.-M. further applications of carbonylations are Dr. Gerald Mehltretter (If OK), Dr. Thomas Basset, R. Mathiew, Z. Jaud, J. Waisser- foreseen due the numerous possibilities Nsenda (Sol vias) for their excellent work in the mann, Organorneta/lics 1991, 10, 4005. for the selective synthesis of various aro- area of aryl-X carbonylations. We gratefully ac- [16] (a) M. Huser, M.-T. Youinou, J.A. Os- knowledge financial support of the Novartis born, Angew. Chern. Int. Ed. Eng. 1989, matic carboxylic acid derivatives and the Forschungsstiftung and from the State of Meck- often existing high pressure equipment in 28, 1386; b) V.V. Grushin, H. Alper, J. lenburg-West Pomerania. Drs. H.-V. Blaser, M. Chem. Soc., Chem. Coml1lun. 1992, 611; industry. Studer, A. Schnyder, F. Spindler (all So]vias c) T. Miyawaki, K. Nomura, M. Hazama, Apart from this aspect, there are inter- AG) are thanked for helpful discussions. We G. Suzukamo, J. Mol. Cata/. 1997, 120, esting research goals for academia and thank Mrs. M. Heyken (IfOK) for her excellent L9. industry. Further catalyst improvements technical assistance and Mrs. S. Buchholz [17] Y. Ben-David, M. Portnoy, D. Milstein, J. (IfOK) for her analytical services. are desirable, and regarding reductive Am. Chern. Soc. 1989,11 1,8742. carbonylation (synthesis of aldehydes) [18] RJ. Perry, B.D. Wilson, J. Org. Chem. there is a need for more general methods. Received: July 13,2001 1996,61, 7482. [19] W. Magerlein, A.F. Indolese, M. Beller, In addition, efficient asymmetric carbon- Angew. Chern. Int. Ed. 2001,40, 2856. ylations of benzyl- and allyl-X deriva- [20] M. Beller, W. Magerlein, A.F. Indolese, tives have not achieved so far. C. Fischer, Synthesis 2001, 1098.