大规模生产--氧化反应综述

Large-Scale Oxidations in the Pharmaceutical Industry† Ste´phane Caron, Robert W. Dugger, Sally Gut Ruggeri, John A. Ragan, and David H. Brown Ripin* Chemical Research and Development, Pfizer Global Research Division, Pfizer Inc., Eastern Point Road, Groton, Connecticut 06340 Received August 1, 2005 Contents 1. Introduction 2944 2. Oxidation of Carbon−Carbon Bonds 2946 2.1. Oxidative Cleavage of Olefins 2946 2.1.1. Ozonolysis 2946 2.1.2. Periodate, Peroxide, and Other Reagents 2947 2.2. Dihydroxylation of Olefins 2949 2.2.1. Asymmetric Dihydroxylations 2949 2.2.2. Non-Asymmetric Dihydroxylations 2950 2.3. Epoxidation of Olefins 2951 2.3.1. Asymmetric Epoxidations (Including Kinetic Resolutions) 2951 2.3.2. Non-Asymmetric Epoxidations 2952 2.4. Halohydrin Formation 2953 2.5. Furan Oxidations 2954 2.6. Hydroboration of Olefins 2955 2.7. Oxidative Rearrangements 2955 2.7.1. Curtius Rearrangements 2956 2.7.2. Hofmann Rearrangements 2956 2.7.3. Lossen Rearrangement 2957 2.7.4. Beckmann Rearrangement 2957 2.7.5. Baeyer−Villiger Reaction 2958 2.7.6. Other Oxidative Rearrangements 2959 2.8. Aromatic Ring Oxidation 2960 3. Oxidation of Carbon−Hydrogen Bonds 2961 3.1. Oxidation R to a Carbonyl 2961 3.1.1. By Halogen 2961 3.1.2. By Oxygen 2963 3.1.3. By Nitrogen 2963 3.1.4. R,â Unsaturation 2964 3.2. Benzylic and Allylic Oxidations 2965 3.2.1. Halogenations 2965 3.2.2. Oxidations 2965 3.3. Oxidation R to a Heteroatom 2966 3.4. Aromatization 2966 4. Oxidation of Nitrogen and Carbon−Nitrogen Bonds 2967 4.1. Diazotization 2967 4.1.1. Diazotization Followed by Reduction to C−H 2967 4.1.2. Conversion of Amino Group to Halide via Diazotization 2967 4.1.3. Diazotization Followed by Cross-Coupling 2967 4.1.4. Diazotization Followed by Carbonylation or Sulfonylation 2968 4.1.5. Conversion of an Amine to an Alcohol via Diazotization 2968 4.1.6. Diazotization and Reduction to the Hydrazine 2968 4.1.7. Preparation of Diazo Dyes 2969 4.2. Oxidative Cyclization of O-Nitroanilines 2970 4.3. Pyridine N-Oxides 2970 4.4. Amination of Nitrogen 2971 4.5. Halogenation of Amines 2971 4.6. Nitrone Formation 2972 4.7. Nitrile Oxides from Aldoximes 2972 4.8. Imine Oxidation 2972 4.9. Oxidation of Amines to Imines 2972 5. Oxidation of Carbon−Oxygen Bonds 2973 5.1. Oxidation of Primary Alcohols to Aldehydes 2973 5.1.1. Metal-Mediated Processes 2973 5.1.2. Moffatt and Modified-Moffatt Processes 2973 5.1.3. TEMPO-Mediated Processes 2974 5.2. Oxidation of Secondary Alcohols to Ketones 2975 5.2.1. Metal-Mediated Processes 2975 5.2.2. Moffatt and Modified-Moffatt Processes 2976 5.2.3. TEMPO-Mediated Processes 2978 5.2.4. Alternative Processes 2978 5.3. Oxidation of Benzylic and Allylic Alcohols 2979 5.3.1. MnO2 Oxidation 2979 5.3.2. DDQ Oxidation 2979 5.4. Oxidation to Carboxylic Acids and Derivatives 2979 5.4.1. Metal-Mediated Oxidations of Aldehydes to Carboxylic Acids and Derivatives 2979 5.4.2. Hydrogen Peroxide Oxidation of Aldehydes to Carboxylic Acids and Derivatives 2979 5.4.3. Sodium Chlorite Oxidation of Aldehydes to Carboxylic Acids and Derivatives 2979 5.4.4 TEMPO/Sodium Chlorite Oxidation of Alcohols to Carboxylic Acids and Derivatives 2980 5.4.5. Metal-Mediated Oxidation of Alcohols to Carboxylic Acids and Derivatives 2980 6. Oxidation of Sulfur 2980 6.1. Oxidation of a Sulfide to a Sulfoxide 2980 6.1.1. Peroxide-Based Reagents 2980 6.1.2. Peracid Oxidations 2980 6.1.3. Inorganic Oxidants 2981 6.1.4. Stereoselective Oxidations 2981 6.2. Oxidation of a Sulfide to a Sulfone 2982 6.2.1. Peroxide-Based Reagents 2982 † Dedicated to Dr. Frank J. Urban, a source of inspiration to the process research group at Pfizer, on the occasion of his retirement. * Correspondingauthor.Fax: (860)441-3630.E-mail: david.b.ripin@pfizer.com. 2943Chem. Rev. 2006, 106, 2943−2989 10.1021/cr040679f CCC: $59.00 © 2006 American Chemical Society Published on Web 06/27/2006 6.2.2. Peracid Oxidations 2983 6.2.3. Inorganic Oxidants 2983 6.3. Oxidation of a Sulfide to a Sulfonic Acid or Sulfonamide 2984 6.3.1. Peroxide-Based Oxidations 2984 6.3.2. Chlorine Oxidations 2984 6.4. Oxidation To Form a Sulfur-Containing Heterocycle 2984 6.5. Oxidation of a Sulfide to a Disulfide 2984 6.5.1. Peroxide-Based Oxidations 2984 6.5.2. Oxidations with Oxygen 2984 7. References 2985 1. Introduction Oxidation reactions are powerful tools to convert a position that is protected in a lower oxidation state to the desired functionality and for the functionalization of otherwise unfunctionalized positions. Yet despite their power as a synthetic tool and abundant use in academic research, oxidation reactions as a whole comprise as little as 3% of the reactions used on a preparative scale in the pharmaceuti- cal industry.1 This disparity is likely due to a mixture of factors. To streamline production processes, every effort is made to develop routes that introduce functionality in the correct oxidation state and without protection. While the byproducts of many oxidants are fairly environmentally benign, many of the more selective reagents produce undesirable waste products. Perhaps the greatest factor influencing the hesitation to employ oxidation reactions on a large scale is the safety of these processes. The majority of reactions run in a production facility are in flammable organic solvents, and on a large scale, the potential for static electric discharge in the solvent charge lines or the reactor itself is significant. In the case of reactions other than oxidations, fires and explosions are prevented by starving the solvent and reactants of oxidants. In the case of oxidation reactions, this safety factor is absent and all three elements Ste´phane Caron was born in 1967 in L’Isle-Verte, Quebec, Canada. He received a B.Sc. and a M.Sc. at l’Universite´ Laval under the supervision of Robert H. Burnell. He completed his Ph.D. on the total synthesis of zaragozic acid A with Clayton H. Heathcock at the University of California at Berkeley. He joined Pfizer in 1995 as a process research chemist and is currently a Director in Chemical Research and Development. He previously served on the grant selection committee of the inorganic and organic division of the National Science and Engineering Research Council of Canada. Bob Dugger was born in St. Louis, Missouri. He received his B.S. from the University of Missouri, St. Louis, in 1975. He received his Ph.D. from the University of California, Berkeley, in 1980, working with Professor Clayton Heathcock. He then joined the faculty at Miami University (Ohio). In 1986, he left Miami and joined the Process Research group at Pfizer in Groton, CT. Sally Gut Ruggeri was born in San Francisco, California, in 1960. She received her A.B. from Cornell University in 1982, carrying out research under the direction of Dave Collum. She briefly carried out research at the Scripps Oceanographic Institute with Bill Fenical prior to moving to the Unviersity of WisconsinsMadison, where she received her Ph.D. in 1987 with Barry Trost. After a National Cancer Institute postdoctoral fellowship at Stanford University with Paul Wender, she joined the Pfizer Process Research Group in 1989. John Ragan was born in 1962 in Kansas City, Missouri. He received his S.B. degree in chemistry from MIT (1985), where he did undergraduate research with Rick Danheiser. He completed his Ph.D. with Stuart Schreiber at Yale (1985−1988) and Harvard Universities (1988−1990), working on the total synthesis of FK506. Following an American Cancer Society postdoctoral fellowship with Clayton Heathcock at the University of CaliforniasBerkeley, he joined Pfizer in 1992, initially in Discovery Chemistry, moving to Chemical Research & Development in 1996, where he is an Associate Research Fellow. He is currently serving on the Board of Editors of Organic Syntheses. 2944 Chemical Reviews, 2006, Vol. 106, No. 7 Caron et al. required for combustion (fuel, energy, and oxidant) are present in the reactor. While these issues can be addressed in order to run oxidations routinely and safely on scale, common errors that can occur in a manufacturing environ- ment, such as charging reagents out of order or too rapidly, must be taken into consideration when designing a safe process. These considerations, in addition to the highly reactive nature of many of the oxidants and thermal instabil- ity of some products (such as diazo compounds, N-oxides, and the like) likely limit the use of oxidative processes in production syntheses. Caution: Many of the oxidants, oxidation reactions, and products described herein haVe the potential to release large amounts of energy in an uncontrolled fashion. InVestigators considering running a large-scale oxidation reaction should consult the literature, run appropriate safety tests, and take proper precautions when running the reaction. Despite the challenges, oxidation reactions are routinely run in production facilities to make many of the commercial pharmaceuticals available today. Some classes of drugs require a high number of oxidations to be produced, such as steroids and prostaglandins. A particularly oxidation-rich process depicted in Scheme 1 is the Merck process for converting desoxycholic acid (1) to cortisone acetate (15). This process was run on a multihundred kilogram scale in the 1950s and 1960s and utilized 11 separate oxidations.2 This review covers oxidation reactions run in 1980 or later, on a scale of around 100 g or larger (as demonstrated in the experimental section of the publication), or clearly developed by a process chemistry group to be run on a large scale. Excluded from discussion are biotransformations and oxida- tive aromatic substitution reactions (such as nitrations and halogenations of aromatic rings). The review is divided into sections based on the functional groups being oxidized, allowing for easy comparison of reagents and substrates used in various transformations. David Ripin was born in 1970 in Silver Spring, Maryland. He received his B.A. in 1988 from Washington University in St. Louis, carrying out research under the direction of Professor Kevin D. Moeller. David then received his Ph.D. in 1997 with Professor David A. Evans (Harvard University) after completing the total synthesis of miyakolide. David joined the Pfizer Process Research Group in 1997, where he continues research. Scheme 1. Merck Process for Cortisone Acetate (15)a a (a) CrO3; (b) (i) Br2; (ii) NaOAc; (c) Br2; (d) Na2Cr2O7, CrO3; (e) (i) HBr; (ii) Ac2O; (f) (i) NBS, hî; (ii) heat; (g) Na2Cr2O7, H2SO4; (h) monoperphthalic acid, NaOH; (i) Br2, KOAc; (j) dibromodimethyl hydantoin; (k) Br2. Large-Scale Oxidations in the Pharmaceutical Industry Chemical Reviews, 2006, Vol. 106, No. 7 2945 2. Oxidation of Carbon−Carbon Bonds Oxidation of carbon-carbon bonds is frequently utilized in the synthesis of pharmaceutical agents, because the resulting functionalities (e.g., vicinal diols, epoxides, carbonyl compounds) provide valuable synthetic intermediates and are themselves frequently present in the active pharmaceutical ingredient (API). Asymmetric variants of several oxidations (e.g., epoxidations, dihydroxylations) have found numerous applications as exemplified in this section. 2.1. Oxidative Cleavage of Olefins 2.1.1. Ozonolysis There are several examples of oxidative cleavage of olefins by ozonolysis. Hansen and co-workers utilized a chiral- auxiliary-based Diels-Alder cycloaddition to generate bi- cyclic olefin 17, which was cleaved by ozonolysis to generate diol 18 following reductive workup (Scheme 2).3 The suitability of this route for further scale-up was suggested. Further transformations converted this diol to cis-perhy- droisoquinoline LY235959 (19), an N-methyl-D-aspartate (NMDA) receptor antagonist. The scale-up of the ozonolysis of olefin 20 (Scheme 3) has been described.4 The primary ozonide was trapped by methanol to generate the methoxy-hydroperoxide,5,6 which was treated with aqueous sodium bisulfite (NaHSO3) to effect simultaneous peroxide reduction and bisulfite formation to generate 21 (57% yield on 2.3 kg scale). This bisulfite adduct could be used directly in a reductive amination to generate amines such as 22. Varie has described the conversion of (R)-carvone to protected alcohols 23 (Scheme 4, R ) TBS, TBDPS, or CO- t-Bu).7 The propenyl side chain was then cleaved by ozonolysis followed by Criegee rearrangement of the inter- mediate methoxy-hydroperoxide to generate the acetate of the desired alcohols (24).8 McWhorter described the ozonolysis of olefin 27 to prepare alcohol 28 (Scheme 5), which was converted to diamine 29, an intermediate in the synthesis of premafloxacin (30).9 Interestingly, this reaction was executed in water. Several related ozonolyses were also described. The olefin substrates were prepared by asymmetric Michael addition of chiral benzylic amines to esters of crotonic acid. Workers at Sumitomo reported ozonolytic cleavage of indole 31 and cyclization of the resulting keto-amine to generate benzodiazepine 32 (Scheme 6, 72% yield on 250 g scale).10 Chromium trioxide also effected this oxidative cleavage, but with lower efficiency (28%). Kleinman utilized ozonolysis of bicyclic carbamate 33 to generate bis-aldehyde 34, which was treated with benzyl- amine and sodium cyanoborohydride to generate the bicyclic amine 35 (Scheme 7).11 Researchers at Lilly utilized ozonolysis followed by NaBH4 workup to provide diol 37, an intermediate in the synthesis of protein kinase C (PKC) inhibitors such as 38 (Scheme 8).12 Alcohol 40 was prepared from olefin 39 by ozonolysis followed by NaBH4 workup (Scheme 9, ca. 89% yield, 288 Scheme 2. Olefin Ozonolysis in the Synthesis of LY235959 (19) Scheme 3. Ozonolysis Followed by in-Situ Bisulfite Adduct Formation Scheme 4. Ozonolysis Followed by Criegee Rearrangement Scheme 5. Synthesis of Premafloxacin (30) Scheme 6. Ozonolytic Cleavage of an Indole Followed by Cyclization To Generate a Benzodiazepine Scheme 7. Conversion of 33 to 35a a (a) (i) O3; (ii) Me2S; (b) BnNH2, sieves, NaCNBH3. 2946 Chemical Reviews, 2006, Vol. 106, No. 7 Caron et al. g scale; yield estimated from a 51% overall yield for a six step sequence).13 It is noteworthy that the least substituted olefin was selectively oxidized (the proximity of the electron- withdrawing sulfone to the other two olefins may also contribute to this selectivity). Sulfone 40 was a precursor to 24(S)-hydroxyvitamin D2, a metabolite of vitamin D2. Ozonolysis of tetrasubstituted enamide 41 has been employed to generate R-hydroxyester 43 (Scheme 10, PNB ) p-nitrobenzyl). This sequence was scaled to 250 kg and proceeded in 70-75% yields. Although the intermediate R-dicarbonyl lactam 42 was not isolated, its direct formation from the ozonolysis (i.e., prior to addition of any reducing agent) indicated that the initial ozonide breaks down to generate this product plus the carbonyl oxide of acetone.14,15 Scheme 11 summarizes five further examples (44-48), all of which were described by process research groups on laboratory scale for the conversion of terminal olefins to aldehydes (44, 47),16 a 1,2-disubstituted olefin to aldehyde (46),17 a 1,1,2-trisubstituted olefin to an aldehyde (45),18 and a 1,1-disubstituted olefin to a secondary alcohol (48).19 Interestingly, in this last example the product chiral alcohol was C2-symmetric, such that the secondary carbinol generated in the oxidation was a chirotopic, nonstereogenic center.20 2.1.2. Periodate, Peroxide, and Other Reagents Electron-rich olefins (e.g., enamines) can be oxidatively cleaved by treatment with periodate reagents. A strategy based on this oxidation was developed by Coe and co- workers at Pfizer for oxidation of activated aromatic methyl groups (e.g., o-nitrotoluenes).21 The example shown (Scheme 12, 51 to 52) proceeds in 95% yield on a 48 g scale. This strategy has been utilized by colleagues in Pfizer’s Chemical Research & Development group to prepare alde- hyde 54, a precursor to 55, a cyclooxygenase (COX)-2 inhibitor, and to prepare bisulfite adduct 57 from 2,4-lutidine (Scheme 13). Nitrobenzaldehyde 54 was prepared in 99% yield on a 25 g scale from the corresponding 2-nitrotoluene derivative.22 Bisulfite adduct 57 was prepared in 64% overall Scheme 8. The Synthesis of PKC Inhibitor 38 Scheme 9. Ozonolysis in the Preparation of 24(S)-Hydroxyvitamin D2 Scheme 10. Ozonolysis to Prepare r-Hydroxyester 43 Scheme 11. Ozonolysis of Olefins to Aldehydes and Ketones Scheme 12. Oxidation of Activated Aromatic Methyl Groups Scheme 13. Oxidation of Activated Aromatic Methyl Groupsa a (a) (MeO)2CHNMe2, DMF, 140 °C; (b) BuLi, Et2NH, DMF; (c) (i) NaIO4; (ii) H2SO3. Large-Scale Oxidations in the Pharmaceutical Industry Chemical Reviews, 2006, Vol. 106, No. 7 2947 yield from 2,4-lutidine on a 2.8 kg scale.23 The regioselective lithiation of 2,4-lutidine had been earlier noted by Evans.24 Oxidative cleavage of vicinal diols is frequently utilized in the preparation of aldehydes and ketones. Researchers at Lilly have described extensive optimization and scale-up studies on the preparation of 2,3-O-isopropylidene-D-glyc- eraldehyde (59) from D-mannitol, which included the oxida- tive cleavage of diol 58 (Scheme 14).25 The periodate cleavage proceeded in 67% yield on a 10 g scale and has been successfully scaled to >100 kg quantities. In a subsequent publication,26 Schmid described an improved preparation of the analogous pentylidene-protected aldehydes from 60 using a buffered potassium periodate system; both enantiomers were prepared. The sequence shown in Scheme 15 was utilized for conversion of salicylaldehyde 62 to â-aminoester 65,27 in which oxidative cleavage of amino alcohol 63 to imine 64 was executed on a 150 kg scale. The overall yield for the sequence ranged from 48% to 57%. Sodium periodate was preferred over lead tetraacetate for this oxidation for obvious reasons. Aminoester 65 is a precursor to an Rvâ3 integrin antagonist. Hydrogen peroxide-mediated oxidation of sodium ery- thorbate (66) was utilized to generate dihydroxylactone 67, which was converted to epoxide 69 by tosylation and ethanolysis (Scheme 16).28 This route was preferred for scale- up over other routes examined (e.g., Sharpless AE and derivatization of diethyl tartrate). Noyori and co-workers have reported the oxidative cleav- age of cyclohexene to adipic acid (HO2C(CH2)4CO2H) with 30% hydrogen peroxide, catalytic Na2WO4â2H2O, and a phase-transfer catalyst (Me(n-octyl)3NHSO4), both 1 mol %.29 The crystalline product is isolated by filtration in 90% yield (100 g scale), and the aqueous phase can be recycled into another oxidation by addition of peroxide and phase- transfer catalyst. Dimethyl-1,3-acetonedicarboxylate (71) has been prepared by oxidative decarboxylation of citric acid (70); this was executed on a 400 g scale in 52% yield (Scheme 17).30 This procedure is a modification of an Organic Syntheses procedure, which utilized fuming H2SO4.31,32 Several oxidative protocols for the conversion of olefin 72 to bicyclic amine 74 have been described (Scheme 18).33 Dihydroxylation of 72 can be effected with catalytic OsO4 (0.126 mol %) and either N-methylmorpholine-N-oxide or sodium chlorite as stoichiometric oxidants. The former provides an 89% yield of diol 73 on a 400 g scale. Oxidative cleavage with NaIO4 in aqueous dichloroethane (DCE) generates a solution of the bis-aldehyde, which is condensed with benzylamine and reduced with NaBH(OAc)3 directly; the overall yield for this sequence is 86% on a 40 g scale. An alternative ozonolysis sequence is also described (Scheme 19), in which the methoxyhydroperoxide is reduced by hydrogenation over Pt/C, and benzylamine and HCO2H are added to effect reductive amination after further hydro- genation over the same catalyst. Hydrogenolysis over Pearl- man’s catalyst in the presen