March's Advanced Organic Chemistry: Reactions, Mechanisms, and Structure, 7th Edition (2013)

Chapter 16. Addition to Carbon–Hetero Multiple Bonds

16.B. Reactions

Many of the reactions in this chapter are simple additions to carbon–hetero multiple bonds, with the reaction ending when the two groups have been added. But in many other cases subsequent reactions take place. There are generally two types:

In type A, the initially formed adduct loses water (or, in the case of addition to C=NH, ammonia, etc.), and the net result of the reaction is the substitution of C=Y for C=O (or C=NH, etc.). In type B there is a rapid substitution, and the OH (or NH₂, etc.) is replaced by another group Z, which is often another YH moiety. This substitution is nucleophilic is most cases: Y usually has an unshared pair and S_N1 reactions occur very well on this type of compound (Sec. 10.G.i, category 2), even when the leaving group is as poor as OH or NH₂. In this chapter, reactions will be classified according to what is initially adding to the carbon–heteroatom multiple bond, even if subsequent reactions take place so rapidly that it is impossible to isolate the initial adduct.

Most of the reactions considered in this chapter can be reversed. In many cases, we will consider the reverse reactions with the forward ones, in the same section. The reverse of some of the other reactions are considered in other chapters. In still other cases, one of the reactions in this chapter is the reverse of another (e.g., 16-2 and 16-13). For reactions that are reversible, the principle of microscopic reversibility (Sec. 6.H) applies.

First, reactions in which hydrogen or a metallic ion (or in one case phosphorus or sulfur) adds to the heteroatom will be discussed. Second, reactions in which carbon adds to the heteroatom will be discussed. Within each group, the reactions are classified by the nature of the nucleophile. Additions to isocyanides, which are different in character, follow. Acyl substitution reactions that proceed by the tetrahedral mechanism, which mostly involve derivatives of carboxylic acids, are treated at the end.

16.B.i. Reactions in which Hydrogen or a Metallic Ion Adds to the Heteroatom

A. Attack by OH (Addition of H₂O)

16-1 The Addition of Water to Aldehydes and Ketones: Formation of Hydrates

O-Hydro-C-hydroxy-addition

The adduct formed upon addition of water to an aldehyde or ketone is called a hydrate or gem-diol.³⁸ These compounds are usually stable only in water solution and decompose on distillation; that is, the equilibrium shifts back toward the carbonyl compound, usually via formation of an enol and tautomerization to the carbonyl. The position of the equilibrium is greatly dependent on the structure of the hydrate. Thus, formaldehyde in water at 20 °C exists 99.99% in the hydrated form, while for acetaldehyde this figure is 58%, and for acetone the hydrate concentration is negligible.³⁹ It has been found, by exchange with ¹⁸O, that the reaction with acetone is quite rapid when catalyzed by acid or base, but the equilibrium lies on the side of acetone and water.⁴⁰ Since methyl, a +I group, inhibits hydrate formation, it may be expected that electron-attracting groups would have the opposite effect, and this is indeed the case. The hydrate of chloral (trichloroacetaldehyde)⁴¹ is a stable crystalline substance. In order for it to revert to chloral, ⁻OH or H₂O must leave. This is made difficult by the electron-withdrawing character of the Cl₃C group and by the absence of a proton on the α carbon, which is required for loss of water to form an enol. Some other⁴² polychlorinated and polyfluorinated aldehydes and ketones⁴³ and α-keto aldehydes also form stable hydrates, as do cyclopropanones.⁴⁴ In the last case,⁴⁵ formation of the hydrate relieves some of the I strain (Sec. 9.B) of the parent ketone.

The reaction is subject to both general-acid and general-base catalysis; the following mechanisms can be written for basic (B) and acidic (BH) catalysis, respectively:⁴⁶

In mechanism a, as the H₂O attacks, the base pulls off a proton, and the net result is addition of ⁻OH. This can happen because the base is already hydrogen bonded to the H₂O molecule before the attack. In mechanism b, because HB is already hydrogen bonded to the oxygen of the carbonyl group, it gives up a proton to the oxygen as the water attacks. In this way, B and HB accelerate the reaction even beyond the extent that they form ^–OH or H₃O⁺ by reaction with water. Reactions in which the catalyst donates a proton to the electrophilic reagent (in this case the aldehyde or ketone) in one direction and removes it in the other are called class e reactions. Reactions in which the catalyst does the same to the nucleophilic reagent are called class n reactions.⁴⁷ Thus the acid-catalyzed process here is a class e reaction, while the base-catalyzed process is a class n reaction.

For the reaction between ketones and H₂O₂, see 17-37.

There are no OS references, but see OS VIII, 597, for the reverse reaction.

16-2 Hydrolysis of the Carbon–Nitrogen Double Bond⁴⁸

Oxo-de-alkylimino-bisubstitution, and so on

Compounds containing carbon–nitrogen double bonds can be hydrolyzed to the corresponding aldehydes or ketones.⁴⁹ For imines (W = R or H) the hydrolysis is easy and can be carried out with water. When W = H, the imine is seldom stable enough for isolation, and in aqueous media hydrolysis usually occurs in situ, without isolation. The hydrolysis of Schiff bases (W = Ar) is more difficult and requires acid or base catalysis. Oximes (W = OH), arylhydrazones (W = NHAr), and, most easily, semicarbazones (W = NHCONH₂) can also be hydrolyzed. Often a reactive aldehyde (e.g., formaldehyde), is added to combine with the liberated amine.

A number of reagents⁵⁰ have been used to cleave C=N bonds, especially those not easily hydrolyzable with acidic or basic catalysts or those that contain other functional groups that are attacked under these conditions.

Oximes have been converted to the corresponding aldehyde or ketone⁵¹ by treatment with aq phosphoric acid without an organic cosolvent,⁵² periodic acid,⁵³ DABCO–Br₂,⁵⁴ NBS in water,⁵⁵ Chloramine-T⁵⁶ HCO₂H on SiO₂ with microwave irradiation,⁵⁷ and 20% I₂ in water containing SDS (sodium dodecyl sulfate)⁵⁸ or in an ionic liquid on SiO₂.⁵⁹ Transition metal compounds have been used, including those of Sb,⁶⁰ Co⁶¹ Hg,⁶² Bi,⁶³ Cu,⁶⁴ or Zn.⁶⁵ Oxidizing agents can be quite effective, including KMnO₄ on Al₂O₃,⁶⁶ tetraalkylammonium permanganates,⁶⁷ and quinolinium dichromate.⁶⁸ Alkaline H₂O₂⁶⁹ has also been used.

Phenylhydrazones can be converted to a ketone using Oxone and KHCO₃,⁷⁰ polymer-bound iodonium salts,⁷¹ or KMnO₄ on wet SiO₂.⁷² Dimethylhydrazones have been converted to ketones with FeSO₄·7 H₂O in chloroform,⁷³and Me₃SiCl/NaI in acetonitrile with 1% water.⁷⁴ Hydrazones (e.g., RAMP or SAMP, Reaction 10-68, category 4) can be hydrolyzed with aq CuCl₂.⁷⁵ Tosylhydrazones can be hydrolyzed to the corresponding ketones with aq acetone and BF₃–etherate,⁷⁶ as well as with other reagents.⁷⁷

Semicarbazones have been cleaved with ammonium chlorochromates on alumina⁷⁸ (Bu₄N)₂S₂O_8,⁷⁹ Mg(HSO₄)₂ on wet silica,⁸⁰ or by SbCl₃ with microwave irradiation.⁸¹

The hydrolysis of carbon–nitrogen double bonds involves initial addition of water and elimination of a nitrogen moiety:

It is thus an example of reaction type A (see above). The sequence shown is generalized.⁸² In specific cases, there are variations in the sequence of the steps, depending on acid or basic catalysis or other conditions.⁸³ Which step is rate determining also depends on acidity and on the nature of W and of the groups connected to the carbonyl.⁸⁴

Iminium ions (10)⁸⁵ would be expected to undergo hydrolysis quite readily, since there is a resonance contributor with a positive charge on the carbon. Indeed, they react with water at room temperature.⁸⁶ Acid-catalyzed hydrolysis of enamines (the last step of the Stork enamine reaction, 10-69, involves conversion to iminium ions).⁸⁷

The mechanism of enamine hydrolysis is thus similar to that of vinyl ether hydrolysis (Reaction 10-6).

OS I, 217, 298, 318, 381; II, 49, 223, 234, 284, 310, 333, 395, 519, 522; III, 20, 172, 626, 818; IV, 120; V, 139, 277, 736, 758; VI, 1, 358, 640, 751, 901, 932; VII, 8; 65, 108, 183; 67, 33; 76, 23.

OS II, 24; IV, 819; V, 273; VI, 910.

16-3 Hydrolysis of Aliphatic Nitro Compounds

Oxo-de-hydro,nitro-bisubstitution

Primary or secondary aliphatic nitro compounds can be hydrolyzed, respectively, to aldehydes or ketones, by treatment of their conjugate bases with sulfuric acid. This is called the Nef reaction.⁸⁸ Tertiary aliphatic nitro compounds do not give the reaction because they cannot be converted to their conjugate bases. Like 16-2, this reaction involves hydrolysis of a C=N double bond. A possible mechanism involves initial formation of the aci form of the nitro compound (11):⁸⁹

Intermediates of type 12 have been isolated in some cases.⁹⁰

The conversion of nitro compounds to aldehydes or ketones has been carried out with better yields and fewer side reactions by several alternative methods.⁹¹ Among these are treatment of the nitro compound with basic H₂O₂ in an ionic liquid,⁹² or 30% H₂O₂–K₂CO₃,⁹³ DBU in acetonitrile,⁹⁴ or CAN.⁹⁵

When primary nitro compounds are treated with sulfuric acid without previous conversion to the conjugate bases, they give carboxylic acids. Hydroxamic acids are intermediates and can be isolated, so that this is also a method for their preparation.⁹⁶ Both the Nef reaction and the hydroxamic acid process involve the aci form; the difference in products arises from higher acidity. For example, a difference in sulfuric acid concentration from 2 to 15.5 M changes the product from the aldehyde to the hydroxamic acid.⁹⁷ The mechanism of the hydroxamic acid reaction is not known with certainty, but if higher acidity is required, it may be that the protonated aci form of the nitro compound is further protonated.

OS VI, 648; VII, 414. See also, OS IV, 573.

16-4 Hydrolysis of Nitriles

Nitriles can be hydrolyzed to give either amides or carboxylic acids.⁹⁸ The amide is formed initially, but since amides are also hydrolyzed with acid or basic treatment, the carboxylic acid is readily formed. When the acid is desired,⁹⁹ the reagent of choice is aq NaOH containing ~ 6–12% H₂O₂, but acid-catalyzed hydrolysis is also carried out. The reaction of nitriles with TFA–acetic acid–sulfuric acid, followed by treatment with water gives the corresponding amide.¹⁰⁰ A Rh catalyzed hydrolysis with aq isopropyl alcohol leads to the amide.¹⁰¹ A "dry" hydrolysis of nitriles has been reported.¹⁰² Enzymatic hydrolysis to give the amide was reported with nitrilase (ZmNIT2).¹⁰³ Nitriles can be hydrolyzed to the carboxylic acids without disturbing carboxylic ester functions also present, by the use of tetrachloro- or tetrafluorophthalic acid.¹⁰⁴

Hydrolysis of cyanohydrins [RCH(OH)CN] is usually carried out under acidic conditions, because basic solutions cause competing reversion of the cyanohydrin to the aldehyde and CN⁻. However, cyanohydrins have been hydrolyzed under basic conditions with borax or alkaline borates.¹⁰⁵

There are a number of procedures for stopping at the amide stage,¹⁰⁶ among them the use of concentrated H₂SO₄; 2 molar equivalents of chlorotrimethylsilane followed by H₂O,¹⁰⁷ aq NaOH with PEG-400 and microwave irradiation,¹⁰⁸ heating on neutral alumina,¹⁰⁹ or dry HCl followed by H₂O. The same result can also be obtained by use of water and certain metal ions or complexes¹¹⁰ including an In,¹¹¹ Au,¹¹² or a Ru catalyst.¹¹³ Other reagents include MnO₂/SiO₂ with microwave irradiation,¹¹⁴ or Hg(OAc)₂ in HOAc.¹¹⁵ The reaction of ferric nitrate and an amine generates the amide.¹¹⁶

Nitriles are converted to thioamides [ArC(=S)NH₂] with ammonium sulfide [(NH₄)₂S] in methanol, with microwave irradiation.¹¹⁷ Thioamides are also prepared using phosphorus pentasulfide.¹¹⁸

Thiocyanates are converted to thiocarbamates in a similar reaction:¹¹⁹ R–S–CN + H₂O → R–S–CO– NH₂. Hydrolysis of cyanamides gives amines, produced by the breakdown of the unstable carbamic acid intermediates: R₂NCN → [R₂NCOOH] → R₂NH.

OS I, 21, 131, 201, 289, 298, 321, 336, 406, 436, 451; II, 29, 44, 292, 376, 512, 586 (see, however, V, 1054), 588; III; 34, 66, 84, 88, 114, 221, 557, 560, 615, 851; IV, 58, 93, 496, 506, 664, 760, 790; V, 239; VI, 932; 76, 169. Also see, OS III, 609; IV, 359, 502; 66, 142.

B. Attack by OR (Addition of ROH)

16-5 The Addition of Alcohols to Aldehydes and Ketones

Dialkoxy-de-oxo-bisubstitution

Dithioalkyl-de-oxo-bisubstitution

Acetals and ketals are formed by treatment of aldehydes and ketones, respectively, with alcohols in the presence of acid catalysts.¹²⁰ Lewis acid derivatives of Ti,¹²¹ Cu,¹²² In,¹²³ Ru,¹²⁴ or Co¹²⁵ can be used in conjunction with alcohols. Organocatalysts have been used for this conversion under acid-free conditions.¹²⁶ This reaction is reversible, and acetals and ketals can be hydrolyzed by treatment with acid.¹²⁷ With small unbranched aldehydes the equilibrium lies to the right. If ketals or acetals of larger molecules must be prepared the equilibrium must be shifted, usually by removal of water. This can be done by azeotropic distillation, ordinary distillation, or the use of a drying agent (e.g., Al₂O₃ or a molecular sieve).¹²⁸ The reaction is not catalyzed in either direction by bases, so most acetals and ketals are quite stable to bases, though they are easily hydrolyzed by acids. This reaction is therefore a useful method of protection of aldehyde or ketone functions from attack by bases. The reaction is of wide scope.

Most aldehydes are easily converted to acetals,¹²⁹ but the reaction with ketones is more difficult, presumably for steric reasons. While the reaction often fails, many ketals, especially from cyclic ketones, have been made in this manner.¹³⁰ Many functional groups may be present without being affected. 1,2- and 1,3-Diols form cyclic acetals and ketals (1,3-dioxolanes¹³¹ and 1,3-dioxanes,¹³² respectively), and these are often used to protect aldehydes and ketones. Chiral dioxolanes have been prepared from chiral diols.¹³³ Dioxolanes have been prepared from ketones in ionic liquids.¹³⁴ Ketones are converted with dimethyl ketals by electrolysis with NaBr in methanol.¹³⁵Intramolecular reactions are possible in which a keto diol or an aldehyde diol generates a bicyclic ketal or acetal.

The conversion of acetals back to aldehydes or ketones is accomplished by many reagents, including aq acid. Heating with water under microwave irradiation converts acetals to the corresponding carbonyl compound.¹³⁶Transition metal compounds of Bi¹³⁷ catalyze this conversion as well.

The mechanism for acetal/ketal formation involves initial formation of a hemiacetal,¹³⁸ and it is the reverse of that given for acetal hydrolysis:

In a study of the acid-catalyzed formation of the hemiacetal, Grunwald¹³⁹ showed that the data best fit a mechanism in which the three steps shown here are actually all concerted; that is, the reaction is simultaneously catalyzed by acid and base, with water acting as the base:¹⁴⁰

Hemiacetals themselves are no more stable than the corresponding hydrates (Reaction 16-1). If the original aldehyde or ketone has an α hydrogen, it is possible to lose water from the hemiacetal, and enol ethers can be prepared in this manner:

Similarly, treatment with an anhydride and a catalyst can give an enol ester (see Reaction 16-6).¹⁴¹ As with hydrates, it is noted that hemiacetals of cyclopropanones¹⁴² and of polychloro and polyfluoro aldehydes and ketones may be quite stable.

When acetals or ketals are treated with an alcohol of higher molecular weight than the one already there, transacetalation is possible (see Reaction 10-13). In another type of transacetalation, aldehydes or ketones can be converted to acetals or ketals by treatment with another acetal or ketal or with an ortho ester,¹⁴³ in the presence of an acid catalyst (shown for an ortho ester):

This method is useful for the conversion of ketones to ketals, since the direct reaction of a ketone with an alcohol often gives poor results. Alternatively, the substrate is treated with an alkoxysilane (ROSiMe₃) in the presence of trimethylsilyl trifluoromethanesulfonate.¹⁴⁴ Formic acid reacts with alcohols to give orthoformates.

1,4-Diketones give furans when treated with acids. This is actually an example of an intramolecular addition of an alcohol to a ketone, since it is the enol form that adds:

Similarly, 1,5-diketones give pyrans. Conjugated 1,4-diketones (e.g., 1,4-diphenylbut-2-en-1,4-dione) are converted to 2,5-diphenylfuran with formic acid, 5% Pd/C, PEG-200, and a sulfuric acid catalyst with microwave irradiation.¹⁴⁵ Note that alkynyl ketones are converted to furans with palladium(II) acetate.¹⁴⁶

OS I, 1, 298, 364, 381; II, 137; III, 123, 387, 502, 536, 644, 731, 800; IV, 21, 479, 679; V, 5, 292, 303, 450, 539; VI, 567, 666, 954; VII, 59, 149, 168, 177, 241, 271, 297; VIII, 357. Also see, OS IV, 558, 588; V, 25; VIII, 415.

16-6 Acylation of Aldehydes and Ketones

O-Acyl-C-acyloxy-addition

Aldehydes can be converted to acylals by treatment with an anhydride in the presence of proton acids,¹⁴⁷ NBS,¹⁴⁸ ceric ammonium nitrate,¹⁴⁹ BF_3, LiBF₄,¹⁵⁰ and Lewis acid compounds of Fe,¹⁵¹ In,¹⁵² Cu,¹⁵³ Bi,¹⁵⁴ W,¹⁵⁵ or Zr.¹⁵⁶N-Chlorosuccinimide with thiourea is a highly efficient catalyst.¹⁵⁷ Silica supported perchloric acid is useful for the preparation of acylals.¹⁵⁸ Conjugated aldehydes are converted to the corresponding acylal by reaction with acetic anhydride and a FeCl₃ catalyst.¹⁵⁹ The reaction cannot normally be applied to ketones, though an exception has been reported when the reagent is trichloroacetic anhydride, which gives acylals with ketones without a catalyst.¹⁶⁰

OS IV, 489.

16-7 Reductive Alkylation of Alcohols

C-Hydro-O-alkyl-addition

Aldehydes and ketones can be converted to ethers by treatment with an alcohol and triethylsilane in the presence of a strong acid¹⁶¹ or by hydrogenation in alcoholic acid in the presence of a Pt catalyst.¹⁶² The process can formally be regarded as addition of ROH to give a hemiacetal [RR′C(OH)OR²], followed by reduction of the OH. In this respect, it is similar to Reaction 16-17. Homoallylic ethers are formed by the Fe catalyzed reaction of acetals and aldehydes,¹⁶³ and by reactions in ionic liquids.¹⁶⁴ The In catalyzed reaction of aldehydes with allyltriethoxysilane leads to the corresponding ether.¹⁶⁵

Ethers have also been prepared by the reductive dimerization of two molecules of an aldehyde or ketone (e.g., cyclohexanone → dicyclohexyl ether). This was accomplished by treatment of the substrate with a trialkylsilane and a catalyst.¹⁶⁶

16-8 The Addition of Alcohols to Isocyanates

N-Hydro-C-alkoxy-addition

The reaction of an isocyanate with alcohols gives a carbamate (a substituted urethane). This is an excellent reaction of wide scope that gives good yields. Isocyanic acid (HNCO) gives unsubstituted carbamates. Addition of a second equivalent of HNCO gives allophanates.

The isocyanate can be generated in situ by the reaction of an amine and oxalyl chloride, and subsequent reaction with HCl and then an alcohol gives the carbamate.¹⁶⁷ Combining compounds with two NCO groups with compounds containing two OH groups makes polyurethanes. Cyclic carbamates (e.g., 1,3-oxazine-2-ones), are generated by the reaction of an isocyanate with an oxetane, in the presence of a Pd catalyst.¹⁶⁸ Isothiocyanates similarly give thiocarbamates (RNHCSOR′),¹⁶⁹ but they react slower than the corresponding isocyanates. Isocyanates react with LiAlHSeH and then iodomethane to give the corresponding selenocarbonate (RNHCOSeMe).¹⁷⁰

The details of the mechanism are poorly understood,¹⁷¹ but the oxygen of the alcohol certainly attacks the carbon of the isocyanate. Hydrogen bonding complicates the kinetic picture.¹⁷² Metallic compounds, can also catalyze the addition of ROH to isocyanates¹⁷³ by light,¹⁷⁴ or, for tertiary ROH, by lithium alkoxides¹⁷⁵ or n-butyllithium.¹⁷⁶

OS I, 140; V, 162; VI, 95, 226, 788, 795.

16-9 Alcoholysis of Nitriles

Alkoxy,oxo-de-nitrilo-tersubstitution

The addition of dry HCl to a mixture of a nitrile and an alcohol in the absence of water leads to the hydrochloride salt of an imino ester (imino esters are also called imidates and imino ethers). This reaction is called the Pinner synthesis.¹⁷⁷ The salt can be converted to the free imino ester by treatment with a weak base (e.g., sodium bicarbonate) or it can be hydrolyzed with water and an acid catalyst to the corresponding carboxylic ester. If the latter is desired, water may be present from the beginning, in which case aq HCl can be used and the need for gaseous HCl is eliminated. Imino esters can also be prepared from nitriles with basic catalysts.¹⁷⁸

This reaction is of broad scope and is good for aliphatic, aromatic, and heterocyclic R and for nitriles with oxygen-containing functional groups. The application of the reaction to nitriles containing a carboxyl group constitutes a good method for the synthesis of mono esters of dicarboxylic acids with the desired group esterified and with no diester or diacid present.

Cyanogen chloride reacts with alcohols in the presence of an acid catalyst (e.g., dry HCl or AlCl₃), to give carbamates:¹⁷⁹

ROH can also be added to nitriles in another manner (Reaction 16-91).

OS I, 5, 270; II, 284, 310; IV, 645; VI, 507; VIII, 415.

16.2.1.2.6 16-10 The Formation of Carbonates and Xanthates

Di-C-alkoxy-addition; S-Metallo-C-alkoxy-addition

The reaction of phosgene with an alcohol generates haloformic esters, and reaction with a second equivalent of alcohol gives a carbonate. This reaction is related to the acyl addition reactions of acyl chlorides in Reaction 16-98. An important example is the preparation of carbobenzoxy chloride (PhCH₂OCOCl; CbzCl) from phosgene and benzyl alcohol. When CbzCl reacts with an amine, the product is the benzyl carbamate, N-Cbz, which is widely used for protection of amino groups during peptide synthesis. When an alcohol reacts with certain alkyl halides (e.g., benzyl chloride) and CO₂, in the presence of Cs₂CO₃ and tetrabutylammonium iodide, a mixed carbonate is formed.¹⁸⁰

The addition of alcohols to carbon disulfide (S=C=S) in the presence of a base produces xanthates.¹⁸¹ The base is often HO⁻, but in some cases better results can be obtained by using methylsulfinyl carbanion (MeSOCH₂⁻).¹⁸² If an alkyl halide (RX) is present, the xanthate ester (ROCSSR′) can be produced directly. In a similar manner, alkoxide ions add to CO₂ to give carbonate ester salts (ROCO₂⁻).

OS V, 439; VI, 207, 418; VII, 139.

C. Sulfur Nucleophiles

16-11 The Addition of H₂S and Thiols to Carbonyl Compounds

O-Hydro-C-mercapto-addition, thioxo-de-oxo-bisubstitution,dimercapto-de-oxo-bisubstitution, and carbonyl-trithiane transformation

The addition of H₂S to an aldehyde or ketone can result in a variety of products. The most usual product is the trithiane 16.¹⁸³gem-Dithiols (15) are much more stable than the corresponding hydrates or α-hydroxy thiols.¹⁸⁴ They have been prepared by the treatment of ketones with H₂S under pressure¹⁸⁵ and under mild conditions with HCl as a catalyst.¹⁸⁶ A much more useful application is the addition of thiols to aldehydes and ketones to give hemi-mercaptals [CH(OH)SR] and dithioacetals [CH(SR)₂] (Reaction 16-5). This reaction is generally not a good route to thioketones (14). α-Hydroxythiols (13) can be prepared from polychloro and polyfluoro aldehydes and ketones.¹⁸⁷Apparently gem-hydroxy thiols (e.g., 13) are rather unstable and quite difficult to prepare.

Thiols add to aldehydes and ketones to initially give hemi-mercaptals and dithioacetals. Hemi-mercaptals are ordinarily unstable,¹⁸⁸ but they are usually more stable than the corresponding hemiacetals and can be isolated in certain cases.¹⁸⁹ The isolated product of this reaction is most often the dithioacetal, which like the acetals obtained by reaction with an alcohol, are stable in the presence of base. However, a sufficiently strong base can remove a proton from the carbon between the sulfur atoms (–S–CHR–S–)¹⁹⁰ to form the corresponding carbanion (see Reaction 10-71). The pK_a of such protons is typically 31–37,¹⁹¹ requiring a strong base, and deprotonation is often quite slow. The reaction of aldehydes or ketones with thiols most commonly uses a Lewis acid catalyst (e.g., boron trifluoride etherate, BF₃·OEt₂)¹⁹² to give the dithioacetal¹⁹³ or dithioketal. Dithioacetals can also be prepared from aldehydes or ketones by treatment with thiols in the presence of TiCl₄,¹⁹⁴ SiCl₄,¹⁹⁵ LiBF₄,¹⁹⁶ Al(OTf)₃,¹⁹⁷ tosic acid on silica gel in dichloromethane,¹⁹⁸ and oxalic acid promotes the reaction.¹⁹⁹ Similarly reactions that use 1,2-ethanedithiol or 1,3-propanedithiol lead to 1,3-dithiolanes (17)²⁰⁰ or 1,3-dithianes.²⁰¹ 3-(1,3-Dithian-2-ylidene)pentane- 2,4-dione has been used as a thioacetalization reagent for reactions in water.²⁰²

Dithioacetals and dithioketals are used as protecting groups for aldehydes and ketones, and after subsequent reactions involving the R or R′ group, deprotection regenerates the carbonyl.²⁰³ Simple hydrolysis is the most common method for converting thiocarbonyls to carbonyls, but there are a variety of reagents for this conversion.²⁰⁴ Lewis acids (e.g., aluminum chloride, AlCl₃) and mercuric salts are common reagents (the Corey–Seebach procedure).²⁰⁵Other reagents include BF₃·OEt₂ in aq THF containing mercuric oxide (HgO),²⁰⁶ ceric ammonium nitrate [Ce(NH₄)₂(NO₃)₆],²⁰⁷ chlorotrimethylsilane and H₂O₂,²⁰⁸ PhI(OAc)₂ in aq acetone,²⁰⁹ and Cu salts.²¹⁰ When aldehydes and ketones react with mercapto–alcohols, mixed acetals or ketals are formed.

Thioamides are converted to amides with Caro's acid (H₂SO₅) on SiO₂.²¹¹ The use of 2-mercaptoethanol (HSCH₂CH₂OH), for example, leads to an oxathiolane²¹² and 3-mercaptopropanol (HSCH₂CH₂CH₂OH) leads to an oxathiane. Alternatively, the dithioketal can be desulfurized with Raney nickel (Reaction 14-27), giving the overall conversion C=O → CH₂ (Reaction 19-70).

Thioketones (14) can be prepared from certain ketones (e.g., diaryl ketones) by treatment with H₂S and an acid catalyst, usually HCl. They are often unstable and usually trimerize (to 16) or react with air. Thioaldehydes²¹³ are even less stable and simple ones²¹⁴ apparently have never been isolated, although tert-BuCHS has been prepared in solution, where it exists for several hours at 20 °C.²¹⁵ A high-yield synthesis of thioketones involves treatment of acyclic²¹⁶ ketones with 2,4-bis(4-methoxyphenyl)-1,3,2,4-dithiadiphosphetane-2,4-disulfide²¹⁷ (18, known as Lawesson's reagent).²¹⁸ Thioketones can also be prepared by treatment of ketones with P₄S₁₀,²¹⁹ P₄S₁₀ and hexamethyldisiloxane,²²⁰ P₄S₁₀ on alumina,²²¹ and from oximes or various types of hydrazone (overall conversion C=N– → C=S).²²² Reagent 18 converts the C=O groups of amides and carboxylic esters²²³ to C=S groups.²²⁴Lactones react with 18 in the presence of hexamethyldisiloxane an microwave irradiation to give the thiolactone.²²⁵

Other reagents may be used for this transformation including POCl₃ followed by S(TMS)₂, which converts lactams to thiolactams.²²⁶ Treatment with triflic anhydride, and then H₂S²²⁷ or aq S(NH₄)₂²²⁸ converts amides to thioamides, as does the microwave assisted reaction with PSCl₂/H₂O/Et₃N, without solvent.²²⁹ Carboxylic acids (RCOOH) can be converted directly to dithiocarboxylic esters (RCSSR′),²³⁰ in moderate yield, with P₄S₁₀ and a primary alcohol (R′OH).²³¹

If an aldehyde or ketone possesses an α hydrogen, it can be converted to the corresponding enol thioether (19) by treatment with a thiol in the presence of TiCl₄.²³² Aldehydes and ketones have been converted to sulfides by treatment with thiols and pyridine–borane, RCOR′ + R²SH → RR′CHSR²,²³³ in a reductive alkylation reaction, analogous to Reaction 16-7.

OS II, 610; IV, 927; VI, 109; VII, 124, 372. Also see, OS III, 332; IV, 967; V, 780; VI, 556; VIII, 302.

16-12 Formation of Bisulfite Addition Products

O-Hydro-C-sulfonato-addition

Bisulfite addition products are formed from aldehydes, methyl ketones, cyclic ketones (generally seven-membered and smaller rings), α-keto esters, and isocyanates, upon treatment with sodium bisulfite (NaHSO₃). Most other ketones do not undergo the reaction, probably for steric reasons. The reaction is reversible (by treatment of the addition product with either acid or base)²³⁴ and is useful for the purification of the starting compounds, since the addition products are soluble in water and many of the impurities are not.²³⁵

OS I, 241, 336; III, 438; IV, 903; V, 437.

D. Attack by NH₂, NHR, or NR₂ (Addition of NH₃, RNH₂, R₂NH)

16-13 The Addition of Amines to Aldehydes and Ketones

Alkylimino-de-oxo-bisubstitution

The addition of ammonia²³⁶ to aldehydes or ketones does not generally give useful products. According to the pattern followed by analogous nucleophiles, the initial products would be expected to be hemiaminals,²³⁷ but these compounds are generally unstable. In addition, many imines with a hydrogen on the nitrogen spontaneously polymerize.²³⁸

In contrast to ammonia, primary, secondary, and tertiary amines can add to aldehydes²³⁹ and ketones to give different kinds of products. Primary amines give imines²⁴⁰ and secondary amines give enamines (for reactions, see 10-69). Such imines are stable enough for isolation, but in some cases, especially with simple R groups, they rapidly decompose or polymerize unless there is at least one aryl group on the nitrogen or the carbon. When there is an aryl group, the compounds are quite stable. They are usually called Schiff bases, and this reaction is the best way to prepare them.²⁴¹ Even sterically hindered imines can be prepared.²⁴² The initial N-substituted hemiaminals²⁴³ lose water to give the stable Schiff bases.

In general, ketones react more slowly than aldehydes, and higher temperatures and longer reaction times are often required.²⁴⁴ In addition, the equilibrium must often be shifted, usually by removal of the water, either azeotropically by distillation, or with a drying agent (e.g., TiCl₄),²⁴⁵ or by addition of a molecular sieve.²⁴⁶ Imines have been formed from aldehydes and amines in an ionic liquid.²⁴⁷

The reaction is often used to effect ring closure.²⁴⁸ The Friedländer quinoline synthesis²⁴⁹ is an example where ortho alkenyl aniline derivatives give the quinoline (20).²⁵⁰ The alkene derivative can be prepared in situ from an aldehyde and a suitably functionalized ylid.²⁵¹ Pyrylium ions react with ammonia or primary amines to give pyridinium ions²⁵² (see Reaction 10-57). Primary amines react with 1,4-diketones, with microwave irradiation, to give N-substituted pyrroles.²⁵³ Similar reactions in the presence of Montmorillonite KSF²⁵⁴ or by simply heating the components with tosic acid²⁵⁵ have been reported.

The reaction of secondary amines with ketones leads to enamines (see 10-69).²⁵⁶ When secondary amines are added to aldehydes or ketones, the initially formed N,N-disubstituted hemiaminals (21) cannot lose water in the same way since the iminium ion intermediate does not have a proton on nitrogen, and in some cases it is possible to isolate them.²⁵⁷ However, they are generally unstable, and under the reaction conditions usually react further. If no α hydrogen is present, 21 is converted to the more stable aminal (22).²⁵⁸ However, if an α hydrogen is present, water (from 21) or RNH₂ (from 22) can be lost in that direction to give an enamine (24).²⁵⁹ This is the most common method²⁶⁰ for the preparation of enamines and usually takes place when an aldehyde

or ketone containing an α hydrogen is treated with a secondary amine.²⁶¹ The water is usually removed azeotropically or with a drying agent,²⁶² and molecular sieves can also be used.²⁶³ Silyl carbamates (e.g., Me₂NCO₂SiMe₃) have been used to convert ketones to enamines.²⁶⁴ Stable primary enamines have also been prepared.²⁶⁵ Enamino–ketones have been prepared from diketones and secondary amines using low-molecular-weight amines in water,²⁶⁶ or using microwave irradiation on silica gel.²⁶⁷ Enamines have been prepared by the reaction of an aldehyde, a secondary amine, and a terminal alkyne in the presence of AgI at 100 °C,²⁶⁸ AgI in an ionic liquid,²⁶⁹ CuI with microwave irradiation,²⁷⁰ or a Au catalyst.²⁷¹

Tertiary amines can only give salts (23).

OS I, 80, 355, 381; II, 31, 49, 65, 202, 231, 422; III, 95, 328, 329, 332, 358, 374, 513, 753, 827; IV, 210, 605, 638, 824; V, 191, 277, 533, 567, 627, 703, 716, 736, 758, 808, 941, 1070; VI, 5, 448, 474, 496, 520, 526, 592, 601, 818, 901, 1014; VII, 8, 135, 144, 473; VIII, 31, 132, 403, 451, 456, 493, 586, 597. Also see, OS IV, 283, 464; VII, 197; VIII, 104, 112, 241.

16-14 The Addition of Hydrazine Derivatives to Carbonyl Compounds

Hydrazono-de-oxo-bisubstitution

The product of condensation of a hydrazine and an aldehyde or ketone is called a hydrazone. Hydrazine itself gives hydrazones only with aryl ketones. With other aldehydes and ketones, either no useful product can be isolated, or the remaining NH₂ group condenses with a second equivalent of carbonyl compound to give an azine. This type of product is especially important for aromatic aldehydes:

However, in some cases azines can be converted to hydrazones by treatment with excess hydrazine and NaOH.²⁷² Arylhydrazines, especially phenyl, p-nitrophenyl, and 2,4-dinitrophenyl,²⁷³ are used much more often and give the corresponding hydrazones with most aldehydes and ketones.²⁷⁴ Since these are usually solids, they make excellent derivatives and were commonly employed for this purpose in the past, before the advent of modern spectroscopy methods. Cyclic hydrazones are known,²⁷⁵ as are conjugated hydrazones.²⁷⁶ Azides react with N,N-dimethylhydrazine and ferric chloride to give the N,N-dimethylhydrazone.²⁷⁷ Alkenes react with CO/H₂, phenylhydrazine and a diphosphine catalyst to give a regioisomeric mixture of phenylhydrazones that favored “anti-Markovnikov” addition.²⁷⁸ Oximes are converted to hydrazones with water and hydrazine in refluxing ethanol.²⁷⁹

α-Hydroxy aldehydes and ketones and α-dicarbonyl compounds give osazones, in which two adjacent carbons have carbon–nitrogen double bonds:

Osazones are particularly important in carbohydrate chemistry, and the osazone test²⁸⁰ with phenylhydrazine is used to test for the presence of sugars with an adjacent sterogenic carbon. In contrast to this behavior, β-diketones and β-keto esters give pyrazoles and pyrazolones, respectively (the latter is illustrated for β-keto esters). No azines are formed under these conditions.

Other hydrazine derivatives frequently used to prepare the corresponding hydrazone are semicarbazide (NH₂NHCONH₂), in which case the hydrazone is called a semicarbazone: Girard's reagents T and P are hydrazones that are water soluble because of the ionic group. Girard's reagents are often used for purification of carbonyl compounds.²⁸¹

Simple N-unsubstituted hydrazones can be obtained by an exchange reaction. The N,N-dimethylhydrazone is prepared first and then treated with hydrazine:²⁸²

OS II, 395; III, 96, 351; IV, 351, 377, 536, 884; V, 27, 258, 747, 929; VI, 10, 12, 62, 242, 293, 679, 791; VII, 77, 438. Also see, OS III, 708; VI, 161; VIII, 597.

16-15 The Formation of Oximes

Hydroxyimino-de-oxo-bisubstitution

In a reaction very much like 16-14, oximes can be prepared by the addition of hydroxylamine (NH₂OH) to aldehydes or ketones. Derivatives of hydroxylamine [e.g., H₂NOSO₃H and HON(SO₃Na)₂] have also been used. For hindered ketones (e.g., 2,2,4,4-tetramethyl-3-pentanone), high pressures (as high as 10,000 atm) may be necessary.²⁸³ The reaction of hydroxylamine with unsymmetrical ketones or with aldehydes leads to a mixture of (E)- and (Z)-isomers. For aromatic aldehydes, heating with K₂CO₃ led to the (E)- isomer, whereas heating with CuSO₄ gave the (Z)-hydroxylamine.²⁸⁴ Hydroxylamines react with ketones in ionic liquids²⁸⁵ and on silica gel.²⁸⁶

It has been shown²⁸⁷ that the rate of formation of oximes is at a maximum at a pH that depends on the substrate but is usually ~ 4. The rate also decreases as the pH is either raised or lowered from this point (a bell-shaped curve). In Section 16.A.i, bell-shaped curves like this were shown to be caused by changes in the rate-determining step in many cases. In this case, at low pH values step 2 is rapid (because it is acid catalyzed), and

step 1 is slow (and rate determining), because under these acidic conditions most of the NH₂OH molecules have been converted to the conjugate ⁺NH₃OH ions, which cannot attack the substrate. As the pH is slowly increased, the fraction of free NH₂OH molecules increases and consequently so does the reaction rate, until the maximum rate is reached at ~ pH 4. As the rising pH causes an increase in the rate of step 1, it also causes a decrease in the rate of the acid-catalyzed step 2, although this latter process has not affected the overall rate since step 2 was still faster than step 1. However, when the pH goes above ~ 4, step 2 becomes rate determining, and although the rate of step 1 is still increasing (as it will until essentially all the NH₂OH is unprotonated), it is now step 2 that determines the rate. This step is slowed by the decrease in acid concentration. Thus the overall rate decreases as the pH rises beyond ~ 4. It is likely that similar considerations apply to the reaction of aldehydes and ketones with amines, hydrazines, and other nitrogen nucleophiles.²⁸⁸ There is evidence that when the nucleophile is 2-methylthiosemicarbazide, there is a second change in the rate-determining step: > pH ~ 10 basic catalysis of step 2 has increased the rate of this step to the point where step 1 is again rate determining.²⁸⁹ Still a third change in the rate-determining step has been found at ~ pH 1, showing that at least in some cases step 1 actually consists of two steps: formation of a zwitterion (e.g., HOH₂N⁺–C–O⁻ in the case shown above) and conversion of this to 25.²⁹⁰ The intermediate 25 has been detected by NMR in the reaction between NH₂OH and acetaldehyde.²⁹¹

OS I, 318, 327; II, 70, 204, 313, 622; III, 690, IV, 229; V, 139, 1031; VII, 149. See also, OS VI, 670.

16-16 The Conversion of Aldehydes to Nitriles

Nitrilo-de-hydro,oxo-tersubstitution

Aldehydes can be converted to nitriles in one step by treatment with hydroxylamine hydrochloride and either formic acid,²⁹² KF–Al₂O₃,²⁹³ or NaHSO₄·SiO₂ with microwave irradiation.²⁹⁴ Heating in N-methylpyrrolidinone (NMP) is also effective with aryl aldehydes²⁹⁵ and heating on dry alumina with aliphatic aldehyde.²⁹⁶ The reaction is a combination of 16-15 and 17-29. Direct nitrile formation has also been accomplished with certain derivatives of NH₂OH, notably, NH₂OSO₂OH.²⁹⁷ Treatment with hydroxylamine and NaI²⁹⁸ or certain carbonates²⁹⁹ also converts aldehydes to the nitrile. Another method involves treatment with hydrazoic acid, although the Schmidt reaction(18-16) may compete.³⁰⁰ Microwave irradiation has been used with NH₂OH·HCl and another reagent, which includes phthalic anhydride³⁰¹ or H–Y zeolite.³⁰² The reaction of an aldehyde with hydroxylamine followed by diethyl cholorophosphate (EtOPOCl₂) gives the nitrile.³⁰³ Heating with hydroxylamine in DMSO³⁰⁴ and using hydroxylamine and oxalyl chloride³⁰⁵ have been used. tert-Butanesulfinyl imines are also used for the conversion of aldehydes to nitriles.³⁰⁶ Other reagents include trimethylsilyl azide,³⁰⁷ hydroxylamine hydrochloride/MgSO₄/TsOH,³⁰⁸ or I₂ with aq ammonia,³⁰⁹ The reaction of a conjugated aldehyde with ammonia, CuCl, and 50% H₂O₂ gave the conjugated nitrile.³¹⁰ Aldehydes with IBX (o-iodoxybenzoic acid) and liquid ammonia gives the nitrile.³¹¹ Tetrabutylammonium tribromide in aq ammonia has also been used.³¹² Trichloroisocyanuric acid with a catalytic amount of TEMPO (Sec. 5.C.i) converts aldehydes to nitriles at 0 °C in dichloromethane.³¹³ Aromatic aldehydes are converted to the nitrile by heating 2.2 molar equivalents of NaN(SiMe₃)₂ in 1,3-dimethylimidazolidin-2-one, in a sealed tube.³¹⁴ The aldehydes employed had a hydroxy substituent.

In a related reaction, the reaction of primary alcohols with iodine in ammonia water gives the corresponding nitrile.³¹⁵ Upon treatment with 2 molar equivalents of dimethylaluminum amide (Me₂AlNH₂), carboxylic esters give nitriles: RCO₂R′ → RCN.³¹⁶ This is likely a combination of Reaction 16-75 and 17-30 (see Reaction 19-5).

OS V, 656.

16-17 Reductive Alkylation of Ammonia or Amines

Hydro,dialkylamino-de-oxo-bisubstitution

When an aldehyde or a ketone is treated with ammonia or a primary or secondary amine in the presence of hydrogen gas and an appropriate catalyst (e.g., Rh or Ir; heterogeneous or homogeneous),³¹⁷reductive alkylation of ammonia or the amine (or reductive amination of the carbonyl compound) takes place.³¹⁸ The reaction can formally be regarded as occurring in the following manner (shown for a primary amine), which probably does correspond to the actual sequence of steps:³¹⁹ In this regard, the reaction of an aldehyde with an amine to give an iminium salt (16-31) can be followed in a second chemical step of reduction of the C=N unit (19-42) using NaBH₄ or a variety of other reagents.³²⁰

Primary amines have been prepared from many aldehydes with at least five carbons and from many ketones by treatment with ammonia and a reducing agent. Smaller aldehydes are usually too reactive to permit isolation of the primary amine. Secondary amines have been prepared by both possible procedures: 2 molar equivalents of ammonia and 1 molar equivalent of aldehyde or ketone, and 1 molar equivalent of primary amine and 1 molar equivalent of carbonyl compound, the latter method being better for all but aromatic aldehydes. Tertiary amines can be prepared in three ways. In general, they are prepared from primary or secondary amines.³²¹ The method is seldom carried out with 3 molar equivalents of ammonia and 1 molar equivalent of carbonyl compound. When the reagent is ammonia, it is possible for the initial product to react again and for this product to react again, so that secondary and tertiary amines are usually obtained as side products. Similarly, primary amines give tertiary as well as secondary amines. In order to minimize this, the aldehyde or ketone is treated with an excess of ammonia or primary amine (unless of course the higher amine is desired).

For ammonia and primary amines there are two possible pathways, but when secondary amines are involved, only the hydrogenolysis pathway is possible. The reaction is compatible with amino acids, giving the N-alkylated amino acid.³²² Other reducing agents³²³ can be used instead of hydrogen and a catalyst, among them boranes,³²⁴ PhSiH₃ with 2% Bu₂SnCl₂,³²⁵ triethylsilane with an Ir³²⁶ or an In catalyst,³²⁷ zinc and HCl, or Zn (with formaldehyde for reductive methylation),³²⁸ and polymethylhydrosiloxane.³²⁹ Several hydride reducing agents can be used, including NaBH_4,³³⁰ NaBH₄ with Ti(OiPr)₄³³¹ or NiCl₂,³³² NaBH₄/H₃BO₄,³³³ BER,³³⁴ polymer-bound triethylammonium acetoxyborohydride,³³⁵ LiBH₄,³³⁶ ZnBH₄-N-methylpiperidine,³³⁷ ZrBH₄,³³⁸ NaBH₃CN,³³⁹ or sodium triacetoxyborohydride,³⁴⁰ Aldehydes and primary amines react with allylic halides, in the presence of Zn dust, to give a homoallylic secondary amine.³⁴¹ A Hantzsch dihydropyridine in conjunction with a Sc catalyst has been used,³⁴² and the use of a Hantzsch ester in a reductive amination is sometimes called a hydrogen-bond catalyzed reaction.³⁴³The reaction of an aldehyde and an amine in isopropyl alcohol, in the presence of Ni nanoparticles, undergoes reductive amination via hydrogen transfer.³⁴⁴

Formic acid is an effective reagent for reductive amination³⁴⁵ in what is called the Wallach reaction. Secondary amines react with formaldehyde and NaH₂PO₃ to give the N-methylated tertiary amine³⁴⁶ and microwave irradiation has also been used.³⁴⁷ Conjugated aldehydes are converted to alkenyl-amines with the amine/silica gel followed by reduction with zinc borohydride.³⁴⁸ In the particular case where primary or secondary amines are reductively methylated with formaldehyde and formic acid, the method is called the Eschweiler–Clarke procedure. Heating with paraformaldehyde and oxalyl chloride has been used to give the same result.³⁴⁹ It is possible to use ammonium (or amine) salts of formic acid,³⁵⁰ or formamides, as a substitute for the Wallach conditions. This method is called the Leuckart reaction,³⁵¹ and in this case the products obtained are often the N-formyl derivatives of the amines instead of the free amines. A Rh catalyzed variation has been reported.³⁵²

Allylic silanes react with aldehydes and carbamates, in the presence of Bi catalysts,³⁵³ or BF₃·OEt₂³⁵⁴ to give the corresponding allylic N-carbamoyl derivative. The reaction can be done with aromatic amines in the presence of vinyl ethers and a Cu complex to give β-amino ketones.³⁵⁵ Reductive amination of an aryl amine and an aryl aldehyde that contains an ortho conjugated ketone substituent gives the amine, which adds 1,4- (Reaction 15-AA) to the α,β-unsaturated ketone unit to give a bicyclic amine.³⁵⁶ Alternative methods of reductive alkylation have been developed. Alkylation of an imine formed in situ is also possible.³⁵⁷

Enantioselective reductive amination reactions are known, generating chiral amines. Ketones and anilines react in the presence of an organocatalyst and a catalytic amount of a chiral phosphoric acid to give the chiral amine.³⁵⁸The reaction of an aldehyde with a chiral amine initiated a reaction that gave a chiral primary amine.³⁵⁹ A Yb catalyzed reaction with a ketone gave a chiral secondary amine.³⁶⁰ Aldehydes react with N-diphenylphosphinoylimines and Et₂Zn, in the presence of a chiral Cu precatalyst, to give a chiral amine.³⁶¹ Asymmetric biocatalytic reductive amination reactions are known.³⁶² Asymmetric reductive amination has been attempted using a Hantzsch estermediated reaction.³⁶³

Reductive alkylation has also been carried out on nitro, nitroso, azo, and other compounds that are reduced in situ to primary or secondary amines. Azo compounds react with aldehydes, in the presence of proline, and subsequent reduction with NaBH₄ gives the chiral hydrazine derivative.³⁶⁴

OS I, 347, 528, 531; II, 503; III, 328, 501, 717, 723; IV, 603; V, 552; VI, 499; VII, 27.

16-18 Addition of Amides to Aldehydes

Alkylamido-de-oxo-bisubstitution

Amides can add to aldehydes in the presence of bases (so the nucleophile is actually RCONH⁻) or acids to give acylated amino alcohols, which often react further to give alkylidene or arylidene bisamides.³⁶⁵ If the R′ group contains an α hydrogen, water may split out.

Sulfonamides add to aldehydes to give the N-sulfonyl imine. Benzaldehyde reacts with TsNH₂ (e.g., with trifluoroacetic anhydride, TFAA) in CH₂Cl₂ at reflux,³⁶⁶ or with TiCl₄ in refluxing dichloroethane,³⁶⁷ to give the N-tosylimine (Ts–N=CHPh). In a similar manner, the reaction of TolSO₂Na + PhSO₂Na with an aldehyde in aq formic acid gives the N-phenylsulfonyl imine.³⁶⁸ The reaction of an aldehyde with Ph₃P=NTs and a ruthenium catalyst gives the N-tosyl imine.³⁶⁹

Primary and secondary amines add to ketenes to give, respectively, N-substituted and N,N-disubstituted amides:³⁷⁰ and to ketenimines to give amidines, (26).³⁷¹

16-19 The Mannich Reaction

Acyl,amino-de-oxo-bisubstitution, and so on

In the Mannich reaction, formaldehyde (or sometimes another aldehyde) is condensed with ammonia, in the form of its salt, and a compound containing an active hydrogen.³⁷² This can formally be considered as an addition of ammonia to give H₂NCH₂OH, followed by a nucleophilic substitution. The reaction can be carried out with salts of primary or secondary amines,³⁷³ or with amides³⁷⁴ rather than ammonia, in which cases the product is substituted on the nitrogen with R, R², and RCO, respectively. The product is referred to as a Mannich base. The imine can be generated in situ, and the reaction of a ketone, formaldehyde, and diethylamine with microwave irradiation gave the Mannich product, a β-amino ketone.³⁷⁵ Many active hydrogen compounds give the reaction, including ketones and aldehydes, esters, nitroalkanes,³⁷⁶ and nitriles as well as ortho-carbon atoms of phenols, the carbon of terminal alkynes, the oxygen of alcohols and the sulfur of thiols.³⁷⁷ Arylamines do not normally give the reaction, but hydrazines can be used.³⁷⁸ Vinylogous Mannich reactions are known³⁷⁹ (see Sec. 6.B for vinylogy).

The Mannich base product can react further in three ways. If it is a primary or secondary amine, it may condense with one or two additional molecules of aldehyde and an active compound, for example,

If the active hydrogen compound has two or three active hydrogen atoms, the Mannich base may condense with one or two additional molecules of aldehyde and ammonia or amine, for example,

Another reaction consists of condensation of the Mannich base with excess formaldehyde:

Sometimes it is possible to obtain these products of further condensation as the main products of the reaction. At other times, they are side products.

When the Mannich base contains an amino group β to a carbonyl (and it usually does), ammonia is easily eliminated. This is a route to α,β-unsaturated aldehydes, ketones, esters, and so on.

Studies of the reaction kinetics have led to the following proposals for the mechanism of the Mannich reaction.³⁸⁰ The base-catalyzed reaction:

The acid-catalyzed reaction:

According to this mechanism, it is the free amine, not the salt that reacts, even in acid solution; and the active-hydrogen compound (in the acid-catalyzed process) reacts as the enol when that is possible. This latter step is similar to what happens in Reaction 12-4. There is kinetic evidence for the intermediacy of the iminium ion (27).³⁸¹

When an unsymmetrical ketone is used as the active-hydrogen component, two products are possible. Regioselectivity has been obtained by treatment of the ketone with preformed iminium ions:³⁸² the use of Me₂N⁺=CH₂CF₃COO⁻ in CF₃COOH gives substitution at the more highly substituted position, while with (iPr)₂N⁺=CH₂ ClO₄⁻ the reaction takes place at the less highly substituted position.³⁸³ The preformed iminium compound dimethyl(methylene)ammonium iodide (CH₂=N⁺Me₂ I⁻), called an Eschenmoser's salt,³⁸⁴ has also been used in Mannich reactions.³⁸⁵ The analogous chloride salt has been condensed with an imine to give a β,β′-dimethylamino ketone after acid hydrolysis.³⁸⁶

Another type of preformed reagent (29) has been used to carry out diastereoselective Mannich reactions. The lithium salts (28) are treated with TiCl₄ to give 29, which is then treated with the enolate anion of a ketone.³⁸⁷ The Pd catalyzed Mannich reaction of enol ethers to imines is also known.³⁸⁸ The reaction of silyl enol ethers and imines³⁸⁹ is catalyzed by HBF₄ in aq methanol.³⁹⁰ Similarly, silyl enol ethers react with aldehydes and aniline in the presence of InCl₃ to give the β-amino ketone.³⁹¹

Enantioselective Mannich reactions are known.³⁹² Chiral catalysts are commonly used,³⁹³ including proline,³⁹⁴ proline derivatives, or proline analogues,³⁹⁵ a Pybox-La catalyst,³⁹⁶ chiral aminosulfonamides,³⁹⁷ or Cinchona alkaloids,³⁹⁸ and other chiral amines.³⁹⁹ Chiral Brnsted acids are also used as catalysts,⁴⁰⁰ as well as chiral ammonium salts.⁴⁰¹ Chiral diamine⁴⁰² or phosphine-imine⁴⁰³ ligands have been used, and chiral dinuclear zinc compounds.⁴⁰⁴ Chiral auxiliaries on the carbonyl fragment can be used.⁴⁰⁵ Chiral imines, in the form of chiral hydrazones, have been used with silyl enol ethers and a Sc catalyst.⁴⁰⁶ Chiral amines react with aldehydes, with silyl enol ethers and an InCl₃ catalyst in ionic liquids, to give the Mannich product with good enantioselectivity.⁴⁰⁷ A chiral thiourea catalyst has been used with a vinylogous Mannich reaction⁴⁰⁸ (see Sec. 6.B for vinylogy).

The reaction of nitroalkanes and amines, usually in the presence of a metal catalyst (e.g., CuBr), ⁴⁰⁹ has been called the nitro-Mannich reaction.⁴¹⁰ An asymmetric nitro-Mannich reaction used a Cu–Sm catalyst,⁴¹¹ a Cu catalyst,⁴¹²or a chiral thiourea catalyst.⁴¹³

Also See, 11-22.

OS III, 305; IV, 281, 515, 816; VI, 474, 981, 987; VII, 34. See also, OS VIII, 358.

16-20 The Addition of Amines to Isocyanates

N-Hydro-C-alkylamino-addition

Ammonia and primary and secondary amines can be added to isocyanates⁴¹⁴ to give substituted ureas.⁴¹⁵ Isothiocyanates give thioureas.⁴¹⁶ This is an excellent method for the preparation of ureas and thioureas. These compounds are often used as derivatives for primary and secondary amines. Isocyanic acid (HNCO) also gives the reaction; usually its salts (e.g., NaNCO) are used. Wöhler's famous synthesis of urea involved the addition of ammonia to a salt of this acid.⁴¹⁷

OS II, 79; III, 76, 617, 735; IV, 49, 180, 213, 515, 700; V, 555, 801, 802, 967; VI, 936, 951; VIII, 26.

16-21 The Addition of Ammonia or Amines to Nitriles

N-Hydro-C-amino-addition

Unsubstituted amidines (in the form of their salts) can be prepared by addition of ammonia to nitriles.⁴¹⁸ Many amidines have been made in this way. Dinitriles of suitable chain length can give imidines (30).⁴¹⁹

Primary and secondary amines can be used instead of ammonia, to give substituted amidines, but only if the nitrile contains electron-withdrawing groups; for example, Cl₃CCN gives the reaction. Ordinary nitriles do not react, and, in fact, acetonitrile is often used as a solvent in this reaction.⁴²⁰ Ordinary nitriles can be converted to amidines by treatment with an alkylchloroaluminum amide [MeAl(Cl)NR₂; R = H or Me).⁴²¹ The addition of ammonia to cyanamide (NH₂CN) gives guanidine [(NH₂)₂C=NH]. Guanidines can also be formed from amines.⁴²²

If water is present, in the presence of a Ru⁴²³ or a Pt catalyst,⁴²⁴ the addition of a primary or secondary amine to a nitrile gives an amide: RCN + R¹NHR² + H₂O → RCONR¹R² + NH₃ (R² may be H). When benzonitrile reacts with H₂PO₃Se⁻ in aq methanol, a selenoamide [Ph(C=Se)NH₂], is formed after treatment with aq potassium carbonate.⁴²⁵

OS I, 302 [but also see, OS V, 589]; IV, 245, 247, 515, 566, 769. See also, OS V, 39.

16-22 The Addition of Amines to Carbon Disulfide and Carbon Dioxide

S-Metallo-C-alkylamino-addition

Salts of dithiocarbamic acid can be prepared by the addition of primary or secondary amines to carbon disulfide.⁴²⁶ This reaction is similar to 16-10. Hydrogen sulfide can be eliminated from the product, directly or indirectly, to give isothiocyanates (RNCS). Isothiocyanates can be obtained directly by the reaction of primary amines and CS₂ in pyridine in the presence of dicyclohexylcarbodiimide.⁴²⁷ A tosyl chloride mediated preparation of isothiocyanates is also known.⁴²⁸ Aniline derivatives react with CS₂ and NaOH, and then ethyl chloroformate to give the aryl isothiocyanate.⁴²⁹ In the presence of diphenyl phosphite and pyridine, primary amines add to CO₂ and to CS₂ to give, respectively, symmetrically substituted ureas and thioureas:⁴³⁰ Isoselenoureas [R₂NC(=NR¹)SeR²] can also be formed.⁴³¹

OS I, 447; III, 360, 394, 599, 763; V, 223.

E. Halogen Nucleophiles

16-23 The Formation of gem-Dihalides from Aldehydes and Ketones

Dihalo-de-oxo-bisubstitution

Aliphatic aldehydes and ketones can be converted to gem-dichlorides⁴³² by treatment with PCl₅. The reaction fails for perhalo ketones.⁴³³ If the aldehyde or ketone has an α hydrogen, elimination of HCl may follow and a vinylic chloride is a frequent side product, as shown,⁴³⁴ or even the main product.⁴³⁵ Phosphorus pentabromide (PBr₅) does not give good yields of gem-dibromides,⁴³⁶ but these can be obtained from aldehydes, by the use of Br₂, and triphenyl phosphite.⁴³⁷gem-Dichlorides can be prepared by reacting an aldehyde with BiCl₃.⁴³⁸

The mechanism of gem-dichloride formation involves initial attack on PCl₄⁺, which is present in solid PCl₅, by the carbonyl oxygen, followed by addition of Cl⁻ to the carbon:⁴³⁹

This chloride ion may come from PCl₆⁻, which is also present in solid PCl₅. There follows a two-step S_N1 process. Alternatively, 31 can be converted to the product without going through the chlorocarbenium ion, by an S_Ni process.

This reaction has sometimes been performed on carboxylic esters, though these compounds very seldom undergo any addition to the C=O bond. An example is the conversion of F₃CCOOPh to F₃CCCl₂OPh.⁴⁴⁰ However, formates commonly give the reaction.

Many aldehydes and ketones have been converted to gem-difluoro compounds with sulfur tetrafluoride (SF₄),⁴⁴¹ including quinones, which give 1,1,4,4-tetrafluorocyclohexadiene derivatives. With ketones, yields can be raised and the reaction temperature lowered, by the addition of anhydrous HF.⁴⁴² Carboxylic acids, acyl chlorides, and amides react with SF₂ to give 1,1,1-trifluorides. In these cases, the first product is the acyl fluoride, which then undergoes the gem-difluorination reaction:

The acyl fluoride can be isolated. Carboxylic esters also give trifluorides, but more vigorous conditions are required. In this case, the carbonyl group of the ester is attacked first, and RCF₂OR′ can be isolated from RCO₂R′⁴⁴³ and then converted to the trifluoride. Anhydrides can react in either manner. Both types of intermediate are isolable under the right conditions, and SF₄ even converts CO₂ to CF₄. A disadvantage of reactions with SF₄ is that they require a pressure vessel lined with stainless steel. Selenium tetrafluoride (SeF₄) gives similar reactions, but atmospheric pressure and ordinary glassware can be used.⁴⁴⁴ Another reagent that is often used to convert aldehydes and ketones to gem-difluorides is the commercially available diethylaminosulfur trifluoride (DAST, Et₂NSF₃), and CF₂Br₂ in the presence of zinc.⁴⁴⁵ The mechanism with SF₄ is probably similar in general nature, if not in specific detail, to that with PCl₅. Some dithianes can be converted to gem-difluorides with a mixture of fluorine and iodine in acetonitrile.⁴⁴⁶ Oximes give gem-difluorides with NO⁺BF₄⁻ and pyridinium polyhydrogen fluoride.⁴⁴⁷

Treatment with hydrazine to give the hydrazone, and then CuBr₂/t-BuOLi, generated the gem-dibromide.⁴⁴⁸ Oximes give gem-dichlorides upon treatment with chlorine and BF₃·OEt₂, and then HCl.⁴⁴⁹

In a related process, α-halo ethers can be prepared by treatment of aldehydes and ketones with an alcohol and HX. The reaction is applicable to aliphatic aldehydes and ketones and to primary and secondary alcohols. The addition of HX to an aldehyde or ketone gives α-halo alcohols, which are usually unstable, although exceptions are known, especially with perfluoro and perchloro species.⁴⁵⁰

OS II, 549; V, 365, 396, 1082; VI, 505, 845; VIII, 247. Also see, OS I, 506. For α-halo-ethers, see OS I, 377; IV, 101 (see, however, OS V, 218), 748; VI, 101.

F. Attack at Carbon by Organometallic Compounds⁴⁵¹

16-24 The Addition of Grignard Reagents and Organolithium Reagents to Aldehydes and Ketones

O-Hydro-C-alkyl-addition

Organomagnesium compounds, commonly known as Grignard reagents (RMgX), are formed by the reaction of alkyl, vinyl, or aryl halides with magnesium metal, usually in ether solvents (e.g., diethyl ether or THF; Reaction 12-38). Halogen–Mg exchange can generate a Grignard reagent by reaction of aryl halides with reactive aliphatic Grignard reagents.⁴⁵² Microwave irradiation has been used to facilitate the formation of Grignard reagents from aryl chlorides that are slow to react otherwise.⁴⁵³

The addition of Grignard reagents to aldehydes and ketones⁴⁵⁴ is known as the Grignard reaction.⁴⁵⁵ The initial product of reaction with a carbonyl is a magnesium alkoxide, requiring a hydrolysis step to generate the final alcohol product. Formaldehyde gives primary alcohols; other aldehydes give secondary alcohols; and ketones give tertiary alcohols. The reaction is of very broad scope. In many cases, the hydrolysis step is carried out with dilute HCl or H₂SO₄, but this cannot be done for tertiary alcohols in which at least one R group is alkyl because such alcohols are easily dehydrated under acidic conditions (Reaction 17-1). In such cases (and often for other alcohols as well), an aqueous solution of ammonium chloride is used instead of a strong acid. Grignard reagents have been used in solid-phase synthesis.⁴⁵⁶ Ionic liquids have been used for the Grignard reaction.⁴⁵⁷

Transition metal catalysts can promote 1,2-addition of Grignard reagents to ketones. A catalytic amount of Zn(II) compounds promote the reaction, for example.⁴⁵⁸ In the presence of a catalytic amount of InCl₃, Grignard reagentsreact to give a mixture of 1,2- and 1,4-addition products with the 1,4-product predominating, but there was an increased 1,2-addition relative to the uncatalyzed reaction.⁴⁵⁹

Diastereoselective addition⁴⁶⁰ has been carried out with achiral reagents and chiral substrates,⁴⁶¹ similar to the reduction shown in Reaction 19-36.⁴⁶² Because the attacking atom in this case is carbon, diastereoselective addition is possible with an achiral substrate and an optically active reagent.⁴⁶³ The use of suitable reactants creates, in the most general case, two new stereogenic centers, so the product can exist as two pairs of enantiomers, as shown. Even if the organometallic compound is racemic, it still may be possible to get a diastereoselective reaction; that is, one pair of enantiomers is formed in greater amount than the other.⁴⁶⁴

Asymmetric Grignard reactions are possible under certain circumstances.⁴⁶⁵ Chiral ligands with a chiral Cu catalyst⁴⁶⁶ or a chiral Ti complex⁴⁶⁷ give alcohols with good enantioselectivity. An interesting method formed using an alkylmagnesium halide, dibutylmagnesium (Bu₂Mg) and a chiral diamine, and subsequent reaction with an aldehyde led to the alcohol derived from acyl addition of a butyl group with good enantioselectivity.⁴⁶⁸ N-Heterocyclic carbenes have been used as organocatalysts for asymmetric Grignard reactions.⁴⁶⁹ Aryl iodides undergo halogen–magnesium exchange when pretreated with PhMgCl, and subsequent reaction with an aldehyde gives the alcohol.⁴⁷⁰

The reaction of aldehydes or ketones with alkyl and aryl Grignard reagents was done in the earliest work without preliminary formation of RMgX, by mixing RX, the carbonyl compound, and magnesium metal in an ether solvent. This approach preceded Grignard's work, and is now known as the Barbier reaction.⁴⁷¹ The organolithium analogue of this process is also known.⁴⁷² Yields were generally satisfactory. Carboxylic ester, nitrile, and imide groups in the R are not affected by the reaction conditions.⁴⁷³ Modern versions of the Barbier reaction employ other metals and/or reaction conditions, and will be discussed in Reaction 16-25. However, Mg–Barbier reactions are catalyzed by other metal complexes (e.g., Cu compounds).⁴⁷⁴ Some transition metal compounds are stable in water, so some Grignard–Barbier reactions can be done in water.⁴⁷⁵ A retro-Barbier reaction has been reported in which a cyclic tertiary alcohol was treated to an excess of bromine and potassium carbonate to give 6-bromo-2-hexanone from 1-methylcyclopentanol.⁴⁷⁶

The reaction of RMgX or RLi with α,β-unsaturated aldehydes or ketones can proceed via 1,4-addition as well as normal 1,2-addition (see Michael addition in Reaction 15-25).⁴⁷⁷ In general, alkyllithium reagents give less 1,4 addition than the corresponding Grignard reagents.⁴⁷⁸ In a compound containing both an aldehyde and a ketone, it is possible to add RMgX chemoselectively to the aldehyde without significantly disturbing the carbonyl of the ketone group⁴⁷⁹ (see also, Reaction 16-24). Grignard reagents have been shown to add to some conjugated cyclic ketones with an α,β-OTf group via 1,2-addition, followed by cleavage to give an alkynyl ketone.⁴⁸⁰

In some cases, a Grignard reaction can be performed intramolecularly.⁴⁸¹ For example, treatment of 5-bromo-2-pentanone with magnesium and a small amount of mercuric chloride in THF produced 1-methyl-1-cyclobutanol in 60% yield.⁴⁸² Other four- and five-membered ring compounds were also prepared by this procedure. Similar closing of five- and six-membered rings was achieved by treatment of a δ- or ε-halocarbonyl compound, not with a metal, but with a dianion derived from nickel tetraphenyporphine⁴⁸³ (see Reaction 16-25).

The gem-disubstituted Mg compounds formed from CH₂Br₂ or CH₂I₂ (Reaction 12-38) react with aldehydes or ketones to give alkenes in moderate-to-good yields.⁴⁸⁴Wittig-type reactions also produce alkenes and are discussed in Reaction 16-44. The reaction could not be extended to other gem-dihalides. Similar reactions with gem-dimetallic compounds prepared with metals other than magnesium also have produced alkenes.⁴⁸⁵

Organolithium reagents (RLi), prepared from alkyl halides and Li metal or by exchange of an alkyl halide with a reactive organolithium (Reaction 12-38) react with aldehydes and ketones by acyl addition to give the alcohol,⁴⁸⁶after hydrolysis. Organolithium reagents are more basic than the corresponding Grignard reagent, which leads to problems of deprotonation in some cases. Organolithium reagents are generally more nucleophilic, however, and can add to hindered ketones with relative ease when compared to the analogous Grignard reagent.⁴⁸⁷ These reagents tend to form aggregates, which influences the reactivity and selectivity of the addition reaction.⁴⁸⁸ The addition of lithium amide–butyllithium mixed aggregates has been studied.⁴⁸⁹

Alkyl, vinyl,⁴⁹⁰ and aryl organolithium reagents can be prepared and undergo acyl addition. Structural variations are also possible, including enantioselective 1,2-addition.⁴⁹¹ 1-Bromo-1-lithioethene was prepared, and reacts with an aldehyde to give an allylic alcohol bearing a vinyl bromide unit.⁴⁹² An interesting variation of the fundamental acyl addition reaction of organolithium reagents treated an aldehyde with an acyl-lithio amide [LiC(=O)N(Me)CH₂Me] to give an α-hydroxy amide derivative.⁴⁹³

As with the reduction of aldehydes and ketones (Reaction 19-36), the addition of organometallic compounds to these substrates can be carried out enantioselectively and diastereoselectively.⁴⁹⁴ Chiral secondary alcohols have been obtained with high enantioselectivity by addition of Grignard and organolithium compounds to aromatic aldehydes, in the presence of optically active amino alcohols as ligands.⁴⁹⁵

An interesting variation is the reaction of methyllithium and CH₂I₂ with an aliphatic aldehyde to give an epoxide.⁴⁹⁶ A lithio-epoxide was formed by treating an epoxide with sec-butyllithium in the presence of sparteine,⁴⁹⁷ or with n-butyllithium/TMEDA,⁴⁹⁸ and subsequent reaction with an aldehyde led to an epoxy alcohol. Alkylidene oxetanes react with lithium, and then with an aldehyde to give a conjugated ketone.⁴⁹⁹ The reaction of gem-dihalides with a carbonyl compound and Li or BuLi give epoxides⁵⁰⁰ (see also, Reaction 16-46).

In other uses of gem-dihalo compounds, aldehydes and ketones add the CH₂I group [R₂CO → R₂C(OH)CH₂I] when treated with CH₂I₂ in the presence of SmI₂,⁵⁰¹ and the CHX₂ group when treated with methylene halides and lithium dicyclohexylamide at low temperatures.⁵⁰²

It is possible to add an acyl group to a ketone to give (after hydrolysis) an α-hydroxy ketone.⁵⁰³ This can be done by adding RLi and CO to the ketone at −110 °C:⁵⁰⁴

When the same reaction is carried out with carboxylic esters (R′COOR²), α-diketones (RCOCOR′) are obtained.⁵⁰³

Most aldehydes and ketones react with most Grignard reagents, but there are several potential side reactions⁵⁰⁵ that occur mostly with hindered ketones and with bulky Grignard reagents. The two most important of these are enolization and reduction. The former requires that the aldehyde or ketone have an α hydrogen, and the latter requires that the Grignard reagent have a β hydrogen:

Enolization is an acid–base reaction (12-24) in which a proton is removed from the α carbon by the Grignard reagent, which is a strong base. The carbonyl compound is converted to its enolate anion, which, on hydrolysis, gives the original ketone or aldehyde. Enolization is important not only for hindered ketones but also for those that have a relatively high percentage of enol (e.g., β-keto esters).

The carbonyl compound can be reduced to an alcohol (Reaction 16-24) by the Grignard reagent, which itself undergoes elimination to give an alkene. The Grignard reagent must have a β-carbon that bears a hydrogen atom.

Two other side reactions are condensation (between enolate ion and excess ketone) and Wurtz-type coupling (10-64). Addition of Grignard reagents to ketones cannot be used to prepare highly hindered tertiary alcohols (e.g., triisopropylcarbinol, tri-tert-butylcarbinol, and diisopropylneopentylcarbinol) or they can be prepared only in extremely low yields, because reduction and/or enolization become prominent.⁵⁰⁶ However, these alcohols can be prepared by the use of alkyllithium reagents at –80 °C⁵⁰⁷ because enolization and reduction are much less important.⁵⁰⁸ Other methods of increasing the degree of addition at the expense of reduction include complexing the Grignard reagent with LiClO₄ or Bu₄N⁺ Br⁻,⁵⁰⁹ or using benzene or toluene instead of ether as solvent.⁵¹⁰ Both reduction and enolization can be avoided by adding CeCl₃ to the Grignard reagent.⁵¹¹

There has been controversy regarding the mechanism of addition of Grignard reagents to aldehydes and ketones.⁵¹² The reaction is difficult to study because of the variable nature of the species present in the Grignard solution (Sec. 5.B.ii) and because the presence of small amounts of impurities in the Mg seems to have a great effect on the kinetics of the reaction, making reproducible experiments difficult.⁵¹³ There seem to be two basic mechanisms, depending on the reactants and the reaction conditions. In one of these, the R group is transferred to the carbonyl carbon with its electron pair. A detailed mechanism of this type has been proposed by Ashby et al.,⁵¹⁴ based on the discovery that this reaction proceeds by two paths: one first order in MeMgBr and the other first order in Me₂Mg.⁵¹⁵ According to this proposal, both MeMgBr and Me₂Mg add to the carbonyl carbon, though the exact nature of the step by which MeMgBr or Me₂Mg reacts with the substrate is not certain. One possibility is a four-centered cyclic transition state:⁵¹⁶

The other type of mechanism is a SET process⁵¹⁷ with a ketyl intermediate:⁵¹⁸

This mechanism, which has been mostly studied with diaryl ketones, is more likely for aromatic and other conjugated aldehydes and ketones than it is for strictly aliphatic ones. Among the evidence⁵¹⁹ for the SET mechanism are ESR spectra⁵²⁰ and the fact that Ar₂C(OH)C(OH)Ar₂ side products are obtained (from dimerization of the ketyl; see pinacol coupling in Reaction 19-76).⁵²¹ In the case of addition of RMgX to benzil (PhCOCOPh), ESR spectra of two different ketyl radicals were observed, both reported to be quite stable at room temperature.⁵²² Note that a separate study failed to observe freely defusing radicals in the formation of Grignard reagents.⁵²³ Carbon isotope effect studies with Ph¹⁴COPh showed that the rate-determining step with most Grignard reagents is the carbon–carbon bond-forming step (marked A), although it is the initial electron-transfer step with allylmagnesium bromide.⁵²⁴ In the formation of Grignard reagents from bromocyclopropane, diffusing cyclopropyl radical intermediates were found.⁵²⁵ The concerted versus stepwise mechanism has been probed with chiral Grignard reagents.⁵²⁶

Note that there are similarities in reactivity for the S_RN1 (see Sec. 13.A.iv) and Grignard mechanisms.⁵²⁷ Experimental evidence from this work suggests a linear rather than a chain mechanism.

Mechanisms for the addition of organolithium reagents have been investigated much less.⁵²⁸ Addition of a cryptand that binds Li⁺ inhibited the normal addition reaction, showing that the lithium is necessary for the reaction to take place.⁵²⁹

There is general agreement that the mechanism leading to reduction⁵³⁰ is usually as follows:

There is evidence that the mechanism leading to enolization is also cyclic, but involves prior coordination with magnesium:⁵³¹

Aromatic aldehydes and ketones can be alkylated and reduced in one reaction vessel by treatment with an alkyl- or aryllithium, followed by lithium and ammonia and then by ammonium chloride.⁵³²

A similar reaction has been carried out with N,N-disubstituted amides: RCONR₂' → RR²CHNR₂'.⁵³³

OS I, 188; II, 406, 606; III, 200, 696, 729, 757; IV, 771, 792; V, 46, 452, 608, 1058; VI, 478, 537, 542, 606, 737, 991, 1033; VII, 177, 271, 447; VIII, 179, 226, 315, 343, 386, 495, 507, 556; IX, 9, 103, 139, 234, 306, 391, 472; 75, 12; 76, 214; X, 200.

16-25 Addition of Other Organometallics to Aldehydes and Ketones

O -Hydro-C-alkyl-addition

A variety of organometallic reagents other than RMgX and RLi add to aldehydes and ketones. A simple example is formation of Na, or K alkyne anions (e.g., RCC-M, Reaction 16-38), which undergo acyl addition to ketones or aldehydes to give the propargylic alcohol.⁵³⁴ In the reaction with terminal acetylenes,⁵³⁵ sodium acetylides are the most common reagents (when they are used, While Na is the metal of choice for the addition of acetylenic groups, vinylic alanes (prepared as in Reaction 15-17) are the reagents of choice for the addition of vinylic groups.⁵³⁶ The reagent Me₃Al/⁻CCH Na⁺ also adds to aldehydes to give the ethynyl alcohol.⁵³⁷ A solvent-free reaction was reported that mixed a ketone, a terminal alkyne and potassium tert-butoxide.⁵³⁸ The reaction is often called the Nef reaction, but Li,⁵³⁹ Mg, and other metallic acetylides have also been used. A particularly convenient reagent is the lithium acetylide–ethylenediamine complex,⁵⁴⁰ a stable, free-flowing powder that is commercially available. Alternatively, the substrate may be treated with the alkyne itself in the presence of a base, so that the acetylide is generated in situ. This procedure is called the Favorskii reaction, not to be confused with the Favorskii rearrangement (18-7).⁵⁴¹ Zinc(II) chloride facilitates the addition of a terminal alkyne to an aldehyde to give a propargylic alcohol.⁵⁴²Zinc(II) triflate can also be used for alkyne addition to aldehydes,⁵⁴³ and in the presence of a chiral ligand leads to good enantioselectivity in the propargyl alcohol product.⁵⁴⁴

The reagents Et₃Al, Et₂Zn, and a terminal alkyne react with ketones, and in the presence of a cinchona alkaloid gives the alkynyl alcohol in moderate ee.⁵⁴⁵ Other enantioselective alkynylation reactions are known using various catalysts.⁵⁴⁶ Terminal alkynes add to aryl aldehydes in the presence of InBr₃ and NEt₃,⁵⁴⁷ SmI₂,⁵⁴⁸ or Me₂Zn.⁵⁴⁹ A Zn mediated reaction using iodoalkynes is known.⁵⁵⁰ Catalytically generated zinc acetylides add to aldehydes.⁵⁵¹ An In catalyzed addition of alkynes to aldehydes used a catalytic amount of BINOL and gave the alkynyl alcohol with high enantioselectivity.⁵⁵² Other enantioselective addition reactions of terminal alkynes are known.⁵⁵³ The reaction with In is compatible with the presence of a variety of other functional groups in the molecule, including phosphonate,⁵⁵⁴ propargylic sulfides.⁵⁵⁵

Many organometallic reagents have been reported for the addition of allylic groups,⁵⁵⁶ and there are enantioselective reactions.⁵⁵⁷ One of the most common methods is the Barbier reaction noted in Reaction 16-24, which includes metals and metal compounds other than Mg or Li, although the method is not limited to allylic compounds. Common reagents are allylic In compounds,⁵⁵⁸ which add to aldehydes or ketones in various solvents,⁵⁵⁹ Enantioselective In mediated Barbier reactions are known.⁵⁶⁰ Indium reacts with allylic bromides and ketones in water⁵⁶¹ and in aqueous media. The reaction of a propargyl halide, In, and an aldehyde in aq THF leads to an allenic alcohol.⁵⁶² When allyl iodide is mixed with In and TMSCl, reaction with a conjugated ketone proceed by 1,4-addition, but in the presence of 10% CuI, the major product is that of 1,2-addition.⁵⁶³ Allyl bromide reacts with Mn/TMSCl and an In catalyst in water to give the homoallylic alcohols from aldehydes.⁵⁶⁴ Indium metal is used for the acyl addition of allylic halides with a variety of aldehydes and ketones, including aliphatic aldehydes,⁵⁶⁵ aryl aldehydes,⁵⁶⁶ and α-keto esters.⁵⁶⁷ Elimination of the homoallylic alcohol to a conjugated diene can accompany the addition in some cases.⁵⁶⁸

Allyl bromide reacts with a ketone and Sm⁵⁶⁹ or SmI₂,⁵⁷⁰ to give the homoallylic alcohol. Other metals can be used with allylic halides to give homoallylic alcohols from aldehydes or ketones,⁵⁷¹, including Zn,⁵⁷² La,⁵⁷³ or Mg-Cd⁵⁷⁴ and compounds of Ti,⁵⁷⁵ Mn,⁵⁷⁶ Fe,⁵⁷⁷ Ga,⁵⁷⁸ Ge,⁵⁷⁹ Zr,⁵⁸⁰ Nb,⁵⁸¹ Cd,⁵⁸² Sb,⁵⁸³ Te,⁵⁸⁴ Ba,⁵⁸⁵ Ce,⁵⁸⁶ Nd,⁵⁸⁷ Hg,⁵⁸⁸ Bi,⁵⁸⁹ In,⁵⁹⁰ and Pb.⁵⁹¹ In addition, BiCl₃/NaBH₄,⁵⁹² Mg–BiCl₃,⁵⁹³ and CrCl₂/NiCl₂,⁵⁹⁴ have been used. Allylic alcohols have been converted to organometallic reagents with diethylzinc and a Pd⁵⁹⁵ or a Ru catalyst,⁵⁹⁶ leading to the homoallylic alcohol upon reaction with an aldehyde. Allylic acetates add to aldehydes in the presence of a Ru catalyst.⁵⁹⁷

Enantioselective reactions are known.⁵⁹⁸ A Cu catalyzed reaction is known that proceeds with good enantioselectivity.⁵⁹⁹ Allylzinc bromide adds to aldehydes under solvent-free conditions.⁶⁰⁰ A chiral Cr–Mn complex has been used with allylic bromides in conjunction with trimethylsilyl chloride.⁶⁰¹ Reagents of the type R–Yb have been prepared from RMgX.⁶⁰² The alkyl group of trialkyl aluminum compounds (e.g., AlEt₃) add to aldehydes, enantioselectively in the presence of chiral transition metal complexes.⁶⁰³ Furthermore, organotitanium reagents can be made to add chemoselectively to aldehydes in the presence of ketones.⁶⁰⁴ Organomanganese compounds are also chemoselective in this way.⁶⁰⁵ In a related reaction, organocerium reagents, generated from cerium chloride (CeCl₃ and a Grignard reagent or an organolithium reagent) gives an organometallic reagent that adds chemoselectively.⁶⁰⁶ Aryl halides that have a pendant ketone unit react with a Pd catalyst to give cyclization via acyl addition.⁶⁰⁷

As noted in Reaction 16-24, enolate formation and reduction complicate some Grignard reactions. One way to avoid complications is to add (RO)₃TiCl, TiCl₄,⁶⁰⁸ (RO)₃ZrCl, or (R₂N)₃TiX to the Grignard or organolithium reagent. This produces organotitanium or organozirconium compounds that are much more selective than Grignard or organolithium reagents.⁶⁰⁹ An important advantage of these reagents is that they do not react with NO₂ or CN functions that may be present in the substrate, as Grignard and organolithium reagents do. The reaction of a β-keto amide with TiCl₄, for example, gives a complex that allows selective reaction of the ketone unit with MeMgCl–CeCl₃ to give the corresponding alcohol.⁶¹⁰ Premixing an allylic Grignard reagent with ScCl₃ prior to reaction with the aldehyde gives direct acyl addition without allylic rearrangement as the major product, favoring the trans-alkene unit.⁶¹¹

Allyltin compounds readily add to aldehydes and ketones.⁶¹² Maleic acid promotes the reaction in aqueous media.⁶¹³ Allylic bromides react with Sn to generate the organometallic in situ, which then adds to aldehydes.⁶¹⁴ Allylic chlorides react with aldehydes in the presence of ditin compounds (e.g., Me₃Sn–SnMe₃ and a Pd catalyst).⁶¹⁵ Allyltrialkyltin compounds⁶¹⁶ and tetraallyltin react with aldehydes or ketones in the presence of BF₃–etherate,⁶¹⁷ or compounds of Cu,⁶¹⁸ Ce,⁶¹⁹ Bi,⁶²⁰ Pb,⁶²¹ Ag⁶²² Cd,⁶²³ Cr,⁶²⁴ Pd,⁶²⁵ Re,⁶²⁶ Gd,⁶²⁷ Ti,⁶²⁸ Rh,⁶²⁹ Zr,⁶³⁰ Co,⁶³¹ or La.⁶³² Propargylic tin compounds react with aldehydes to give the alcohol, with good antiselectivity.⁶³³ Tetraallyltin reacts via 1,2-addition to conjugated ketones in refluxing methanol.⁶³⁴ Tetraallyltin reacts with aldehydes in ionic liquids⁶³⁵ and on wet silica.⁶³⁶ In addition, allyltributyltin adds to aldehydes in ionic liquids with InCl₃.⁶³⁷ Tetraallyltin adds to ketones or aldehydes to give homoallylic alcohols with good enantioselectivity in the presence of a chiral Ti complex⁶³⁸ or a chiral In complex.⁶³⁹ Allylic alcohols and homoallylic alcohols add to aldehydes in the presence of Sn(OTf)₂⁶⁴⁰ In/InCl₃,⁶⁴¹ or with a Rh catalyst.⁶⁴²

The tin compound can be prepared in situ using an α-iodo ketone with an aldehyde and Bu₂SnI₂–LiI.⁶⁴³ A similar addition occurs with (allyl)₂SnBr₂ in water.⁶⁴⁴ Asymmetric induction has been reported.⁶⁴⁵ The use of a chiral Rh⁶⁴⁶ or Ti⁶⁴⁷ catalyst leads to enantioselective addition of allyltributyltin to aldehydes. Allyltributyltin reacts with aldehydes in the presence of SiCl₄ and a chiral phosphoramide to give the homoallylic alcohol with moderate enantioselectivity.⁶⁴⁸ Tetravinyltin adds to ketones in the presence of an In catalyst.⁶⁴⁹ Allenyl tin compounds (CH₂=C=CHSnBu₃) also react with aldehydes in the presence of BF₃·OEt₂ to give a 2-dienyl alcohol.⁶⁵⁰ Selectfluor has been used to induce 1,2-addition of the allyl group of allyltributyltin to a conjugated aldehyde.⁶⁵¹

Both aluminum and titanium reagents have been used. Aluminum catalysts, (e.g., [methylaluminum bis(4-bromo-2,6-di-tert-butylphenoxide)], MABR), facilitate addition of allyltributyltin to aldehydes.⁶⁵² Triphenylaluminum reacts with aryl aldehydes in the presence of a Ti catalyst.⁶⁵³ Trimethylaluminum⁶⁵⁴ and dimethyltitanium dichloride⁶⁵⁵ exhaustively methylate ketones to give gem-dimethyl compounds⁶⁵⁶ (see also, Reaction 10-63):

The Ti reagent also dimethylates aromatic aldehydes.⁶⁵⁷ Triethylaluminum reacts with aldehydes, however, to give the mono-ethyl alcohol, and in the presence of a chiral additive the reaction proceeds with good asymmetric induction.⁶⁵⁸ A complex of Me₃Ti·MeLi has been shown to be selective for 1,2-addition with conjugated ketones, in the presence of nonconjugated ketones.⁶⁵⁹

Chiral amides react with aldehydes in the presence of TiCl₄ to give syn-selective addition products.⁶⁶⁰ Titanium-catalyzed enantioselective additions are also known.⁶⁶¹ High %ee values have been obtained with organometallics,⁶⁶²including organotitanium compounds (methyl, aryl, allylic) in which an optically active ligand is coordinated to the Ti,⁶⁶³ allylic boron compounds, and organozinc compounds. Chiral dendritic Ti catalysts have been used to give moderate enantioselectivity.⁶⁶⁴

Organozinc compounds add to aldehydes and ketones. An example is the reaction of an alkylzinc chloride (RZnCl) to give the corresponding alcohol.⁶⁶⁵ Enantioselective reaction of a carbonyl with a dialkylzinc is possible when chiral catalysts are employed,⁶⁶⁶ or when chiral ligands⁶⁶⁷ are employed.⁶⁶⁸ A comparison of the stereoselectivity for reactions of diphenylzinc and diethylzinc has been reported.⁶⁶⁹ Dialkylzinc reagents, in the presence of a chiral Ti complex⁶⁷⁰ a Zn complex,⁶⁷¹ an Al complex,⁶⁷² a chiral Cr complex,⁶⁷³ a chiral Schiff base,⁶⁷⁴ or a chiral bis(sulfonamide⁶⁷⁵ and other chiral complexes⁶⁷⁶ react with aldehydes or ketones to give the corresponding alcohol with good enantioselectivity.⁶⁷⁷ High enantioselectivity was obtained from R₂Zn reagents (R = alkyl)⁶⁷⁸ and aromatic⁶⁷⁹ aldehydes by the use of a small amount of various catalysts.⁶⁸⁰ The enantioselectivity is influenced by additives (e.g., LiCl).⁶⁸¹ Silica-immobilized chiral ligands⁶⁸² can be used in conjunction with dialkylzinc reagents, and polymer-supported ligands have been used.⁶⁸³

Propargylic acetate adds to aldehydes with good antiselectivity in the presence of Et₂Zn and a Pd catalyst.⁶⁸⁴ With other organometallic compounds, active metals (e.g., alkylzinc reagents)⁶⁸⁵ are useful and compounds, such as alkylmercurys, do not react. Dimethylzinc and diethylzinc are probably the most common reagents. An intramolecular version is possible by reaction of an allene–aldehyde with dimethylzinc. Aryl halides react with Zn–Ni complexes to give acyl addition of the aryl group to an aldehyde.⁶⁸⁶ The reaction of an allylic halide and Zn⁶⁸⁷ or Zn/TMSCl⁶⁸⁸ leads to acyl addition of aldehydes.

Organochromium compounds add to aldehydes or ketones.⁶⁸⁹ The reaction of an organochromium reagent with an aldehyde or ketone is known as the Nozaki–Hiyama reaction. In the original version, a Cr(II) solution was prepared by reduction of chromic chloride by LiAlH₄. This product was subsequently treated with an aldehyde and an allylic halide.⁶⁹⁰ The coupling of allylic halides and aldehydes or ketones in the presence of a Cr catalyst and a chiral ligand gives products with good enantioselectivity.⁶⁹¹ Enantioselective coupling reactions catalyzed by Cr compounds are of increasing interest.⁶⁹² The organochromium reagent may be derived from vinyl halides, triflates, or aryl derivatives.⁶⁹³

Other metals facilitate addition of groups to an aldehyde, including the coupling of an alkene to an aldehyde using a Ni catalyst.⁶⁹⁴ Vinyl bromides react with NiBr₂/CrCl₃/TMSCl to give a reagent that adds to aldehydes to give the allylic alcohol.⁶⁹⁵ Vinyl complexes generated from alkynes and SmI₂ add intramolecularly, and eight-membered rings have been formed in this way.⁶⁹⁶ Vinyl reagents are formed in situ via organozirconium compounds with Me₂Zn, Ti compounds, and terminal alkynes.⁶⁹⁷

Lithium dimethylcopper (Me₂CuLi) reacts with aldehydes⁶⁹⁸ and with certain ketones⁶⁹⁹ to give the expected alcohols. The RCu(CN)ZnI reagents also react with aldehydes, in the presence of BF₃–etherate, to give secondary alcohols. Vinyltellurium compounds react with BF₃·OEt₂ and cyanocuprates [R(2-thienyl)CuCNLi₂] to give a reagent that adds 1,2- to the carbonyl of a conjugated ketone.⁷⁰⁰ Vinyl tellurium compounds also react with n-butyllithium to give a reagent that adds to nonconjugated ketones.⁷⁰¹ In conjunction with BeCl₂, organolithium reagents add to conjugated ketones. In THF, 1,4- addition is observed, but in diethyl ether the 1,2-addition product is formed.⁷⁰²

Alkenes and alkynes add to aldehydes or ketones via the π bond by conversion to a reactive organometallic. A radical-type addition is possible using alkenes with BEt₃. Alkynes add to aldehydes elsewhere in the same molecule in the presence of BEt₃ and a Ni catalyst to give a cyclic allylic alcohol.⁷⁰³ Alkene aldehydes react similarly using Me₃SiOTf.⁷⁰⁴ In a similar manner, dienes⁷⁰⁵ or alkynes⁷⁰⁶ add to aldehydes in the presence of a Ni catalyst. Allenes add to aldehydes in the presence of a Ni catalyst, using a chiral imidazolinyl carbene ligand, and the product is trapped as the triethylsilyl ether by addition of Et₃SiI.⁷⁰⁷ Ketenimines add to aldehydes, in the presence of SiCl₄ and an achiral ligand, to give the β-cyanohydrin with good enantioselectivity.⁷⁰⁸ Allylic acetates react with ketones to give the homoallylic alcohol under electrochemical conditions that include bipyridyl, tetrabutylammonium tetrafluoroborate, and FeBr₂.⁷⁰⁹ Terminal alkynes react with Zr complexes and Me₂Zn to give an allylic tertiary alcohol.⁷¹⁰ Internal alkynes also give allylic alcohols in the presence of BEt₃ and a Ni catalyst.⁷¹¹ Reaction of an aldehyde containing a conjugated diene unit with diethylzinc and a Ni catalyst leads to cyclic alcohols having a pendant allylic unit.⁷¹² A similar reaction was reported using a Cu catalyst.⁷¹³ The intramolecular addition of an alkene to an aldehyde leads to a saturated cyclic alcohol using PhSiH₃ and a Co catalyst.⁷¹⁴

Aryl halides react with a Ni complex under electrolytic conditions to add the aryl group to aldehydes.⁷¹⁵ The C position of an indole adds to aldehydes in the presence of a Pd catalyst.⁷¹⁶ The addition of trifluoromethyl to an aldehyde was accomplished photochemically using CF₃I and (Me₂N)₂C=C(NMe₂)₂.⁷¹⁷ α-Iodo phosphonate esters react with aldehydes and SmI₂ to give a β-hydroxy phosphonate ester.⁷¹⁸ Addition to the allene in the presence of a Ni catalyst⁷¹⁹ or a CeCl₃ catalyst⁷²⁰ is followed by addition of the intermediate organometallic to the aldehyde to give the cyclic product.

Intramolecular addition of a conjugated ester (via the β-carbon) to an aldehyde generates a cyclic ketone.⁷²¹ This type of coupling has been called the Stetter reaction,⁷²² which actually involves the addition of aldehydes to activated double bonds (Reaction 15-34), mediated by a catalytic amount of thiazolium salt in the presence of a weak base. The intramolecular addition of the allene moiety to an aldehyde is catalyzed by a Pd complex in the presence of Me₃SiSnBu₃.⁷²³ A highly enantio- and diastereoselective intramolecular Stetter reaction has been developed.⁷²⁴ Alkynyl aldehydes react with silanes (e.g., Et₃SiH) and a nickel catalyst to give a cyclic compound having a silyl ether and an exocyclic vinylidene unit.⁷²⁵ Alkene-aldehydes give cyclic alcohols via intramolecular addition of the C=C unit to the carbonyl under electrolytic conditions using a phase-transfer catalyst.⁷²⁶ A similar cyclization was reported using SnCl₄.⁷²⁷ Vinylidene cycloalkanes react with aldehydes in the presence of a Pd catalyst to give a homoallylic alcohol where addition occurs at the carbon exocyclic to the ring.⁷²⁸ Allenes react with benzaldehyde using HCl–SnCl₂ with a Pd catalyst.⁷²⁹ Silyl allenes react with aldehydes in the presence of a chiral Sc catalyst to give homopropargylic alcohols with good enantioselectivity.⁷³⁰ Intramolecular addition of an allene to aldehyde via addition of phenyl when treated with PhI and a Pd catalyst.⁷³¹ Allenes add to ketones to give homoallylic alcohols in the presence of SmI₂ and HMPA.⁷³² Alkenes having an allylic methyl group in the presence of BF₃·OEt₂ add to formaldehyde to give a homoallylic alcohol.⁷³³ Conjugated dienes react with aldehyde via acyl addition of a terminal carbon of the diene, in the presence of Ni(acac)₂ and Et₂Zn.⁷³⁴

Although organoboranes do not generally add to aldehydes and ketones,⁷³⁵ allylic boranes are exceptions.⁷³⁶ When they add, an allylic rearrangement always takes place. Allylic rearrangements may take place with the other reagents as well. The use of a chiral catalyst leads to asymmetric induction⁷³⁷ and chiral allylic boranes have been prepared.⁷³⁸ Addition of trialkylboranes to aldehydes is catalyzed by a Ni complex.⁷³⁹ Note that chloroboranes (R₂BCl) react with aldehydes via acyl addition of the alkyl group, giving the corresponding alcohol after treatment with water.⁷⁴⁰ Treatment with catecholborane gives addition to the conjugated ketone, and subsequent cyclization of the resulting organometallic at the nonconjugated ketone gives a cyclic alcohol with a pendant ketone unit, after treatment with methanol.⁷⁴¹ In the presence of Ru⁷⁴² Cu,⁷⁴³ Ni,⁷⁴⁴ or Pd⁷⁴⁵ compounds, RB(OH)₂ and arylboronic acids [ArB(OH)₂, see Reaction 12-28] add to aldehydes to give the corresponding alcohol. Arylboronic acids add to aldehydes in the presence of a chiral ligand to give an alcohol with good enantioselectivity.⁷⁴⁶ An enantioselective intramolecular reaction of an arylboronic acid to a pendant ketone moiety used a Pd catalyst.⁷⁴⁷ Polymer-bound aryl borates add an aryl group to aldehydes in the presence of a Rh catalyst.⁷⁴⁸ An intramolecular version of the phenylboronic acid induced reaction is known, where a molecule with ketone and conjugated ketone units is converted to a cyclic alcohol using a chiral Rh catalyst.⁷⁴⁹ Allylic boronates add to aldehydes,⁷⁵⁰ and there are enantioselective reactions.⁷⁵¹

A number of optically active allylic boron compounds have been used, including⁷⁵² B-allylbis(2-isocaranyl)borane (32),⁷⁵³ (E)- and (Z)-crotyl-(R,R)-2,5-dimethylborolanes (33),⁷⁵⁴ and the borneol derivative 34,⁷⁵⁵ all of which add an allyl group to aldehydes, with good enantioselectivity. A recyclable 10-TMS-9-borabicyclo[3.3.2]decane has been used for asymmetric allyl and crotyl addition to aldehydes.⁷⁵⁶ Where the substrate possesses an aryl group or a triple bond, using a metal carbonyl complex of the substrate enhances enantioselectivity.⁷⁵⁷

Potassium alkynyltrifluoroborates (Reaction 12-28) react with aldehydes and a secondary amine, in an ionic liquid, to give a propargylic amine.⁷⁵⁸ Allylic trifluoroborates (Reaction 12-28) react with aldehydes to give the homoallylic alcohol. The Pd catalyzed reaction of aryl aldehydes with PhBF₃K gave a diaryl alcohol.⁷⁵⁹ A Ru catalyzed reaction of aryltrifluoroborates led to sterically hindered diaryl ketones.⁷⁶⁰ Aliphatic aldehydes react with this reagent, in the presence of BF₃·OEt₂, to give the homoallylic alcohol with allylic rearrangement and a preference for the syn diastereomer.⁷⁶¹ Aryl aldehydes react as well.⁷⁶²

16-26 Addition of Trialkylallylsilanes to Aldehydes and Ketones

O-Hydro-C-alkyl-addition

Allylic trialkyl, trialkoxy, and trihalosilanes add to aldehydes to give the homoallylic alcohols in the presence of a Lewis acid⁷⁶³ (including TaCl₅⁷⁶⁴ and YbCl₃⁷⁶⁵), fluoride ion,⁷⁶⁶ proazaphosphatranes,⁷⁶⁷ or a catalytic amount of iodine.⁷⁶⁸ A Ru catalyst has been used in conjunction with an arylsilane and an aldehyde.⁷⁶⁹ Allyl(trimethoxy)silane adds an allyl group to aldehydes using a CdF₂⁷⁷⁰ catalyst or a chiral AgF complex.⁷⁷¹ Allyltrichlorosilanes have also been used in addition reactions with aldehydes.⁷⁷² Hünig's base (iPr₂NEt) and a sulfoxide have also been used to facilitate the addition of an allyl group to an aldehyde from allyltrichlorosilane.⁷⁷³ The mechanism of this reaction has been examined.⁷⁷⁴

Allyltrichlorosilane reacts with aldehydes in the presence of certain additives to give the corresponding alcohol,⁷⁷⁵ and the reaction proceeds with good enantioselectivity by addition of a chiral additive.⁷⁷⁶ Other chiral additives have been used,⁷⁷⁷ as well as chiral catalysts,⁷⁷⁸ and chiral complexes of allyl silanes.⁷⁷⁹ Trimethoxyallylsilanes react with aldehydes in the presence of a Cu catalyst and a chiral ligand to give the chiral alcohol.⁷⁸⁰ Chiral allylic silyl derivatives add to aldehydes to give the chiral homoallylic alcohol.⁷⁸¹

Allylic silanes react with gem-diacetates in the presence of InCl₃ to give a homoallylic acetate⁷⁸² or with dimethyl acetals and TMSOTf in an ionic liquid to give the homoallylic methyl ether.⁷⁸³ Allylic alcohols can be treated with TMS–Cl and NaI, and then Bi to give an organometallic reagent that adds to aldehydes.⁷⁸⁴

16-27 Addition of Conjugated Alkenes to Aldehydes (the Baylis–Hillman Reaction)⁷⁸⁵

O-Hydro-C-alkenyl-addition

In the presence of a base⁷⁸⁶ (1,4-diazabicyclo[2.2.2]octane, DABCO) or trialkylphosphines, conjugated carbonyl compounds (ketones, esters, including lactones)⁷⁸⁷ thioesters,⁷⁸⁸ and amides⁷⁸⁹) add to aldehydes via the α carbon to give α-alkenyl-β-hydroxy esters or amides. This sequence is called the Baylis–Hillman reaction (or Morita–Baylis–Hillman),⁷⁹⁰ and a simple example is the formation of 35.⁷⁹¹ Mechanistic investigations of the reaction have been reported.⁷⁹² Methyl vinyl ketone gave other products in the Baylis–Hillman reaction, whereas conjugated esters do not.⁷⁹³ Methods that are catalytic in base have been developed for the Baylis–Hillman reaction.⁷⁹⁴ Both microwave irradiation⁷⁹⁵ and ultrasound⁷⁹⁶ have been used to induce the reaction.⁷⁹⁷ Baylis–Hillman reactions have been done in aqueous acidic media.⁷⁹⁸ There is a protein-catalyzed reaction.⁷⁹⁹ Under certain conditions, rate enhancements have been observed.⁸⁰⁰ The reaction has been done in ionic liquids⁸⁰¹ and PEG⁸⁰² or sulfolane,⁸⁰³ and without an organic solvent.⁸⁰⁴ Rate acceleration occurs with bis(arylthio)ureas in a DABCO promoted reaction.⁸⁰⁵ Transition metal compounds can facilitate the Baylis–Hillman reaction,⁸⁰⁶ and BF₃·OEt₂ has been used.⁸⁰⁷ A sila variation is known, involving the reaction of vinylsilanes and aldehydes.⁸⁰⁸ An aza-Balyis–Hillman reaction is discussed in Reaction 16-31.

An intramolecular version of the Baylis–Hillman reaction generated cyclopentenone derivatives from alkyne-aldehydes and a Rh catalyst.⁸⁰⁹ Other intramolecular cyclization reactions are known.⁸¹⁰ Cyclization of a conjugated ester using DABCO, where the “alcohol” group contained an aldehyde unit (an α-hydroxy aldehyde derivative) gave a lactone with a hydroxy unit at C-3 relative to the carbonyl and an α-vinylidene.⁸¹¹

Enantioselective Baylis–Hillman reactions⁸¹² are possible using a chiral auxiliary via an amide⁸¹³ or ester.⁸¹⁴ Organocatalysts can be used to give products with good enantioselectivity.⁸¹⁵ Sugars have been used as ester auxiliaries, and in reaction with aryl aldehydes and 20% DABCO gave the allylic alcohol with modest enantioselectivity.⁸¹⁶

Another variation is the Rauhut–Currier cyclization reaction,⁸¹⁷ which involves the reaction of a conjugated carbonyl with the allylic site of a second conjugated system. Intramolecular variations of this latter method are known.⁸¹⁸ With the boron trifluoride induced reaction between an aldehyde and a conjugated ketone, a saturated β-hydroxy ketone was formed with good antiselectivity.⁸¹⁹ The coupling of aldehydes with conjugated ketones used TiCl₄,⁸²⁰ dialkylaluminum halides,⁸²¹ and with (polymethyl)hydrosiloxane and a Cu catalyst.⁸²² Conjugated esters were coupled to aldehydes with DABCO and a La catalyst.⁸²³ Aldehydes are coupled to conjugated esters with a chiral quinuclidine catalyst and a Ti catalyst, as well as in the presence of tosylamine, the final product was the allylic N-tosylamine formed with modest enantioselectivity.⁸²⁴

Alkyl halides are coupled with conjugated carbonyls to give the alkylated derivative in what is known as Morita–Baylis–Hillman alkylation.⁸²⁵ α-Bromomethyl esters react with conjugated ketones and DABCO to give a coupling product, (34).⁸²⁶ A similar DBU induced reaction was reported using α-bromomethyl esters and conjugated nitro compounds.⁸²⁷

The reaction can be modified to give additional products, as with the reaction of o-hydroxybenzaldehyde and methyl vinyl ketone with DABCO, where the initial Baylis–Hillman product cyclized via conjugate addition of the phenolic oxygen to the conjugated ketone (Reaction 15-31).⁸²⁸ Aldehydes and conjugated esters can be coupled with a sulfonamide to give an allylic amine.⁸²⁹

See OS 2010, 87, 201

16-28 The Reformatsky Reaction

O-Hydro-C-α-ethoxycarbonylalkyl-addition

The Reformatsky reaction⁸³⁰ is very similar to Reaction 16-24. An aldehyde or ketone is treated with Zn and a halide; the halide is usually an α-halo ester or a vinylog (see Sec. 6.B) of an α-halo ester (e.g., RCHBrCH=CHCOOEt), although α-halo nitriles,⁸³¹ α-halo ketones,⁸³² and α-halo N,N-disubstituted amides have also been used. Especially high reactivity can be achieved with activated Zn,⁸³³ and with Zn and ultrasound.⁸³⁴ The reaction is catalytic in Zn in the presence of iodine and ultrasound.⁸³⁵ Metals other than Zn can be used, including In,⁸³⁶ Mn,⁸³⁷ low valent Ti,⁸³⁸ metal compounds of Ti,⁸³⁹ Sn,⁸⁴⁰ Sm⁸⁴¹ or Sc.⁸⁴² The use of additives (e.g., Ge⁸⁴³ or Me₂Zn)⁸⁴⁴ can lead to highly selective reactions.⁸⁴⁵ A combination of Zn and an α-bromo ester can be used in conjunction with BF₃·OEt₂, followed by reaction with dibenzoyl peroxide.⁸⁴⁶ The aldehyde or ketone can be aliphatic, aromatic, or heterocyclic, or contain various functional groups. Solvents used are generally ethers, including Et₂O, THF, and 1,4-dioxane, although the reaction can be done in water⁸⁴⁷ using dibenzoyl peroxide and MgClO₄.

With the use of chiral auxiliaries⁸⁴⁸ or chiral catalysts,⁸⁴⁹ good enantioselectivity⁸⁵⁰ can be achieved.

Dialkylzinc compounds are an alternative source of Zn in the Reformatsky reaction. The reaction of an α-bromo ester, an aldehyde, and diethylzinc in THF with a Rh catalyst, gave a β-hydroxy ester.⁸⁵¹

Formally, the reaction is somewhat analogous to the Grignard reaction (16-24), with EtO₂C–C–ZnBr (37) as an intermediate analogous to RMgX.⁸⁵² There is an intermediate derived from Zn and the ester, the structure of which has been shown to be 38, by X-ray crystallography of the solid intermediate prepared from t-BuOCOCH₂Br and Zn.⁸⁵³ As can be seen, it has some of the characteristics of 37.

After hydrolysis, the alcohol is the usual product, but sometimes (especially with aryl aldehydes) elimination follows directly and the product is an alkene. By the use of Bu₃P along with Zn, the alkene can be made the main product,⁸⁵⁴ making this an alternative to the Wittig reaction (16-44). The alkene is also the product when K₂CO₃/NaHCO₃ is used with 2% PEG–telluride.⁸⁵⁵ Since Grignard reagents cannot be formed from α-halo esters, the method is quite useful, but competing reactions sometimes lead to low yields. A similar reaction (called the Blaise reaction) has been carried out on nitriles⁸⁵⁶:

Carboxylic esters have also been used as substrates, but then, as might be expected (Sec. 16.A), the result is substitution and not addition:

The product in this case is the same as with the corresponding nitrile, though the pathways are different. The reaction is compatible with amine substituents, and α-(N,N-dibenzyl)amino aldehydes have been used to prepare β-hydroxy-γ-(N,N-dibenzylamino) esters with good antiselectivity.⁸⁵⁷

For an alternative approach involving ester enolates, see Reaction 16-36.

OS 3, 408; 4, 120, 444; 6, 598; IX, 275.

16-29 The Conversion of Carboxylic Acid Salts to Ketones with Organometallic Compounds

Alkyl-de-oxido-substitution

Good yields of ketones can often be obtained by treatment of the lithium salt of a carboxylic acid with an alkyllithium reagent, followed by hydrolysis.⁸⁵⁸ The carboxylate salt is formed by reaction of the carboxylic acid with 1 molar equivalent of R¹Li. The R′ group may be aryl or primary, secondary, or tertiary alkyl and R may be alkyl or aryl. The compounds MeLi and PhLi have been employed most often. Tertiary alcohols are side products. Lithium acetate can be used, but generally gives low yields. Using R(PrNH)Mg, carboxylic acid salts react to form ketones.⁸⁵⁹

A variation of this transformation reacts the acid with lithium naphthalenide in the presence of 1-chlorobutane. The product is the ketone.⁸⁶⁰ A related reaction treats the lithium carboxylate with lithium metal and the alkyl halide, with sonication, to give the ketone.⁸⁶¹ Phenylboronic acid (Reaction 12-28) reacts with aryl carboxylic acids in the presence of a Pd catalyst and disuccinoyl carbonate to give a diaryl ketone.⁸⁶²

OS V, 775.

16-30 The Addition of Organometallic Compounds to CO₂ and CS₂

C-Alkyl-O-halomagnesio-addition

Grignard reagents add to one C=O bond of CO₂ exactly as they do to an aldehyde or a ketone,⁸⁶³ but the product is the salt of a carboxylic acid. The reaction is usually performed by adding the Grignard reagent to dry ice. Many carboxylic acids have been prepared in this manner, and this constitutes an important way of increasing a carbon chain by one unit. Since labeled CO₂ is commercially available, this is a good method for the preparation of carboxylic acids labeled in the carboxyl group. Other organometallic compounds have also been used (RLi,⁸⁶⁴ RNa, RCaX, RBa,⁸⁶⁵ etc.), but much less often. The formation of the salt of a carboxylic acid after the addition of CO₂ to a reaction mixture is regarded as a positive test for the presence of a carbanion or of a reactive organometallic intermediate in that reaction mixture (see also, Reaction 16-42).

Aryl organoboronates react with CO₂, in the presence of a Ru catalyst, to give the corresponding aryl carboxylic acid.⁸⁶⁶ Aryl and alkenyl boronic esters are carboxylated in the presence of a Cu catalyst.⁸⁶⁷ Vinyl halides react with CO and water, with a Pd catalyst in an ionic liquid, to give the conjugated carboxylic acid.⁸⁶⁸ Organozinc compounds are carboxylated with CO₂ and a Ni catalyst.⁸⁶⁹ Metal-free carboxylation of organozinc reagents is also known.⁸⁷⁰In the presence of CO₂ and an organocatalyst, aromatic aldehydes are converted to the corresponding carboxylic acid.⁸⁷¹ Direct carboxylation of aryl bromides is possible using CO₂ and a Pd catalyst.⁸⁷² Allenes are converted to β,γ-unsaturated acids with CO₂ in the presence of Et₂Zn.⁸⁷³

When chiral additives (e.g., (−)-sparteine) have added to the initial reaction with the organolithium reagent, quenching with CO₂ produces carboxylic acids with good asymmetric induction.⁸⁷⁴

In a closely related reaction, Grignard reagents add to CS₂ to give salts of dithiocarboxylic acids.⁸⁷⁵ These salts can be trapped with amines to form thioamides.⁸⁷⁶ Two other reactions are worthy of note. (1) Lithium dialkylcopper reagents react with dithiocarboxylic esters to give tertiary thiols.⁸⁷⁷ (2) Thiono lactones can be converted to cyclic ethers,⁸⁷⁸ for example:

This is a valuable procedure because medium and large ring ethers are not easily made, while the corresponding thiono lactones can be prepared from the readily available lactones (see, e.g., Reaction 16-63) by reaction 16-11.

A terminal alkyne can be converted to the anion under electrolytic conditions, in the presence of CO₂, to give propargylic acids (R–CC–CO₂H).⁸⁷⁹

OS I, 361, 524; II, 425; III, 413, 553, 555; V, 890, 1043; VI, 845; IX, 317.

16-31 The Addition of Organometallic Compounds to C=N Compounds

N-Hydro-C-alkyl-addition

Aldimines can be converted to secondary amines by treatment with Grignard reagents.⁸⁸⁰ Ketimines generally react with Grignard reagents to give reduction instead of addition. However, organolithium compounds give the normal addition product with both aldimines and ketimines.⁸⁸¹ The solvent and the aggregation state of the organolithium play a role in the addition, however.⁸⁸² For the addition of an organometallic compound to an imine to give a primary amine, R′ in RCH=NR′ would have to be H, and such compounds are seldom stable. However, the conversion has been done for R = aryl, by the use of the masked reagents (ArCH=N)₂SO₂ [prepared from an aldehyde, RCHO, and sulfamide, (NH₂)₂SO₂]. Addition of R²MgX or R²Li to these compounds gives ArCHR²NH₂ after hydrolysis.⁸⁸³ An intramolecular version of the addition of organolithium reagents is known that gave 2-phenylpyrrolidine.⁸⁸⁴ Grignard reagents add to imines in the presence of various transition metal catalysts, including Sc(OTf)₃⁸⁸⁵ or Cp₂ZrCl₂.⁸⁸⁶ Alkynes add to imines yielding propargylic amines.⁸⁸⁷

When chiral additives are used in conjunction with the organolithium reagent, chiral amines are produced⁸⁸⁸ with good asymmetric induction.⁸⁸⁹ Chiral auxiliaries have been used in addition reactions of organometallic compounds to imines,⁸⁹⁰ and to oxime derivatives.⁸⁹¹ Transition metal asymmetric alkylation of imines uses metal compounds of Cu.⁸⁹² Chiral catalysts lead to enantioselective addition of alkynes to imines to give the amine.⁸⁹³ Chiral N-sulfinylimines reaction with lithium silanes give the α-silylsulfinylamine.⁸⁹⁴

Zinc metal reacts with allylic bromides to form an allylic zinc complex, which reacts with imines to give the homoallylic amine.⁸⁹⁵ This reaction is catalyzed by TMSCl.⁸⁹⁶ Dialkylzinc reagents add to functionalized imines to give the functionalized amine, often with transition metal catalysts (e.g., Ni compounds).⁸⁹⁷ Dialkylzinc reagents add to N-tosyl imines to give the alkylated tosylamine.⁸⁹⁸ In the presence of a chiral ligand, the metal-catalyzed reaction proceeds with good enantioselectivity.⁸⁹⁹ With a Cu catalyst and a chiral ligand the product is formed with good enantioselectivity.⁹⁰⁰ The reaction of imines (e.g., ArN=CHCO₂Et), where R = a chiral benzylic substituent and ZnBr₂, followed by R′ZnBr leads to a chiral α-amino ester.⁹⁰¹ Terminal alkynes add to imines using ZnCl₂ and TMSCl, and with a chiral ligand attached to nitrogen the reaction proceeds with some enantioselectivity.⁹⁰² Dimethylzinc has been used to mediate the addition of terminal alkynes to N-tosylimines.⁹⁰³

Other organometallic compounds add to aldimines,⁹⁰⁴ including Sn,⁹⁰⁵ Sm,⁹⁰⁶ Ge,⁹⁰⁷ Zr,⁹⁰⁸ Ga metal with ultrasound,⁹⁰⁹ Yb with Me₃SiCl,⁹¹⁰and In.⁹¹¹ Catalytic amounts of the metal compound can be used with an allylic stannane.⁹¹² Catalytic enantioselective addition reactions with these organometallics are well known,⁹¹³ including reactions in an ionic liquid.⁹¹⁴ Aryltrialkylstannanes add the aryl group to N-tosyl imines using a Rh catalyst and sonication.⁹¹⁵ Allylic halides react with imines in the presence of In metal⁹¹⁶ or InCl₃⁹¹⁷ to give the homoallylic amine, and with N-sulfonyl imines to give the homoallylic sulfonamide.⁹¹⁸ In this latter reaction, antiselectivity was observed when the reaction was done in water, and syn selectivity when done in aq THF.⁹¹⁹ Aryl iodides add to N-aryl imines in the presence of a Rh catalyst.⁹²⁰ Arylation of N-tosyl ketimines proceeds with good enantioselectivity using a Rh catalyst.⁹²¹ Propargylic halides add to imines in the presence of In metal, in aq THF.⁹²²

Terminal alkynes react with aryl aldehydes and aryl amines to give propargylic amine without a catalyst.⁹²³ Alternatively, with an Ir⁹²⁴ or a Cu catalyst⁹²⁵ they also lead to a propargylic amine.⁹²⁶ Terminal alkynes add to N-substituted imines to give a propargylic amine with good enantioselectivity using a chiral Cu complex.⁹²⁷ Lithium alkyne anions add to chiral N-sulfinylimines, in the presence of Me₃Al, to give the chiral propargyl compound.⁹²⁸Diynes add to N-sulfinyl imines in the presence of a chiral Rh catalyst to give the corresponding allylic sulfinylamine.⁹²⁹ Alkenes add to N-sulfonyl imines in the presence of a chiral Rh catalyst to give the alkylated sulfonamide.⁹³⁰

Triethylaluminum adds an ethyl group to an imine in the presence of a Eu catalyst. Reaction with PhSnMe₃ and N-tosylimines with a Rh catalyst leads to addition of a phenyl group to the carbon of the C=N bond.⁹³¹ Other N-sulfonyl imines react similarly to give the corresponding sulfonamide, and in the presence of a chiral ligand the reaction proceeds to good enantioselectivity.⁹³²N-Tosyl imines also react with dialkylzinc reagents, giving the sulfonamide with modest enantioselectivity.⁹³³N-Sulfinyl imines [R₂CH=NS(=O)R′]⁹³⁴ react with Grignard reagents at carbon to give the corresponding N-sulfinylamine.⁹³⁵ Furan derivatives add via C-2 with good enantioselectivity using a chiral phosphoric acid catalyst.⁹³⁶ Alkenes add to N-tosyl imines with a Yb catalyst⁹³⁷ and allenes add to N-carbamoyl imines in the presence of a V catalyst.⁹³⁸N-Carbamoyl imines, formed in situ, react with allylic silanes in the presence of an iodine catalyst.⁹³⁹N-Carbamoyl imines add acetonitrile (via carbon) using DBU and a Ru catalyst.⁹⁴⁰

Arylboronates (Reaction 12-28) add to N-sulfonyl imines in the presence of a Rh catalyst to give the corresponding sulfonamide.⁹⁴¹ Chiral Cu complexes have also been used for effective allylation of ketimines.⁹⁴² Aryl boronic acids (Reaction 12-28) add the aryl group to N-tosyl imines using a Rh⁹⁴³ or Pd⁹⁴⁴ catalyst. Arylboronic acids react similarly, and in the presence of a chiral Rh⁹⁴⁵ or Ir⁹⁴⁶ catalyst give a chiral sulfinamide or sulfonamide. Allylic boronates also add to aldehydes, and subsequent treatment with ammonia gives the homoallylic amine.⁹⁴⁷ Vinyl boronates add to nitrones in the presence of Me₂Zn, transferring the vinyl group to the C=N unit.⁹⁴⁸ Potassium allyltrifluoroborates react with N-tosyllimines in the presence of a Pd catalyst.⁹⁴⁹

Allylic silanes (e.g., allyltrimethylsilane) add to N-substituted imines in the presence of a Pd catalyst to give the homoallylic amine.⁹⁵⁰ Similar results are obtained when the allylic silane and imine are treated with a catalytic amount of tetrabutylammonium fluoride.⁹⁵¹ Allylic trichlorosilanes add to hydrazones to give homoallylic hydrazine derivatives with excellent antiselectivity⁹⁵² and with good enantioselectivity using a chiral ligand.⁹⁵³ Chiral allylic silane derivatives have been developed, and add to hydrazones with good enantioselectivity.⁹⁵⁴

There is an aza-Balyis–Hillman reaction that converts imines and conjugated carbonyl derivatives to the α-amino conjugated derivative.⁹⁵⁵N-Tosyl imines can be used in place of aldehydes, and the reaction of the imine, a conjugated ester and DABCO gave the allylic N-tosylimine.⁹⁵⁶ A “double Baylis–Hillman” reaction has also been reported using N-tosylimines and conjugated ketones.⁹⁵⁷ The use of chiral catalysts leads enantioselective product formation.⁹⁵⁸ An enantioselective aza-Baylis–Hillman reaction⁹⁵⁹ was reported using as chiral reaction medium.⁹⁶⁰ Aldehydes add via the α carbon using proline, to give β-amino aldehydes with good selectivity to give chiral β-amino aldehydes.⁹⁶¹

Silyl enol ethers add to hydrazones in the presence of ZnF₂ and a chiral ligand to give chiral β-hydrazino ketones.⁹⁶² Similar addition to imine derivatives was accomplished using ketene silyl acetals and Amberlyst-15.⁹⁶³Alternatively, an imine is reacted first with Zn(OTf)₂ and then with a ketene silyl acetal.⁹⁶⁴

Nitro compounds add to N-carbamoyl imines with a chiral diamine catalyst with some enantioselectivity.⁹⁶⁵ Nitro compounds add via carbon using a Cu catalyst, and with good enantioselectivity when a chiral ligand is used.⁹⁶⁶The conjugate bases of nitro compounds (formed by treatment of the nitro compound with BuLi) react with Grignard reagents in the presence of ClCH=NMe₂⁺ Cl⁻ to give oximes: RCH=N(O)OLi + R′MgX → RR′C=NOH.⁹⁶⁷

Iminium salts⁹⁶⁸ give tertiary amines directly, via 1,2-addition to the C=N unit. Chloroiminium salts [ClCH=NR₂'Cl⁻, generated in situ from an amide (HCONR₂') and phosgene (COCl₂)] react with 2 molar equivalents of a Grignard reagent RMgX, one adding to the C=N and the other replacing the Cl, to give tertiary amines R₂CHNR₂'.⁹⁶⁹

Many other C=N systems (phenylhydrazones, oxime ethers, etc.) give 1,2-addition when treated with Grignard reagents, while some gave reductions and others gave miscellaneous reactions. Organocerium reagents add to hydrazones.⁹⁷⁰ Indium metal promotes the addition of alkyl iodides to hydrazones.⁹⁷¹ A hydrazone can be formed in situ by reacting an aldehyde with a hydrazine derivative. In the presence of tetrallyltin and a Sc catalysts, homoallylic hydrazine derivatives are formed.⁹⁷² Hydrazone derivatives react with iodoalkenes in the presence of InCl₃ and Mn₂(CO)₁₀ under photochemical conditions to give the hydrazine derivative.⁹⁷³ Ketene dithioacetals add to hydrazones using a chiral Zr catalyst to give a pyrazolidine.⁹⁷⁴ Oximes can be converted to hydroxylamines (39) by treatment with 2 molar equivalents of an alkyllithium reagent, followed by methanol.⁹⁷⁵ Oxime ethers add an allyl group upon reaction with allyl bromide and In metal in water.⁹⁷⁶ Nitrones [R₂C=N⁺(R′)–O⁻] react with allylic bromides and Sm to give homoallylic oximes,⁹⁷⁷ and with terminal alkynes and a Zn catalyst to give propargylic oximes.⁹⁷⁸ Grignard reagents also add to nitrones.⁹⁷⁹ Nitrones react with CH₂=CHCH₂InBr in aq DMF to give the homoallylic oxime⁹⁸⁰ and silyl ketene acetals add in the presence of a chiral Ti catalyst to good enantioselectivity.⁹⁸¹

Allylic alcohols add to imines in the presence of a Pd catalyst to give the homoallylic amine.⁹⁸²

OS IV, 605; VI, 64. Also see, OS III, 329.

16-32 Addition of Carbenes and Diazoalkanes to C=N Compounds

In the presence of metal catalysts [e.g., Yb(OTf)₃], diazoalkanes add to imines to generate aziridines.⁹⁸³ The reaction is somewhat selective for the cis-diastereomer. The use of chiral additives in this reaction leads to aziridines enantioselectively.⁹⁸⁴ Imines can be formed by the reaction of an aldehyde and an amine, and subsequent treatment with Me₃SiI and butyllithium gives an aziridine.⁹⁸⁵N-Tosyl imines react with diazoalkenes to form N-tosyl aziridines, with good cis selectivity⁹⁸⁶ and modest enantioselectivity in the presence of a chiral Cu catalyst,⁹⁸⁷ but give excellent enantioselectivity with a chiral Rh catalyst⁹⁸⁸ or with the use of an organocatalyst.⁹⁸⁹ Note that N-tosyl aziridines are formed by the reaction of an alkene with PhI=NTs and a Cu catalyst.⁹⁹⁰N-Acylimines react with diazoesters via C–H insertion using a Pt catalyst.⁹⁹¹ The reaction of alkenes with diazo compounds is discussed in Reaction 15-53.

16-33 The Addition of Grignard Reagents to Nitriles and Isocyanates

Alkyl,oxo-de-nitrilo-tersubstitution (Overall transformation)

N-Hydro-C-alkyl-addition

Ketones can be prepared by addition of Grignard reagents to nitriles, followed by hydrolysis of the initially formed imine anion. Many ketones have been made in this manner, though when both R groups are alkyl, yields are not high.⁹⁹² Yields can be improved by the use of Cu(I) salts⁹⁹³ or by using benzene containing 1 equiv of ether as the solvent, rather than ether alone.⁹⁹⁴ In general, the ketimine salt does not react with Grignard reagents: Hence, tertiary alcohols or tertiary alkyl amines are not often side products.⁹⁹⁵ By careful hydrolysis of the salt, it is sometimes possible to isolate ketimines (RR′C=NH),⁹⁹⁶ especially when R and R′ = aryl. The addition of Grignard reagents to the CN group is normally slower than to the C=O group, and cyano group containing aldehydes add the Grignard reagent without disturbing the CN group.⁹⁹⁷ Organolithium reagents add to nitriles, mediated by LiBr, to form N-acetyl enamines.⁹⁹⁸ Other metal compounds have been used, including Sm with allylic halides⁹⁹⁹ and organocerium compounds (e.g., MeCeCl₂).¹⁰⁰⁰ Allylic halides react with an excess of Zn metal in the presence of 40% AlCl₃ and in the presence of a nitrile, to give homoallylic ketones after hydrolysis.¹⁰⁰¹

Addition of Grignard reagents¹⁰⁰² or organolithium reagents¹⁰⁰³ to ω-halo nitriles leads to 2-substituted cyclic imines.

The Blaise reaction is the reaction of the organozinc reagent derived from a α-bromoester with Zn metal and a nitrile to give the corresponding β-ketoester.¹⁰⁰⁴

The following mechanism has been proposed for the reaction of the methyl Grignard reagent with benzonitrile¹⁰⁰⁵:

Arylboronic acids add to nitriles in the presence of a Pd catalyst.¹⁰⁰⁶ Aryl and alkenylboronic acids add to isocyanates in the presence of a Pd¹⁰⁰⁷ or Rh¹⁰⁰⁸ catalyst.

Arenes add to nitriles in the presence of a Pd catalyst in DMSO/trifluoroacetic acid to give a diaryl ketone.¹⁰⁰⁹

The addition of Grignard reagents to isocyanates gives, after hydrolysis, N-substituted amides.¹⁰¹⁰ This is a very good reaction and can be used to prepare derivatives of alkyl and aryl halides. The reaction has also been performed with alkyllithium compounds.¹⁰¹¹ Isothiocyanates give N-substituted thioamides. Other organometallic compounds add to isocyanates. Vinyltin reagents lead to conjugate amides.¹⁰¹²

Note that terminal alkynes add to the carbon of an isonitrile in the presence of a uranium complex, giving a propargylic imine.¹⁰¹³

OSCV III, 26, 562; V, 120, 520.

G. Carbon Attack by Active Hydrogen Compounds

Reactions 16-34–16-50 are base-catalyzed condensations (although some of them are also catalyzed to acids).¹⁰¹⁴ In Reactions 16-34–16-44, a base removes a C–H proton to give a carbanion, which then adds to a C=O. The oxygen acquires a proton, and the resulting alcohol may or may not be dehydrated, depending on whether an α hydrogen is present and on whether the new double bond would be in conjugation with double bonds already present:

The reactions differ in the nature of the active hydrogen component and the carbonyl component. Table 16.2 illustrates the differences. Reaction 16-50 is an analogous reaction involving addition to CN.

Table 16.2 Base-Catalyzed Condensations Showing the Active-Hydrogen Components and the Carbonyl Compounds.

16-34 The Aldol Condensation¹⁰¹⁵

O-Hydro-C-(α-acylalkyl)-addition; α-Acylalkylidine-de-oxo-bisubstitution

In the aldol reaction or aldol condensation¹⁰¹⁶ the α carbon of one aldehyde or ketone molecule adds to the carbonyl carbon of another.¹⁰¹⁷ Although acid-catalyzed aldol reactions are known,¹⁰¹⁸ the most common form of the reaction uses a base. There is evidence that an SET mechanism can intervene when the substrate is an aromatic ketone.¹⁰¹⁹ Although hydroxide was commonly used in early versions of this reaction, stronger bases [e.g., alkoxides (RO⁻) or amides (R₂N⁻)] are also common. Amine bases have been used to catalyze the aldol condensation.¹⁰²⁰ Hydroxide ion is not a strong enough base to convert substantially all of an aldehyde or ketone molecule to the corresponding enolate ion; that is, the equilibrium lies well to the left, for both aldehydes

and ketones. Nevertheless, enough enolate ion is present for the reaction to proceed:

This equilibrium lies further to the right with alkoxide relative to hydroxide, but the equilibrium still lies predominantly to the left. With amide bases, and particularly with aprotic solvents, the equilibrium usually lies much more to the right when compared with alkoxides or hydroxide. Protic solvents, (e.g., water or alcohol) are acidic enough to react with the enolate anion and shift the equilibrium to the left. As noted, in an aprotic solvent (e.g., ether or THF), with a strong amide base (e.g., LDA, sec. Sec. 8.F, category 7), the equilibrium lies more to the right.¹⁰²¹ A variety of amide bases can be used to deprotonate the ketone or aldehyde, and in the case of an unsymmetrical ketone removal of the more acidic proton leads to the kinetic enolate anion.¹⁰²² Note that a polymer-bound amide base has been used¹⁰²³ and solid-phase chiral lithium amides are known.¹⁰²⁴ A polymer-supported phosphoramide has been used as a catalyst for the aldol condensation.¹⁰²⁵

The product of an aldol condensation is a β-hydroxy aldehyde (called an aldol) or a β-hydroxy ketone, which in some cases is dehydrated during the course of the reaction. An aldol is readily isolated unless the substrate is an aromatic aldehyde or ketone when the reaction is done in aprotic solvents with a mild workup procedure. The aldol reaction has been done in ionic liquids.¹⁰²⁶ Even if the dehydration is not spontaneous, it can usually be done easily, since the new double bond is in conjugation with the C=O bond; so that this is a method of preparing α,β-unsaturated aldehydes and ketones, as well as β-hydroxy aldehydes and ketones. One-pot procedures have been reported to give the conjugated product.¹⁰²⁷ The entire reaction is an equilibrium (including the dehydration step), and α,β-unsaturated and β-hydroxy aldehydes and ketones can be cleaved by treatment with ⁻OH (the retrograde aldol reaction). The retro-aldol condensation has been exploited for crossed-aldol reactions.¹⁰²⁸ A vinylogous (Sec. 6.B) aldol reaction is known¹⁰²⁹ as is a ‘double’ aldol.¹⁰³⁰ Enzyme-mediated aldol reactions have been reported using two aldehydes, including formaldehyde.¹⁰³¹

Under the principle of vinylogy (Sec. 6.B), the active hydrogen can be one in the γ position of an α,β-unsaturated carbonyl compound:

The scope of the aldol reaction may be discussed under five headings:

1. Reaction between Two Molecules of the Same Aldehyde. Homo-coupling. Hydroxide or alkoxide bases are used in protic solvents,¹⁰³² and the reaction is quite feasible. Nowadays, the use of dialkylamide bases in aprotic solvents (e.g., ether or THF) is more common. Many aldehydes have been converted to aldols and/or their dehydration products in this manner. The most effective catalysts are basic ion-exchange resins. Of course, the aldehyde must possess an α hydrogen.

2. Reaction between Two Molecules of the Same Ketone. Homo-coupling. With hydroxide or alkoxide bases in protic solvents the equilibrium lies well to the left,¹⁰³³ and the reaction is feasible only if the equilibrium can be shifted. This can often be done by allowing the reaction to proceed in a Soxhlet extractor (e.g., See OS I, 199). As with aldehydes, the use of dialkylamide bases (e.g., LDA or lithium hexamethyldisilazide, Sec. 8.F., category 7) in aprotic solvents (e.g., ether or THF) are more common. Unsymmetrical ketones condense on the side that has the most hydrogens with dialkylamide bases in aprotic solvents, but on the side with the fewest hydrogens with alkoxide bases in alcohol solvents.

3. Reaction between Two Different Aldehydes. Cross-coupling. In protic solvents with an alkoxide base, this will produce a mixture of four products (eight, if the alkenes are counted). However, if one aldehyde does not have an α hydrogen, only two aldols are possible, and in many cases the crossed product is the main one. The crossed-aldol reaction is often called the Claisen–Schmidt reaction.¹⁰³⁴ The crossed-aldol reaction is readily accomplished using amide bases in aprotic solvent. The first aldehyde is treated with LDA in THF at –78 °C, for example, to form the enolate anion. Subsequent treatment with a second aldehyde leads to the mixed-aldol product. The crossed aldol of two aldehydes has been done using potassium tert-butoxide and Ti(OBu)₄.¹⁰³⁵

4. Reaction between Two Different Ketones. Cross-coupling. This is seldom attempted with hydroxide or alkoxide bases in protic solvents since similar considerations apply to those discussed for aldehydes. This reaction is commonly done with amide bases in aprotic solvents, but with somewhat more difficulty than with aldehydes.

5. Reaction between an Aldehyde and a Ketone. This is usually feasible with hydroxide or alkoxides bases in protic solvents when the aldehyde has no α hydrogen, since there is no competition from ketone condensing with itself.¹⁰³⁶ This is also called the Claisen–Schmidt reaction. Even when the aldehyde has an α hydrogen, it is generally the α carbon of the ketone that adds to the carbonyl of the aldehyde, not the other way around. Mixtures are usually produced, however. If the ketone or the aldehyde is treated with an amide base in aprotic solvents, a second aldehyde or ketone can be added to give the aldolate with high regioselectivity. The reaction can be also made regioselective by preparing an enol derivative of the ketone separately¹⁰³⁷ and then adding this to the aldehyde (or ketone). Other types of preformed derivatives that react with aldehydes and ketones are enamines (with a Lewis acid catalyst),¹⁰³⁸ and enol borinates (R′CH=CR²–OBR₂),¹⁰³⁹ which can be synthesized by Reaction 15-27 or directly from an aldehyde or ketone¹⁰⁴⁰. Preformed metallic enolates are also used. For example, lithium enolates¹⁰⁴¹(prepared by Reaction 12-23) react with the substrate in the presence of ZnCl₂.¹⁰⁴² In this case, the aldol product is stabilized by chelation of its two oxygen atoms with the zinc ion.¹⁰⁴³ Other metallic enolates can be used for aldol reactions, either preformed or generated in situ with a catalytic amount of a metal compound. Compounds used for this purpose include metal complexes of Mg,¹⁰⁴⁴ Ti,¹⁰⁴⁵ Zr,¹⁰⁴⁶ Pd,¹⁰⁴⁷ In,¹⁰⁴⁸ Sn,¹⁰⁴⁹ La,¹⁰⁵⁰ and Sm,¹⁰⁵¹all of which give products with moderate-to-excellent diastereoselectivity¹⁰⁵² and regioselectivity. α-Alkoxy ketones react with lithium enolates particularly fast.¹⁰⁵³ A bis(aldol) condensation has been reported with epoxy ketones and aldehydes using SmI₂.¹⁰⁵⁴

There are discussions that relate to the transition state of the aldol condensation. There is experimental evidence for chair-like transition states in the aldol reactions of methyl ketone lithium enolate anions.¹⁰⁵⁵A computational study gave the gas-phase activation energies for lithium enolate anions in an aldol-type reaction.¹⁰⁵⁶ There is a computational study of the mechanism for an aldol reaction in pure water.¹⁰⁵⁷

The reactions with pre-formed enol derivatives provide a way to control the stereoselectivity of the aldol reaction.¹⁰⁵⁸ As with the Michael reaction (15-24), the aldol reaction creates two new stereogenic centers. In the most general case, there are four stereoisomers of the aldol product (two racemic diastereomers), which can be represented as the syn and anti diastereomers shown. The reaction may be diastereoselective, however, if one is preferred over the other.

Among the preformed enol, derivatives used for diastereoselective aldol condensations have been enolates of Li,¹⁰⁵⁹ Mg, Ti,¹⁰⁶⁰ and Sn,¹⁰⁶¹ silyl enol ethers,¹⁰⁶² enol borinates,¹⁰⁶³ and enol borates [R′CH=CR²OB(OR)₂].¹⁰⁶⁴ The nucleophilicity of silyl enol ethers has been examined¹⁰⁶⁵ and reactions of these compounds are discussed in Reaction 16-35.

Base-induced formation of the enolate anion generally leads to a mixture of (E)- and (Z)-isomers, and dialkyl amide bases are used in most cases. The (E/Z) stereoselectivity depends on the structure of the lithium dialkylamide base, with the highest (E/Z) ratios obtained with LiTMP–butyllithium mixed aggregates in THF.¹⁰⁶⁶ The use of LiHMDS resulted in a reversal of the (E/Z) selectivity. In general, metallic (Z)-enolates give the syn (or erythro) pair, and this reaction is highly useful for the diastereoselective synthesis of these products.¹⁰⁶⁷ The (E) isomers generally react nonstereoselectively. However, anti-stereoselectivity has been achieved in a number of cases, with Ti enolates,¹⁰⁶⁸ with Mg enolates,¹⁰⁶⁹ with certain enol borinates,¹⁰⁷⁰ and with lithium enolates at –78 °C.¹⁰⁷¹ Enolization accounts for syn–anti isomerization of aldols.¹⁰⁷² In another variation, a β-keto Weinreb amide (see Reaction 16-82) reacted with TiCl₄ and Hünig's base (iPr₂NEt) and then an aldehyde to give the β-hydroxy ketone.¹⁰⁷³

These reactions are enantioselective¹⁰⁷⁴ (in which case only one of the four isomers predominates)¹⁰⁷⁵ by using¹⁰⁷⁶ chiral enol derivatives,¹⁰⁷⁷ chiral aldehydes or ketones,¹⁰⁷⁸ or both.¹⁰⁷⁹ Chiral bases¹⁰⁸⁰ can be used (e.g., proline),¹⁰⁸¹ proline derivatives,¹⁰⁸² or chiral additives used in conjunction with an organobase.¹⁰⁸³ Indeed, chiral organocatalysts are increasingly important,¹⁰⁸⁴ including those that can be used in aqueous media.¹⁰⁸⁵ Chiral auxiliaries¹⁰⁸⁶ have been developed that can be used in conjunction with the aldol condensation, as well as chiral transition metal complexes¹⁰⁸⁷ and chiral ligands¹⁰⁸⁸ in catalytic reactions. The enantioselective condensation of methyl vinyl ketone and an aldehyde used a chiral Zn catalyst.¹⁰⁸⁹ Structural variations in the aldehyde or ketone are compatible with many enantioselective condensation reactions. An α-hydroxy ketone was condensed with an aldehyde using a chiral Zn catalyst to give the aldol (an α,β-dihydroxy ketone) with good syn selectivity and good enantioselectivity.¹⁰⁹⁰ Chiral vinylogous aldol reactions (Sec. 6.B) have been reported.¹⁰⁹¹ Formation of the magnesium enolate anion of a chiral amide, adds to aldehydes to give the alcohol enantioselectively.¹⁰⁹² Diamine protonic acids have been used for catalytic asymmetric aldol reactions.¹⁰⁹³

Silyl enol ethers react with aldehydes in the presence of chiral boranes¹⁰⁹⁴ or other additives¹⁰⁹⁵ to give aldols with good asymmetric induction (see the Mukaiyama aldol Reaction in 16-35). Chiral boron enolates have been used.¹⁰⁹⁶ Since both new stereogenic centers are formed enantioselectively, this kind of process is called double asymmetric synthesis.¹⁰⁹⁷ Where both the enolate derivative and substrate were achiral, carrying out the reaction in the presence of an optically active boron compound¹⁰⁹⁸ or a diamine coordinated with a Sn compound¹⁰⁹⁹ gave the aldol product with excellent enantioselectivity for one stereoisomer. Boron triflate (R₂BOTf) derivatives have been used for the condensation of ketals and ketone to give β-alkoxy ketones.¹¹⁰⁰

It is possible to make the α carbon of the aldehyde add to the carbonyl carbon of the ketone, by using an imine instead of an aldehyde, and LiN(iPr)₂ as the base to form the α-lithio imine (40).¹¹⁰¹ This is known as a directed aldol reaction. Similar reactions have been performed with α-lithiated dimethylhydrazones of aldehydes or ketones¹¹⁰² and with α-lithiated aldoximes.¹¹⁰³

The aldol reaction can also be performed with acid catalysts, as mentioned above, in which case dehydration usually follows. Here there is initial protonation of the carbonyl group, which attacks the α carbon of the enol form of the other molecule¹¹⁰⁴:

With respect to the enol, this mechanism is similar to that of halogenation (Reaction 12-4). A side reaction that is sometimes troublesome is further condensation, since the product of an aldol reaction is still an aldehyde or ketone. The aldol condensation of aldehydes has also been done using a mixture of pyrrolidine and benzoic acid.¹¹⁰⁵

The intramolecular aldol condensation is well known, and aldol reactions are often used to close five- and six-membered rings. Because of the favorable entropy (Sec. 6.D), such ring closures generally take place with ease¹¹⁰⁶when using hydroxide or alkoxide bases in protic solvents. In aprotic solvents with amide bases, formation of the enolate anion occurs by deprotonation of the more acidic site, followed by cyclization to the second carbonyl. The acid-catalyzed intramolecular aldol condensation is known, and the mechanism has been studied.¹¹⁰⁷ Stereoselective proline-catalyzed intramolecular aldol reactions give the cyclized product with good enantioselectivity.¹¹⁰⁸ Chiral ligands, in conjugation with transition metal compounds of Cu¹¹⁰⁹, lead to asymmetric intramolecular aldol condensation reactions. An asymmetric intramolecular aldol reaction was catalyzed by a chiral amine.¹¹¹⁰ The regioselectivity of an intramolecular aldol condensation of unsaturated 1,5-diketones is strongly influenced by the presence or absence of a trialkylphosphine.¹¹¹¹

An important extension of the intramolecular aldol condensation is the Robinson annulation reaction,¹¹¹² which has often been used in the synthesis of steroids and terpenes. In original versions of this reaction, a cyclic ketone is converted to another cyclic ketone under equilibrium conditions using hydroxide or alkoxide bases in a protic solvent, forming one additional six-membered ring containing a double bond. The reaction can be done in a stepwise manner using amide bases in aprotic solvents. In the reaction with hydroxide or alkoxide bases in alcohol or water solvents, the substrate is treated with methyl vinyl ketone (or a simple derivative of methyl vinyl ketone) and a base.¹¹¹³ The enolate ion of the substrate adds to the methyl vinyl ketone in a Michael reaction (15-24) to give a diketone that undergoes or is made to undergo an internal aldol reaction and subsequent dehydration to give the product.¹¹¹⁴ The Robinson annulation can be combined with alkylation.¹¹¹⁵ Enantioselective Robinson annulation techniques have been developed, including a proline-catalyzed reaction.¹¹¹⁶ The Robinson annulation has been done in ionic liquids¹¹¹⁷ and a solvent-free version of the reaction is known.¹¹¹⁸

Because methyl vinyl ketone has a tendency to polymerize, surrogates are often used instead (i.e., compounds that will give methyl vinyl ketone when treated with a base). One common example, MeCOCH₂CH₂NEt₂Me⁺ I⁻ (see Reaction 17-9), is easily prepared by quaternization of MeCOCH₂CH₂NEt₂, which itself is prepared by a Mannich Reaction (16-19) involving acetone, formaldehyde, and diethylamine. α-Silylated vinyl ketones [RCOC(SiMe₃)=CH₂] have also been used successfully in annulation reactions¹¹¹⁹ because the SiMe₃ group is easily removed. When the ring closure of a 1,5-diketone is catalyzed by the amino acid (S)-proline, the product is optically active with high ee.¹¹²⁰Stryker's reagent,¹¹²¹[(Ph₃P)CuH]₆, has been used for an intramolecular addition where a ketone enolate anion adds to a conjugated ketone, giving cyclic alcohol with a pendant ketone unit.¹¹²²

OS I, 77, 78, 81, 199, 283, 341; II, 167, 214; III, 317, 353, 367, 747, 806, 829; V, 486, 869; VI, 496, 666, 692, 781, 901; VII, 185, 190, 332, 363, 368, 473; VIII, 87, 208, 241, 323, 339, 620; IX, 432, 610; X, 339.

16-35 Mukaiyama Aldol and Related Reactions¹¹²³

O-Hydro-C-(α-acylalkyl)-addition

An important variation of the aldol condensation involves treatment of an aldehyde or ketone with a silyl ketene acetal [R₂C=C(OSiMe₃)OR′]¹¹²⁴ or a silyl enol ether, in the presence of TiCl₄¹¹²⁵, to give 41. This variation is known as the Mukaiyama aldol reaction, or simply the Mukaiyama reaction. The silyl ketene acetal can be considered a pre-formed enolate that gives aldol product with TiCl₄ in aqueous solution, or with no

catalyst at all.¹¹²⁶ A combination of TiCl₄ and N-tosyl imine has also been used to facilitate the Mukaiyama aldol reaction.¹¹²⁷ Reaction at the carbonyl of saturated carbonyl compounds is significantly faster than 1,2-addition to unsaturated carbonyl compounds.¹¹²⁸ The mechanism of this reaction has been explored.¹¹²⁹ Other catalysts have been used for this reaction, including InCl₃,¹¹³⁰ SmI₂,¹¹³¹ HgI₂,¹¹³² Yb(OTf)₃,¹¹³³ Cu(OTf)₂,¹¹³⁴ LiClO₄,¹¹³⁵ VOCl₃,¹¹³⁶an Fe catalyst,¹¹³⁷ and Bi(OTf)₃.¹¹³⁸ Lithium perchlorate in acetonitrile (5 M) can be used for the reaction of an aldehyde and a silyl enol ether.¹¹³⁹ The reaction can be done in water using a Sc catalyst¹¹⁴⁰ or on a Montmorillonite K-10 clay.¹¹⁴¹ Silyl enol ethers react with aq formaldehyde in the presence of TBAF to give the aldol product.¹¹⁴² A catalytic amount of Me₃SiCl facilitates the Ti mediated reaction.¹¹⁴³ Sulfonamides (e.g., HNTf₂) have been used as a catalyst¹¹⁴⁴ as has pyridine N-oxide¹¹⁴⁵ and N-methylimidazole.¹¹⁴⁶ An ab initio study of the uncatalyzed Mukaiyama aldol reaction showed that the nucleophilicity of silyl enol ether and the electrophilicity of the aldehyde are important in promoting the reactivity.¹¹⁴⁷ This reaction can also be run with the aldehyde or ketone in the form of its acetal [R³R⁴C(OR′)₂], in which case the product is the ether R¹COCHR₂CR³R⁴OR′ instead of 41.¹¹⁴⁸ Vinylogous (Sec. 6.B) silyl ketene acetals with Ti,¹¹⁴⁹ Cu,¹¹⁵⁰ In,¹¹⁵¹ Fe,¹¹⁵² or Zn¹¹⁵³ catalysts or an organocatalyst,¹¹⁵⁴ give the product with good enantioselectivity.

Silyl enol ethers¹¹⁵⁵ derived from esters (silyl ketene acetals) react with aldehydes in the presence of various catalysts to give β-hydroxy esters. Water accelerates the reaction of an aldehyde and a ketene silyl acetal with no other additives.¹¹⁵⁶ The reaction is catalyzed by triphenylphosphine¹¹⁵⁷ and also by SiCl₄ with a chiral bis(phosphoramide) catalyst.¹¹⁵⁸ The reaction was done without a catalyst in an ionic liquid.¹¹⁵⁹ A vinylogous reaction (Sec. 6.B) is known that gives δ-hydroxy-α,β-unsaturated esters.¹¹⁶⁰ An interesting variation in this reaction combined an intermolecular Mukaiyama aldol followed by an intramolecular reaction (a “domino” Mukaiyama aldol) that gave cyclic conjugated ketone products.¹¹⁶¹ Under different conditions, silyl ketene acetals of conjugated esters react with aldehydes to give conjugated lactones.¹¹⁶² Imines react with silyl ketene acetals in the presence of SmI₃ to give β-amino esters.¹¹⁶³ Imines react with silyl enol ethers in the presence of BF₃·OEt₂ to give β-amino ketones.¹¹⁶⁴ Silyl ketene acetals also undergo conjugate addition in reactions with conjugated ketones.¹¹⁶⁵ Propargylic acetals react with silyl enol ethers and a Sc catalyst to give β-alkoxy ketones.¹¹⁶⁶ α-Silyl silyl enol ethers [RCH=CH(OTMS)SiMe₃] react with acetals in the presence of SnCl₄ to give β-alkoxy silyl ketones.¹¹⁶⁷ Borane derivatives [e.g., C=C–OB(NMe₂)₂] react with aldehydes to give β-amino ketones.¹¹⁶⁸

Asymmetric Mukaiyama aldol reactions and reactions of silyl ketene acetals have been reported,¹¹⁶⁹ usually using chiral additives¹¹⁷⁰ although chiral auxiliaries have also been used.¹¹⁷¹ Chiral catalysts, usually transition metal complexes using chiral ligands, are quite effective¹¹⁷² but chiral organocatalysts¹¹⁷³ are increasingly important.

Enol acetates and enol ethers also give this product when treated with acetals and TiCl₄ or a similar catalyst.¹¹⁷⁴ A variation of this condensation uses an enol acetate with an aldehyde in the presence of Et₂AlOEt to give the aldol product.¹¹⁷⁵

16-36 Aldol-Type Reactions between Carboxylic Acid Derivatives and Aldehydes or Ketones

O-Hydro-C-(α-alkoxycarbonylalkyl)-addition; α-Alkoxycarbonylalkylidene-de-oxo-bisubstitution

In the presence of a strong base, removal of a proton from the α carbon of a carboxylic ester or other acid derivative generates an enolate anion that can condense with the carbonyl carbon of an aldehyde or ketone to give a β-hydroxy ester,¹¹⁷⁶ amide, and so on. These products may or may not be dehydrated to the α,β-unsaturated derivative. This reaction is sometimes called the Claisen reaction,¹¹⁷⁷ an unfortunate usage since that name is more firmly connected to Reaction 16-85. Claisen condensation is a better descriptor. Early reactions used hydroxide or an alkoxide base in water or alcohol solvents, where self-condensation was the major process. Under such conditions, the aldehyde or ketone was usually chosen for its lack of an α-proton. Much better control of the reaction was achieved when dialkylamide bases in aprotic solvents (e.g., ether or THF) were used. The reaction of tert-butyl acetate and LDA¹¹⁷⁸ in hexane or more commonly THF at –78 °C gives the enolate anion of tert-butyl acetate,¹¹⁷⁹ (Reaction 12-23, e.g., although self-condensation is occasionally a problem even here). Additives play an important role in the LDA mediated enolization of esters¹¹⁸⁰. Subsequent reaction of a ketone provides a simple rapid alternative to the Reformatsky Reaction (16-28) as a means of preparing β-hydroxy tert-butyl esters. It is also possible for the α carbon of an aldehyde or ketone to add to the carbonyl carbon of a carboxylic ester, but this is a different reaction (16-86) involving nucleophilic substitution and not addition to a C=O bond. It can, however, be a side reaction if the aldehyde or ketone has an α hydrogen.

Transition metal mediated condensation of esters with aldehydes is known. The reaction of a thioester and an aryl aldehyde with TiCl₄–NBu₃, for example, gave a β-hydroxy thioester with good syn selectivity.¹¹⁸¹ Selenoamides [RCH₂C(=Se)NR′₂] react with LDA and then an aldehyde to give β-hydroxy selenoamides.¹¹⁸² The reaction of an α,β-unsaturated ester and benzaldehyde with a chiral Rh catalyst gave a β-hydroxy ester with good diastereoselectivity and good enantioselectivity.¹¹⁸³

Besides ordinary esters (containing an α hydrogen), the reaction can also be carried out with lactones and, as in Reaction 16-34, with the γ position of α,β-unsaturated esters (vinylogy; Sec. 6.B). The enolate anion of an amide can be condensed with an aldehyde.¹¹⁸⁴ Thioesters undergo aldol-type condensations.¹¹⁸⁵

For most esters, a much stronger base is needed than for aldol reactions; (iPr)₂NLi (LDA, Sec. 8.F, category 7), Ph₃CNa and LiNH₂ are among those employed. However, esters of malonic and succinic acid react more easily, and such strong bases are not needed. For example, diethyl succinate and its derivatives condense with aldehydes and ketones in the presence of bases (e.g., NaOEt, NaH, or KOCMe₃). This reaction is called the Stobbe condensation.¹¹⁸⁶One of the ester groups (sometimes both) is hydrolyzed in the course of the reaction. The following mechanism accounts for (1) the fact the succinic esters react so much better than others; (2) one ester group is always cleaved; and (3) the alcohol is not the product but the alkene. In addition, intermediate lactones (42) have been isolated from the mixture.¹¹⁸⁷ The Stobbe condensation has been extended to di-tert-butyl esters of glutaric acid.¹¹⁸⁸ The boron-mediated reaction is known.¹¹⁸⁹

The intramolecular reaction of an aldehyde–carboxylic acid in the presence of triethylamine and a pyridinium salt led to the ring-forming condensation reaction, followed by formation of a β-lactone.¹¹⁹⁰

Amides participate in this condensation reaction, reacting with aldehydes in the presence of a Ba catalyst to give a β-hydroxy amide derivative.¹¹⁹¹ The reaction of an amide with LDA in the presence of an acyl silane, followed by reaction with an alkyl halide, leads to the β-hydroxy amide with the additional alkyl group at the β-carbon.¹¹⁹²

Chiral additives (e.g., diazaborolidines) can be added to an ester, and subsequent treatment with a base, and then an aldehyde leads to a chiral β-hydroxy ester.¹¹⁹³ A variety of chiral amide or oxazolidinone derivatives have been used to form amide linkages to carboxylic acid derivatives. These chiral auxiliaries lead to chirality transfer from the enolate anion of such derivatives, in both alkylation reactions and acyl substitution reactions with aldehydes and ketones. The so-called Evans auxiliaries (43–45) are commonly used and give good enantioselectivity.¹¹⁹⁴ A variation is the magnesium halide-catalyzed anti-aldol reaction of chiral N-acylthiazolidinethiones (see 46).¹¹⁹⁵ The use of chiral N-acyloxazolidinthiones with TiCl₄ and sparteine also gave good selectivity in the acyl addition.¹¹⁹⁶ Chiral diazaboron derivatives have also been used to facilitate the condensation of a α-phenylthio ester with an aldehyde.¹¹⁹⁷

The condensation of an ester enolate and a ketone¹¹⁹⁸ can be used as part of a Robinson annulation-like sequence (see Reaction 16-34).

OS I, 252; III, 132; V, 80, 564; 70, 256; X, 437; 81, 157. Also see, OS IV, 278, 478; V, 251.

16-37 The Henry Reaction¹¹⁹⁹

The classical condensation of an aliphatic nitro compound with an aldehyde or ketone is usually called the Henry reaction¹²⁰⁰ or the Kamlet reaction, and is essentially a nitro aldol reaction. A variety of conditions have been reported, including the use of a recoverable polymer catalyst,¹²⁰¹ a silica catalyst,¹²⁰² a tetraalkylammonium hydroxide,¹²⁰³ proazaphosphatranes,¹²⁰⁴ and it has been done in an aqueous media¹²⁰⁵ or an ionic liquid.¹²⁰⁶ A solvent-free Henry reaction was reported in which a nitroalkane and an aldehyde were reacted on KOH powder.¹²⁰⁷ A solvent-free microwave assisted reaction was reported.¹²⁰⁸ Potassium phosphate has been used with nitromethane and aryl aldehydes.¹²⁰⁹ The Henry reaction has been done using ZnEt₂ and 20% ethanolamine.¹²¹⁰ A gel-entrapped base has been used to catalyze this reaction.¹²¹¹ Biocatalysts have been used.¹²¹²

Catalytic enantioselective Henry reactions are known,¹²¹³ including the use of a chiral Cu,¹²¹⁴ Zn,¹²¹⁵ Nd,¹²¹⁶ or Ti catalyst.¹²¹⁷ The Henry reaction of nitromethane and a chiral aldehyde under high pressure gives the β-nitro alcohol with excellent enantioselectivity.¹²¹⁸ Enantioselective nitro–aldol reactions are catalyzed by organocatalysts (e.g., Cinchona alkaloids¹²¹⁹ or other organocatalysts).¹²²⁰

A variation of this reaction converts nitro compounds to nitronates [RCH=N⁺(OTMS)–O⁻], which subsequently react with aldehydes in the presence of a Cu catalyst to give the β-nitro alcohol.¹²²¹ aza-Henry reactions condense nitroalkanes with imine derivatives, and the resulting amino nitro compounds are formed with good enantioselectivity in the presence of organocatalysts¹²²² or Brnsted acid catalysts.¹²²³ Aza-Henry products are also formed by the reaction of amines with activated unsaturated compounds.¹²²⁴

16-38 The Knoevenagel Reaction

Bis(ethoxycarbonyl)methylene-de-oxo-bisubstitution, and so on

The condensation of aldehydes or ketones, usually not containing an α hydrogen, with compounds of the form Z–CH₂–Z' or Z–CHR–Z' is called the Knoevenagel reaction.¹²²⁵ Both Z and Z' may be CHO, COR, CO₂H, CO₂R, CN, NO₂, SOR, SO₂R, SO₂OR, or similar groups. The presence of two electron-withdrawing groups makes the α-proton much more acidic (Table 8.1 in Sec. 8.A.i), and such compounds have a significantly higher enol content.¹²²⁶ When Z = CO₂H, decarboxylation of the product often takes place in situ.¹²²⁷ As shown in the example, the reaction of β-keto esters and aldehydes to give 47 is promoted by diethylamine at 0 °C. Nitroalkanes,¹¹⁹⁹ as well as β-keto sulfoxides,¹²²⁸ undergo the reaction.

As with Reaction 16-34, these reactions have sometimes been mediated by an acid catalyst.¹²²⁹ Ionic liquid solvents have been used,¹²³⁰ and heating on quaternary ammonium salts without solvent leads to a Knoevenagel reaction.¹²³¹ Other solvent-free reactions are known.¹²³² Ultrasound has been used to promote the reaction,¹²³³ and it has also been done using microwave irradiation¹²³⁴ or on silica,¹²³⁵ with microwave irradiation. Another solid-state variation is done on moist LiBr,¹²³⁶ heating with sodium carbonate and molecular sieves 4 Å promotes the reaction,¹²³⁷ as do zeolites.¹²³⁸ High-pressure conditions have been used.¹²³⁹ Transition metal compounds of Pd,¹²⁴⁰Sm¹²⁴¹ Ce,¹²⁴² Ti,¹²⁴³ or Bi¹²⁴⁴ have been used to promote the Knoevenagel reaction.

With most of these reagents the alcohol is not isolated (only the alkene) if the alcohol has a hydrogen in the proper position,¹²⁴⁵ but with a careful workup the alcohol may be the major product. With suitable reactants, the Knoevenagel reaction, like the aldol condensation (16-34), has been carried out diastereoselectively¹²⁴⁶ and enantioselectively.¹²⁴⁷ When the reactant is of the form ZCH₂Z', aldehydes react much better than ketones and few successful reactions with ketones have been reported. However, it is possible to get good yields of the alkene from the condensation of diethyl malonate [CH₂(CO₂Et)₂] with ketones, as well as with aldehydes, if the reaction is run with TiCl₄ and pyridine in THF.¹²⁴⁸ In reactions with ZCH₂Z', the catalyst is most often a secondary amine (piperidine is the most common, but see formation of 47), but many other catalysts have been used. Alkoxides are also common catalysts. When the catalyst is pyridine (to which piperidine may or may not be added) the reaction is known as the Doebner modification of the Knoevenagel reaction and the product is usually the conjugated acid (48). Microwave-induced Doebner condensation reactions are known.¹²⁴⁹

A number of special applications of the Knoevenagel reaction follow:

1. The dilithio derivative of N-methanesulfinyl-p-toluidine¹²⁵⁰ (49) adds to aldehydes and ketones to give, after hydrolysis, the hydroxysulfinamides (50). Subsequent heating leads to a stereospecific syn eliminations to give an alkene.¹²⁵¹ The reaction is thus a method for achieving the conversion RR′CO → RR′C=CH₂ and represents an alternative to the Wittig reaction.¹²⁵² Note that sulfones with an amide group at the α-position, [ArSO₂CH(R)N(R)C=O] react with ketones via acyl addition in the presence of SmI₂.¹²⁵³

2. The reaction of ketones with tosylmethylisocyanide (51) gives different products,¹²⁵⁴ depending on the reaction conditions. When the reaction is run with potassium tert-butoxide in THF at –5 °C, one obtains (after hydrolysis) the normal Knoevenagel product (52), except that the isocyano group has been hydrated (Reaction 16-97).¹²⁵⁵ With the same base but with 1,2-dimethoxyethane (DME) as solvent the product is the nitrile (53).¹²⁵⁶When the ketone is treated with 51 and thallium(I) ethoxide in a 4: 1 mixture of absolute ethanol and DME at room temperature, the product is a 4-ethoxy-2-oxazoline (54).¹²⁵⁷ Since 53 can be hydrolyzed to a carboxylic acid¹¹⁹⁸ and 54 to an α-hydroxy aldehyde,¹²⁵⁷ this versatile reaction provides a means for achieving the conversion of RCOR′ to RCHR′CO₂H, RCHR′CN, or RCR′(OH)CHO. The conversions to RCHR′COOH and to RCHR′CN¹²⁵⁸ have also been carried out with certain aldehydes (R′ = H).

3. Aldehydes and ketones (RCOR′) react with α-methoxyvinyllithium [CH₂=C(Li)OMe] to give hydroxy enol ethers [RR′C(OH)C(OMe)=CH₂], which are easily hydrolyzed to acyloins [RR′C(OH)COMe].¹²⁵⁹ In this reaction, the CH₂=C(Li)OMe is a synthon for the unavailable H₃C–C=O,¹²⁶⁰ and is termed an acyl anion equivalent. The reagent also reacts with esters (RCOOR′) to give RC(OH)(COMe=CH₂)₂. A synthon for the Ph–C=O ion is PhC(CN)OSiMe₃, which adds to aldehydes and ketones (RCOR′) to give, after hydrolysis, the α-hydroxy ketones [RR′C(OH)COPh].¹²⁶¹

4. Lithiated allylic carbamates (55) (prepared as shown) react with aldehydes or ketones (R⁶COR⁷), in a reaction accompanied by an allylic rearrangement, to give (after hydrolysis) γ-hydroxy aldehydes or ketones.¹²⁶² The reaction is called the homoaldol reaction, since the product is a homologue of the product of Reaction 16-34. The reaction has been performed enantioselectively.¹²⁶³

5. The lithium salt of an active hydrogen compound adds to the lithium salt of the tosylhydrazone of an aldehyde to give product 56. If X = CN, SPh, or SO₂R, 56 spontaneously loses N₂ and LiX to give the alkene 57. The entire process is done in one reaction vessel: The active hydrogen compound is mixed with the tosylhydrazone and the mixture is treated with (iPr)₂NLi to form both salts at once.¹²⁶⁴ This process is another alternative to the Wittig reaction for forming double bonds.

OS I, 181, 290, 413; II, 202; III, 39, 165, 317, 320, 377, 385, 399, 416, 425, 456, 479, 513, 586, 591, 597, 715, 783; IV, 93, 210, 221, 234, 293, 327, 387, 392, 408, 441, 463, 471, 549, 573, 730, 731, 777; V, 130, 381, 572, 585, 627, 833, 1088, 1128; VI, 41, 95, 442, 598, 683; VII, 50, 108, 142, 276, 381, 386, 456; VIII, 258, 265, 309, 353, 391, 420; X, 271. Also see, OS III, 395; V, 450.

16-39 The Perkin Reaction

α-Carboxyalkylidene-de-oxo-bisubstitution

The condensation of aromatic aldehydes with anhydrides is called the Perkin reaction.¹²⁶⁵ When the anhydride has two α hydrogen atoms (as shown), dehydration almost always occurs; the β-hydroxy acid salt is rarely isolated. In some cases, anhydrides of the form (R₂CHCO)₂O have been used, and then the hydroxy compound is the product since dehydration cannot take place. The base in the Perkin reaction is nearly always the salt of the acid corresponding to the anhydride. Although the Na and K salts have been most frequently used, higher yields and shorter reaction times have been reported for the Cs salt.¹²⁶⁶ Besides aromatic aldehydes, their vinylogs (ArCH=CHCHO) also give the reaction (see Sec. 6.B). Otherwise, the reaction is not suitable for aliphatic aldehydes.¹²⁶⁷

OS I, 398; II, 61, 229; III, 426.

16-40 Darzens Glycidic Ester Condensation

(2+1) OC,CC-cyclo-α-Alkoxycarbonylmethylene-addition

Aldehydes and ketones condense with α-halo esters in the presence of bases to give α,β-epoxy esters, called glycidic esters. This is called the Darzens condensation.¹²⁶⁸ The reaction consists of an initial Knoevenagel-typeReaction (16-38), followed by an internal S_N² Reaction (10-9)¹²⁶⁹:

Although the intermediate halo alkoxide is generally not isolated,¹²⁷⁰ it has been done, not only with α-fluoro esters (since fluorine is such a poor leaving group in nucleophilic substitutions), but also with α-chloro esters.¹²⁷¹ This is only one of several types of evidence that rule out a carbene intermediate.¹²⁷² Sodium ethoxide is often used as the base, but other bases, including sodium amide, are sometimes used. Aromatic aldehydes and ketones give good yields. The reaction can be made to give good yields (~ 80%) with simple aliphatic aldehydes, as well as with aromatic aldehydes and ketones, by treatment of the α-halo ester with the base lithium bis(trimethylsilyl)amide [LiN(SiMe₃)₂] in THF at –78 °C (to form the conjugate base of the ester) and addition of the aldehyde or ketone to this solution.¹²⁷³ If a preformed dianion of an α-halo carboxylic acid (Cl⁻CRCO₂⁻) is used instead, α,β-epoxy acids are produced directly.¹²⁷⁴ The Darzens reaction has also been carried out on α-halo ketones, α-halo nitriles,¹²⁷⁵ α-halo sulfoxides¹²⁷⁶ and sulfones,¹²⁷⁷ α-halo N,N-disubstituted amides,¹²⁷⁸ α-halo ketimines,¹²⁷⁹ and even on allylic¹²⁸⁰and benzylic halides. Phase-transfer catalysis has been used.¹²⁸¹ Note that the reaction of a β-bromo-α-oxo ester and a Grignard reagent leads to the glycidic ester.¹²⁸² Acid-catalyzed Darzens reactions have also been reported.¹²⁸³(see also, Reaction 16-46).

Diastereoselective Darzens condensations are possible.¹²⁸⁴ The Darzens reaction has been performed with good enantioselectivity,¹²⁸⁵ and chiral additives have proven to be effective.¹²⁸⁶ Chiral phase-transfer agents have been used to give epoxy ketones with modest enantioselectivity.¹²⁸⁷

Glycidic esters can easily be converted to aldehydes (Reaction 12-40). The reaction has been extended to the formation of analogous aziridines by treatment of an imine with an α-halo ester or an α-halo N,N-disubstituted amide and t-BuOK in the solvent 1,2-dimethoxyethane.¹²⁸⁸ However, yields were not high.

OS III, 727; IV, 459, 649.

16-41 The Peterson Alkenylation Reaction

Alkylidene-de-oxo-bisubstitution

In the Peterson alkenylation reaction^1289, the lithio (or sometimes magnesio) derivative of a trialkylsilane adds to an aldehyde or ketone to give a β-hydroxysilane, which spontaneously eliminates water, or can be made to do so by treatment with acid or base, to produce an alkene. This reaction is still another alternative to the Wittig reaction (16-44), and is sometimes called the silyl–Wittig reaction.¹²⁹⁰ The R group can also be a COOR group, in which case the product is an α,β-unsaturated ester,¹²⁹¹ or an SO₂Ph group, in which case the product is a vinylic sulfone.¹²⁹² The stereochemistry of the product can often be controlled by whether an acid or a base is used to achieve elimination. The role of Si–O interactions has also been examined.¹²⁹³ Use of a base generally gives syn elimination (Ei mechanism, see Sec. 17.C.i), while an acid usually results in anti elimination (E2 mechanism, see Sec. 17.A.i).¹²⁹⁴ Samarium(II) iodide in HMPA has also been used for elimination of the hydroxy sulfone.¹²⁹⁵

When aldehydes or ketones are treated with reagents of the form 58, the product is an epoxy silane (Reaction 16-46), which can be hydrolyzed to a methyl ketone.¹²⁹⁶ For aldehydes, this is a method for converting RCHO to a methyl ketone (RCH₂COMe).

The reagents Me₃SiCHRM (M = Li or Mg) are often prepared from Me₃SiCHRCl¹²⁹⁷ (by Reaction 12-38 or 12-39), but they have also been made by Reaction 12-22 and other procedures.¹²⁹⁸ Lithio alkenylsilanes have been used for this reaction.¹²⁹⁹

A new version of the reaction has been developed, reacting Me₃SiCH₂CO₂Et with an aldehyde and a catalytic amount of CsF in DMSO.¹³⁰⁰ A seleno-amide derivative has been used in a similar manner.¹³⁰¹

There are no references in Organic Syntheses, but see OS VIII, 602, for a related reaction.

16-42 The Addition of Active Hydrogen Compounds to CO₂ and CS₂

α-Acylalkyl-de-methoxy-substitution (Overall reaction)

Ketones of the form RCOCH₃ and RCOCH₂R′ can be carboxylated indirectly by treatment with magnesium methyl carbonate (59).¹³⁰² Because formation of the chelate (60) provides the driving force of the reaction, carboxylation cannot be achieved at a disubstituted α position. The reaction has also been performed on CH₃NO₂, on compounds of the form RCH₂NO₂¹³⁰³ and on certain lactones.¹³⁰⁴ Direct carboxylation has been reported in a number of instances. Ketones have been carboxylated in the α position to give β-keto acids.¹³⁰⁵ The base here was lithium 4-methyl-2,6-di-tert-butylphenoxide.

Ketones (RCOCH₂R′), as well as other active hydrogen compounds, undergo base-catalyzed addition to CS₂¹³⁰⁶ to give a dianion intermediate (RCOC⁻R′CSS^2-), which can be dialkylated with a halide (R²X) to produce α-dithiomethylene ketones [RCOCR′=C(SR²)₂].¹³⁰⁷ Compounds of the form ZCH₂Z' also react with bases and CS₂ to give analogous dianions.¹³⁰⁸

Although reactions with N=O derivatives do not formally fall into this category of reactions, it is somewhat related. Nitroso compounds react with activated nitriles in the presence of LiBr and microwave irradiation to give a cyano imine [ArN=C(CN)Ar].¹³⁰⁹ This transformation has been called the Ehrlich–Sachs reaction.¹³¹⁰

OS VII, 476. See also, OS VIII, 578.

16-43 Tollens' Reaction

O-Hydro-C (β-hydroxyalkyl)-addition

In the Tollens' reaction, an aldehyde or ketone containing an α hydrogen is treated with formaldehyde in the presence of Ca(OH)₂ or a similar base. The first step is a mixed-aldol reaction (16-34).

The reaction can be stopped at this point, but more often a second equivalent of formaldehyde is permitted to reduce the newly formed aldol to a 1,3-diol, in a crossed-Cannizzaro Reaction (19-81). If the aldehyde or ketone has several α hydrogen atoms, they can all be replaced. An important use of the reaction is to prepare pentaerythritol from acetaldehyde:

OS I, 425; IV, 907; V, 833.

16-44 The Wittig Reaction

Alkylidene-de-oxo-bisubstitution

In the Wittig reaction, an aldehyde or ketone is treated with a phosphorus ylid (a phosphorane; also spelled ylide) to give an alkene.¹³¹¹ The conversion of a carbonyl compound to an alkene with a phosphorus ylid is called the Wittig reaction. Phosphorus ylids are usually prepared by treatment of a phosphonium salt with a base,¹³¹² and phosphonium salts (61) are usually prepared from a triaryl phosphine and an alkyl halide (Reaction 10-31):

The reaction of triphenylphosphine and an alkyl halides is facilitated by the use of microwave irradiation.¹³¹³ Indeed, the Wittig reaction itself is assisted by microwave irradiation.¹³¹⁴ Phosphonium salts are also prepared by addition of phosphines to Michael alkenes (like Reaction 15-8) and in other ways.

The phosphonium salts are most often converted to the ylids by treatment with a strong base (e.g., butyllithium, sodium amide,¹³¹⁵ sodium hydride, or a sodium alkoxide, though weaker bases can be used if the salt is acidic enough. In some cases, and excess of fluoride ion is sufficient.¹³¹⁶ For (Ph₃P⁺)₂CH₂, sodium carbonate is a strong enough base.¹³¹⁷ When the base used does not contain lithium, the ylid is said to be prepared under “salt-free” conditions¹³¹⁸ because the lithium halide (where the halide counterion comes from the phosphonium salt) is absent. Wittig reactions can be done in aqueous media in the presence of surfactants.¹³¹⁹

When the phosphorus ylid reacts with the aldehyde or ketone to form an alkene, a phosphine oxide is also formed. When triphenylphosphine is used to give Ph₃P=CRR′, for example, the byproduct is triphenylphosphine oxide (Ph₃PO), which is sometimes difficult to separate from the other reaction products. Other triarylphosphines¹³²⁰ and trialkylphosphines¹³²¹ have been used. Phosphines that have an α-hydrogen should be avoided, so that reaction with the chosen alkyl halide will lead to a phosphonium salt (61) with the α-proton at the desired position. This limitation is essential if a specific ylid is to be formed from the alkyl halide precursor. The Wittig reaction has been carried out with polymer-supported ylids.¹³²² It has also been done on silica gel.¹³²³ Polymer-bound aryldiphenylphosphino compounds¹³²⁴ have been used in reactions with alkyl halides to complete a Wittig reaction.

If an alkyl halide is viewed as the starting material (alkyl halide phosphonium salt → phosphorus ylid → alkene), the halogen-bearing carbon of an alkyl halide must contain at least one hydrogen, as in 62 (for deprotonation at the phosphonium salt stage).

The reaction is very general.¹³²⁵ The aldehyde or ketone may be aliphatic, alicyclic, or aromatic (including diaryl ketones). Wittig reactions are known in which the ylid and/or the carbonyl substrate contain double or triple bonds; various functional groups may be present (e.g., OH, OR, NR₂, aromatic nitro or halo, acetal, amide,¹³²⁶ or even ester groups).¹³²⁷ Note, however, that a Wittig reaction has been reported in which the carbonyl group of an ester was converted to a vinyl ether.¹³²⁸ An important advantage of the Wittig reaction is that the position of the new double bond is always certain, in contrast to the result in most of the base-catalyzed condensations (Reactions 16-34–16-43). Ylids have been shown to react with lactones, however, to form ω-alkenyl alcohols.¹³²⁹ β-Lactams have also been converted to alkenyl-azetidine derivatives using phosphorus ylids.¹³³⁰ Double or triple bonds conjugated with the carbonyl also do not interfere, the attack being at the C=O carbon. The carbonyl partner can be generated in situ, in the presence of an ylid; the reaction of an alcohol with a mixture of an oxidizing agent and an ylid generates an alkene. Oxidizing agents used in this manner include BaMnO₄,¹³³¹ MnO₂,¹³³² and PhI(OAc)₂.¹³³³ Weinreb amides (Reaction 16-82) can be converted to the corresponding aldehyde via the Wittig reaction.¹³³⁴

As noted above, the phosphorus ylid may also contain double or triple bonds and certain functional groups. Simple ylids (R, R′ = hydrogen or alkyl) are highly reactive, reacting with oxygen, water, hydrohalic acids, and alcohols, as well as carbonyl compounds and carboxylic esters, so the reaction must be run under conditions where these materials are absent. When an electron-withdrawing group (e.g., COR, CN, CO₂R, CHO) is present in the α position, the ylids are much more stable, because the charge on the carbon is delocalized by

resonance as in 63. Such ylids react readily with aldehydes, but slowly or not at all with ketones.¹³³⁵ In extreme cases (e.g., 64 where the carbanion unit is part of the aromatic cyclopentadienyl anion), the ylid does not react with ketones or aldehydes. Besides these groups, the ylid may contain one or two α halogens¹³³⁶ or

an α OR or OAr group. In the latter case, the product is an enol ether, which can be hydrolyzed (Reaction 10-6) to an aldehyde,¹³³⁷ so that this reaction is a means of achieving the conversion RCOR′ → RR′CHCHO.¹³³⁸ However, the ylid may not contain an α nitro group. If the phosphonium salt contains a potential leaving group (e.g., Br or OMe) in the β position, treatment with a base gives elimination, instead of the ylid:

However, a β COO⁻ group may be present, and the product is a β,γ-unsaturated acid:¹³³⁹ This is the only convenient way to make these compounds, since elimination by any other route gives the thermodynamically more stable α,β-unsaturated isomers. This is an illustration of the utility of the Wittig method for the specific location of a double bond. Another illustration is the conversion of cyclohexanones to alkenes containing exocyclic double bonds, for example:¹³⁴⁰

Still another example is the formation of anti-Bredt bicycloalkenones¹³⁴¹ (see Sec. 4.Q.iii). As indicated above, α,α'-dihalophosphoranes can be used to prepare 1,1-dihaloalkenes. Another way to prepare haloalkenes¹³⁴² is to treat the carbonyl compound with a mixture of CX₄ (X = Cl, Br, or I) and triphenylphosphine, either with or without the addition of zinc dust, which allows less Ph₃P to be used.¹³⁴³ Aryl aldehydes react with these dihalophosphoranes to give aryl alkynes after treatment of the initially formed vinyl halide with potassium tert-butoxide.¹³⁴⁴

The mechanism¹³⁴⁵ of the key step of the Wittig reaction is as follows:¹³⁴⁶

The energetics of ylid formation and their reaction in solution has been studied.¹³⁴⁷ For many years it was assumed that a diionic compound, called a betaine, is an intermediate on the pathway from the starting compounds to the oxaphosphetane, but it has been argued that there is little evidence for it.¹³⁴⁸ However, “betaine” precipitates have been isolated in certain Wittig reactions,¹³⁴⁹ although these are betaine–lithium halide adducts, and might just as well have been formed from the oxaphosphetane as from a true betaine.¹³⁵⁰ There is one report of an observed betaine lithium salt during the course of a Wittig reaction.¹³⁵¹ An X-ray structure was determined for a gauche betaine from a thio-Wittig reaction.¹³⁵² In contrast, there is much evidence for the presence of the oxaphosphetane intermediates, at least with unstable ylids. For example, ³¹P NMR spectra taken of the reaction mixtures at low temperatures¹³⁵³ are compatible with an oxaphosphetane structure that persists for some time but not with a tetra-coordinated phosphorus species. Since a betaine, an ylid, and a phosphine oxide all have tetracoordinated phosphorus, these species could not be causing the spectra, leading to the conclusion that an oxaphosphetane intermediate is present in the solution. In certain cases, oxaphosphetanes have been isolated.¹³⁵⁴ It has even been possible to detect cis and trans isomers of the intermediate oxaphosphetanes by NMR spectroscopy.¹³⁵⁵ According to this mechanism, an optically active phosphonium salt (RR¹R²P⁺CHR²) should retain its configuration all the way through the reaction, and it should be preserved in the phosphine oxide (RR¹R²PO). This has been shown to be the case.¹³⁵⁶

Note that the proposed betaine intermediates can be formed, in a completely different manner, by nucleophilic substitution by a phosphine on an epoxide (10-35):

Betaines formed in this way can then be converted to the alkene. This is one reason why betaine intermediates were long accepted in the Wittig reaction. It is also noteworthy that stable phosphonium enolate zwitterions have been formed by the reaction of an aryl aldehyde, a phosphine, and a propargylic ester.¹³⁵⁷

Phosphorus is not the only key element used to produce useful ylids. Triphenylarsine¹³⁵⁸ has been used. Tellurium ylids have been prepared in situ from α-halo esters and BrTeBu₂OTeBu₂Br and react with aldehydes to give conjugated esters.¹³⁵⁹

The Wittig reaction has been carried out with phosphorus ylids other than phosphoranes, the most important being prepared from phosphonate esters (e.g., 65).¹³⁶⁰

This method, sometimes called the Horner–Emmons, Wadsworth–Emmons, Wittig–Horner, or Horner–Wadsworth–Emmons reaction,¹³⁶¹ has several advantages over the use of phosphoranes, including selectivity.¹³⁶² These ylids are more reactive than the corresponding phosphoranes, and when R¹ or R² is an electron-withdrawing group, these compounds often react with ketones that are inert to phosphoranes. High pressure has been used to facilitate this reaction.¹³⁶³ In addition, the phosphorus product is a phosphate ester, and hence soluble in water, unlike Ph₃PO, which makes it easy to separate it from the alkene product. Phosphonates are also cheaper than phosphonium salts and can easily be prepared by the Arbuzov reaction:¹³⁶⁴

Phosphonates have also been prepared from alcohols and (ArO)₂P(=O)Cl, NEt₃, and a TiCl₄ catalyst.¹³⁶⁵ The reaction of (RO)₂P(=O)H and aryl iodides with a CuI catalyst leads to aryl phosphonates.¹³⁶⁶ Polymer-bound phosphonate esters have been used for olefination.¹³⁶⁷ Dienes are produced when allylic phosphonate esters react with aldehydes.¹³⁶⁸ Nucleophilicity parameters have been determined for phosphoryl-stabilized carbanions.¹³⁶⁹ A Zn promoted reaction is known using diprotic phosphonates.¹³⁷⁰ Wittig reactions of stabilized phosphorus ylids have also been done in water.¹³⁷¹

Stereoselective alkenylation reactions have been achieved using chiral additives¹³⁷² or auxiliaries.¹³⁷³ Ylids formed from phosphine oxides [Ar₂P(=O)CHRR′], phosphonic acid bis(amides), (R₂²N)₂POCHRR¹],¹³⁷⁴ and alkyl phosphonothionates [(MeO)₂PSCHRR¹],¹³⁷⁵ share some of these advantages. Reagents (e.g., Ph₂POCH₂NR₂') react with aldehydes or ketones (R²COR³) to give good yields of enamines (R²R³C=CHNR).¹³⁷⁶ (Z)-Selective reagents are also known,¹³⁷⁷ including the use of a di(2,2,2-trifluoroethoxy)phosphonate with KHMDS and 18-crown-6.¹³⁷⁸ An interesting intramolecular version of the Horner–Emmons reaction leads to alkynes.¹³⁷⁹ The reaction of a functionalized aldehyde (R–CHO) with (MeO)₂POCHN₂, leads to an alkyne (R–CCH).¹³⁸⁰

Some Wittig reactions give the (Z)-alkene; some the (E), and others give mixtures. The question of which factors determine the stereoselectivity has been much studied.¹³⁸¹ It is generally found that ylids containing stabilizing groups or formed from trialkylphosphines give (E)-alkenes. The origin of the (E) selectivity in salt-free stabilized ylids may be related to dipole–dipole interactions.¹³⁸² The energy of the elimination transition state must also be taken into account.¹³⁸³ It has been shown that ylids formed from triarylphosphines and not containing stabilizing groups often give (Z) or a mixture of (Z)- and (E)-alkenes.¹³⁸⁴ One explanation for this¹³⁵⁶ is that the reaction of the ylid with the carbonyl compound is a [2+2]-cycloaddition, which in order to be concerted must adopt the [_π2_s + _π2_a]-pathway. As discussed in Reaction 15-63, this pathway leads to the formation of the more sterically crowded product, in this case the (Z)-alkene. If this explanation is correct, it is not easy to explain the predominant formation of (E) products from stable ylids, but (E) compounds are of course generally thermodynamically more stable than the (Z) isomers, and the stereochemistry seems to depend on many factors.

The (E/Z) ratio of the product can often be changed by a change in solvent or by the addition of salts.¹³⁸⁵ Another way of controlling the stereochemistry of the product is by use of the aforementioned phosphonic acid bis(amides). In this case, the betaine (66) does form and when treated with water gives the β-hydroxyphosphonic acid bis(amides) (67), which can be crystallized and then cleaved to R¹R²C=CR³R⁴ by refluxing in benzene or toluene in the presence of silica gel.¹³⁷⁴ β-Hydroxy products (67) are generally formed as mixtures of diastereomers, and these mixtures can be separated by recrystallization. Each diastereomer will give one of the two isomeric alkenes. Optically active phosphonic acid bis(amides) have been used to give optically active alkenes.¹³⁸⁶ Another method of controlling the stereochemistry of the alkene [to obtain either the (Z)- or (E)-isomer] starting with a phosphine oxide (Ph₂POCH₂R) has been reported.¹³⁸⁷

In reactions where the betaine–lithium halide intermediate is present, it is possible to extend the chain further if a hydrogen is present α to the phosphorus. For example, reaction of ethylidenetriphenylphosphorane with heptanal at −78 °C gave 68, and subsequent treatment with butyllithium gave the ylid, (69). Treatment of 69 with an aldehyde (R′CHO) gave the intermediate (70) that gave, after workup, 71¹³⁸⁸ stereoselectively. Alkoxide 69 also reacts with other electrophiles. For example, treatment of 69 with NCS or PhICl₂ gave the vinylic chloride (RCH=CMeCl) stereoselectively: NCS gave the cis and PhICl₂ the trans-isomer.¹³⁸⁹ The use of Br₂ and FClO₃ (several explosions¹³⁹⁰have been observed with this reagent) gives the corresponding bromide or fluoride, respectively.¹³⁹¹ Reactions of 69 with electrophiles have been called scoopy reactions (α substitution plus carbonyl alkenylation via β-oxido phosphorus ylids).¹³⁹²

The reaction of a phosphonate ester, DBU, NaI, and HMPA with an aldehyde leads to a conjugated ester with excellent (Z)-selectivity.¹³⁹³ A (Z)-selective reaction was reported using a trifluoroethyl phosphonate in a reaction with an aldehyde and potassium tert-butoxide.¹³⁹⁴

The Wittig reaction has been carried out intramolecularly, to prepare rings containing from 5 to 16 carbons,¹³⁹⁵ both by single-ring closure to give alkenes (e.g., 70), or by double-ring closure, as in the conversion of 71 to 72).¹³⁹⁶

The Wittig reaction has proved very useful in the synthesis of natural products, some of which are quite difficult to prepare in other ways.¹³⁹⁷

Phosphorus ylids also react in a similar manner with the C=O bonds of ketenes,¹³⁹⁸ isocyanates,¹³⁹⁹ certain anhydrides¹⁴⁰⁰ lactones,¹⁴⁰¹ and imides,¹⁴⁰² the N=O of nitroso groups, and the C=N of imines,¹⁴⁰³ as shown in the composite reaction. Phosphorus ylids react with carbon dioxide to give the isolable salts (74),¹⁴⁰⁴ which can be hydrolyzed to the carboxylic acids (75) (thus achieving the conversion RR′CHX → RR′CHCOOH) or (if neither R nor R′ is hydrogen) dimerized to allenes (76).

Although phosphorus ylids are most commonly used for alkenylation reactions, nitrogen ylids can occasionally be used. However, nitrogen ylids are often difficult to form, are unstable and highly reactive. The reaction of N-benzyl-N-phenylpiperidinium bromide with base to give an N-ylid is one example, and it reacted with benzaldehyde to form styrene.¹⁴⁰⁵ The structure has been determined for an intermediate in an aza-Wittig reaction.¹⁴⁰⁶ An aza-Wittig reaction¹⁴⁰⁷ has been used to prepare pyrrolines and tetrahydropyridines.¹⁴⁰⁸

OS V, 361, 390, 499, 509, 547, 751, 949, 985; VI, 358; VII, 164, 232; VIII, 265, 451; 75, 139, OS IX, 39, 230.

16-45 Tebbe, Petasis, and Alternative Alkenylations

Methylene-de-oxo-bisubstitution

A useful alternative to phosphorus ylids are Ti reagents (e.g., 77) prepared from dicyclopentadienyltitanium dichloride and trimethylaluminum.¹⁴⁰⁹ Treatment of a carbonyl compound with 77 (Tebbe reagent) in toluene–THF containing a small amount of pyridine¹⁴¹⁰ leads to the alkene. Dimethyltitanocene (Me₂TiCp₂), called the Petasis reagent, is a convenient and highly useful alternative to 77.¹⁴¹¹ The mechanism of Petasis olefination has been examined.¹⁴¹² Both the Tebbe and the Petasis reagent give good results with ketones.¹⁴¹³ An important feature of these new reagents is that carboxylic esters and lactones¹⁴¹⁴ can be converted to the corresponding enol ethers in good yields. The enol ether can be hydrolyzed to a ketone (Reaction 10-6), so this is also an indirect method for making the conversion RCO₂R′ → RCOCH₃ (see also, Reaction 16-82). Conjugated esters are converted to alkoxy-dienes with this reagent.¹⁴¹⁵ Lactams, including β-lactams, are converted with alkylidene cycloamines (alkylidene azetidines from β-lactams, which are easily hydrolyzed to β-amino ketones).¹⁴¹⁶

Besides stability and ease of preparation, another advantage of the Petasis reagent is that structural analogues can be prepared, including Cp₂Ti(C₃H₅)₂¹⁴¹⁷ (C₃H₅ = cyclopropyl), CpTi(CH₂SiMe₃)₃,¹⁴¹⁸ and Cp₂TiMe(CH=CH₂).¹⁴¹⁹In another variation, 2 molar equivalents of Cp₂Ti[P(OEt)₃]₂ reacted with a ketone in the presence of 1,1-diphenylthiocyclobutane to give the alkenylcyclobutane derivative.¹⁴²⁰ An alternative Ti reagent was prepared using TiCl₄, Mg metal and dichloromethane, reacting with both ketones¹⁴²¹ and esters¹⁴²² to give alkenes or vinyl ethers, respectively. Alkenes are generated form ketones and alkyl iodides in the presence of a catalytic amount of Cp₂Ti[POEt)₃]₂.¹⁴²³

Carboxylic esters undergo the conversion C=O → C=CHR (R = primary or secondary alkyl) when treated with RCHBr₂, Zn,¹⁴²⁴ and TiCl₄ in the presence of N,N,N',N'-tetramethylethylenediamine.¹⁴²⁵ Metal carbene complexes¹⁴²⁶R₂C=ML_n (L = ligand), where M is a transition metal (e.g., Zr, W, or Ta) have also been used to convert the C=O of carboxylic esters and lactones to CR₂.¹⁴²⁷ It is likely that the complex Cp₂Ti=CH₂ is an intermediate in the reaction with the Tebbe reagent. Indeed, Ti carbenoids have been used to convert carbonyl groups to the corresponding alkene.¹⁴²⁸

There are a few other methods for converting ketones or aldehydes to alkenes.¹⁴²⁹ Carbonyl compounds react with bis(iodozincio)methane to give alkenes.¹⁴³⁰ When a ketone is treated with CH₃CHBr₂/Sm/SmI₂, with a catalytic amount of CrCl₃, the alkene is formed.¹⁴³¹ α-Halo esters also react with CrCl₂ in the presence of a ketone to give vinyl halides.¹⁴³² Organozinc reagents have been used to convert carbonyl compounds to alkenes in the presence of Lewis acids.¹⁴³³ α-Diazo esters react with ketones in the presence of an iron catalyst to give the corresponding alkene.¹⁴³⁴ α-Diazo silylalkanes react similarly in the presence of a Rh catalyst.¹⁴³⁵ Ketone olefination has been accomplished using methyltrioxorhenium.¹⁴³⁶ α-Halosulfones react with aldehydes in the presence of LiHMDS and MgBr₂·OEt₂ to give a vinyl chloride.¹⁴³⁷

OS VIII, 512, IX, 404; X, 355.

16-46 The Formation of Epoxides from Aldehydes and Ketones

(1+2)OC,CC-cyclo-Methylene-addition

Aldehydes and ketones can be converted to epoxides¹⁴³⁸ in good yields with the sulfur ylids¹⁴³⁹ dimethyloxosulfonium methylid (78)¹⁴⁴⁰ and dimethylsulfonium methylid (79).¹⁴⁴¹ Ylid 79 is much less stable

and ordinarily must be used as soon as it is formed, while 78 can be stored several days at room temperature. When diastereomeric epoxides can be formed, 79 usually attacks from the more hindered and 78 from the less-hindered side. Thus, 4-tert-butylcyclohexanone, treated with 78 gave exclusively 80 while 79 gave mostly 81.¹⁴⁴² Another difference in behavior between the two reagents is that with α,β-unsaturated ketones, 78 gives

only cyclopropanes (Reaction 15-64), while 79 gives oxirane formation. Other sulfur ylids have been used in an analogous manner, to transfer CHR or CR₂.¹⁴⁴³ Other sulfur ylids convert aldehydes to epoxides.¹⁴⁴⁴ High yields have been achieved by the use of sulfonium ylids anchored to insoluble polymers under phase-transfer conditions.¹⁴⁴⁵ A solvent-free version of this reaction has been developed using powdered K tert-butoxide and Me₃S⁺I^–.¹⁴⁴⁶ Note that treatment of epoxides with 2 equiv of Me₂S=CH₂ leads to allylic alcohols.¹⁴⁴⁷

Chiral sulfur ylids¹⁴⁴⁸ have been prepared, giving the epoxide with good asymmetric induction,¹⁴⁴⁹ and chiral additives have also been used.¹⁴⁵⁰ Chiral Se ylids have been used in a similar manner.¹⁴⁵¹

The generally accepted mechanism for the reaction between sulfur ylids and aldehydes or ketone is formation of 82, with displacement of the Me₂S leaving group by the alkoxide.¹⁴⁵² This mechanism is similar to that of the reaction of sulfur ylids with C=C double bonds (Reaction 15-64).¹⁴⁵³ The stereochemical difference in the behavior of 78 and 79 has been attributed to formation of the betaine (82) being reversible for 78, but not for the less stable 79, so that the more-hindered product is the result of kinetic control and the less hindered of thermodynamic control.¹⁴⁵⁴

Phosphorus ylids do not give this reaction, but give 16-44 instead.

Aldehydes and ketones can also be converted to epoxides by treatment with a diazoalkane.¹⁴⁵⁵ Most commonly it is diazomethane, but an important side reaction is the formation of an aldehyde or ketone with one more carbon than the starting compound (Reaction 18-9). The reaction can be carried out with many aldehydes, ketones, and quinones, usually with a Rh catalyst.¹⁴⁵⁶ A mechanism that accounts for both products follows:

Compound 83 or nitrogen-containing derivatives of it have sometimes been isolated.

An alternative route to epoxides from ketones uses α-chloro sulfones and potassium tert-butoxide to give α,β-epoxy sulfones.¹⁴⁵⁷ A similar reaction was reported using KOH and 10% of a chiral phase-transfer agent, giving moderate enantioselectivity in the epoxy sulfone product.¹⁴⁵⁸

Dihalocarbenes and carbenoids, which readily add to C=C bonds (Reaction 15-64), do not generally add to the C=O bonds of ordinary aldehydes and ketones.¹⁴⁵⁹ See also, Reaction 16-91.

There is a report of a strained azetidinium ylid that has been used for epoxidation.¹⁴⁶⁰

OS V, 358, 755.

16-47 The Formation of Aziridines from Imines

(1+2)NC,CC-cyclo-Methylene-addition

Just as sulfur ylids (e.g., 78) react with the carbonyl of an aldehyde or ketone to give an epoxide, Te ylids react with imines to give an aziridine. The reaction of an allylic Te salt (RCH=CHCH₂Te⁺Bu₂ Br⁻), with lithium hexamethyldisilazide in HMPA/toluene leads to the tellurium ylid via deprotonation. In the presence of an imine, the ylid add to the imine and subsequent displacement of Bu₂Te generates an aziridine with a pendant vinyl group.¹⁴⁶¹Catalytic aziridination of tosylimines was reported and mediated by arsonium ylids.¹⁴⁶²

16-48 The Formation of Episulfides and Episulfones¹⁴⁶³

Epoxides can be converted directly to episulfides by treatment with NH₄SCN and ceric ammonium nitrate.¹⁴⁶⁴ Diazoalkanes, treated with sulfur, give episulfides.¹⁴⁶⁵ It is likely that R₂C=S is an intermediate, which is attacked by another molecule of diazoalkane, in a process similar to that shown in Reaction 16-46. Thioketones do react with diazoalkanes to give episulfides,¹⁴⁶⁶ and have also been converted to episulfides with sulfur ylids.¹⁴⁴² Carbenes (e.g., the dichlorocarbene from CHCl₃) and base react with thioketones to give an α,α-dichloro episufide.¹⁴⁶⁷

Alkanesulfonyl chlorides, when treated with diazomethane in the presence of a base (usually a tertiary amine), give episulfones (85).¹⁴⁶⁸ The base removes HCl from the sulfonyl halide to produce the highly reactive sulfene (84) (see Reaction 17-14), which then adds CH₂. The episulfone can then be heated to give off SO₂ (see Reaction 17-20), making the entire process a method for achieving the conversion RCH₂SO₂Cl → RCH=CH₂.¹⁴⁶⁹

OS V, 231, 877.

16-49 Cyclopropanation of Conjugated Carbonyl Compounds

Double-bond compounds that undergo the Michael reaction (15-24) can be converted to cyclopropane derivatives with sulfur ylids.¹⁴⁷⁰ Among the most common of these is dimethyloxosulfonium methylid (78),¹⁴⁷¹

which is widely used to transfer CH₂ to activated double bonds, but other sulfur ylids have also been used. A combination of DMSO and KOH in an ionic liquid converts conjugated ketones to α,β-cyclopropyl ketones.¹⁴⁷² Both CHR and CR₂ can be added in a similar manner with certain nitrogen-containing compounds. For example, ylids¹⁴⁷³ (e.g., 86), add various groups to activated double bonds.¹⁴⁷⁴ Sulfur ylids react with allylic alcohols in the presence of MnO₂ and molecular sieve 4 Å to give the cyclopropyl aldehyde.¹⁴⁷⁵ Similar reactions have been performed with phosphorus ylids,¹⁴⁷⁶ with pyridinium ylids,¹⁴⁷⁷ and with the compounds (PhS)₃CLi and Me₃Si(PhS)₂CLi.¹⁴⁷⁸ The reactions with ylids such as these of course involve nucleophilic acyl addition. Enantioselective cyclopropanation occurs in the presence of certain organocatalysts¹⁴⁷⁹ or chiral metal catalysts.¹⁴⁸⁰

Other reagents can be used to convert an aldehyde or ketone to a cyclopropane derivative. Tellurium ylids react with conjugated imines to the cyclopropyl aldehyde after hydrolysis of the imine.¹⁴⁸¹ Conjugated ketones react with Cp₂Zr(CH₂=CH₂) and PMe₃ to give a vinyl cyclopropane derivative after treatment with aq H₂SO₄.¹⁴⁸²

16-50 The Thorpe Reaction

N-Hydro-C-(α-cyanoalkyl)-addition

In the Thorpe reaction, the α carbon of one nitrile molecule is added to the CN carbon of another, so this reaction has analogies with the aldol reaction (16-34). The C=NH bond is, of course, hydrolyzable (Reaction 16-2), so β-keto nitriles can be prepared in this manner. The Thorpe reaction can be done intramolecularly, in which case it

is called the Thorpe–Ziegler reaction.¹⁴⁸³ As with other cyclization methods, yields are high for 5–8-membered rings, fall off to about zero for rings of 9–13 members, but are high again for 14-membered and larger rings, if high-dilution techniques are employed. The product in the Thorpe–Ziegler reaction is not the imine, but the tautomeric enamine (e.g., 87); if desired this can be hydrolyzed to an α-cyano ketone (Reaction 16-2), which can in turn be hydrolyzed and decarboxylated (Reactions 16-4 and 12-40). Other active-hydrogen compounds can also be added to nitriles.¹⁴⁸⁴

OS VI, 932.

H Other Carbon or Silicon Nucleophiles

16-51 Addition of Silanes

O-Hydro-C-alkyl-addition

Allylic silanes react with aldehydes, in the presence of Lewis acids, to give a homoallylic alcohol.¹⁴⁸⁵ In the case of benzylic silanes, this addition reaction has been induced with Mg(ClO₄)₂ under photochemical conditions.¹⁴⁸⁶Cyclopropylcarbinyl silanes add to acetals in the presence of TMSOTf to give a homoallylic alcohol.¹⁴⁸⁷ Allyltrichlorosilane adds an allyl group to an aldehyde in the presence of a cyclic urea and AgOTf.¹⁴⁸⁸ In a related reaction, allylic silanes react with acyl halides to produce the corresponding carbonyl derivative. The reaction of phenyl chloroformate, allyltrimethylsilane, and AlCl₃, for example, gave phenyl but-3-enoate.¹⁴⁸⁹ The use of chiral additives leads to the alcohol with good asymmetric induction.¹⁴⁹⁰

Allylic silanes also add to imines, in the presence of TiCl₄, to give amines.¹⁴⁹¹ Silanes also add to iminium salts, mediated by alkyltin compounds.¹⁴⁹²

16-52 The Formation of Cyanohydrins

O-Hydro-C-cyano-addition

The addition of HCN to aldehydes or ketones produces cyanohydrins.¹⁴⁹³ This is an equilibrium reaction, and for aldehydes and aliphatic ketones the equilibrium lies to the right; therefore the reaction is quite feasible, except with sterically hindered ketones (e.g., diisopropyl ketone). However, ketones (ArCOR) give poor yields, and the reaction cannot be carried out with ArCOAr since the equilibrium lies too far to the left. With aromatic aldehydes the benzoin condensation (Reaction 16-55) competes. With α,β-unsaturated aldehydes and ketones, 1,4-addition competes (Reaction 15-38).

Ketones of low reactivity (e.g., ArCOR) can be converted to cyanohydrins by treatment with diethylaluminum cyanide (Et₂AlCN) (see OS VI, 307) or, indirectly, with cyanotrimethylsilane (Me₃SiCN)¹⁴⁹⁴ in the presence of a Lewis acid or base,¹⁴⁹⁵ followed by hydrolysis of the resulting O-trimethylsilyl cyanohydrin (88). Both direct formation of the cyanohydrin (hydrocyanation) and formation of the cyano-O-silyl ether have been carried out enantioselectively using chiral catalysts,¹⁴⁹⁶ including chiral organocatalysts,¹⁴⁹⁷ or chiral additives.¹⁴⁹⁸ Biocatalysts have been used.¹⁴⁹⁹ Hydrogen cyanide adds to aldehydes in the presence of a lyase to give the cyanohydrin with good enantioselectivity.¹⁵⁰⁰ Cyanohydrins have been formed using a lyase in an ionic liquid.¹⁵⁰¹

Solvent-free conditions have been reported using TMSCN, an aldehyde, and potassium carbonate.¹⁵⁰² Amine N-oxides catalyze the reaction,¹⁵⁰³ as does tetrabutylammonium cyanide.¹⁵⁰⁴ Lithium perchlorate in ether facilitates this reaction,¹⁵⁰⁵ and LiCl catalyzes the reaction with Me₃SiCN.¹⁵⁰⁶N-Heterocyclic carbenes catalyze the reaction,¹⁵⁰⁷ as do certain ionic liquids.¹⁵⁰⁸ With MgBr₂ as a catalyst, the reaction proceeds with good syn selectivity.¹⁵⁰⁹ Other useful catalysts include Pt¹⁵¹⁰ Au,¹⁵¹¹ or Ti¹⁵¹² compounds, and InBr₃.¹⁵¹³ The use of chiral additives leads to cyanohydrins with good asymmetric induction.¹⁵¹⁴ Chiral transition metal catalysts have been used to give O-trialkylsilyl cyanohydrins with good enantioselectivity.¹⁵¹⁵ A venedium catalyst has been used in an ionic liquid.¹⁵¹⁶ Note that the reaction of an aldehyde and TMSCN in the presence of aniline and a BiCl₃ catalyst leads to an α-cyano amine.¹⁵¹⁷Potassium cyanide and acetic anhydride react with an aldehyde in the presence of a chiral Ti catalyst to give an α-acetoxy nitrile.¹⁵¹⁸

Rather than direct reaction with an aldehyde or ketone, the bisulfite addition product is often treated with cyanide. The addition is nucleophilic and the actual nucleophile is ⁻CN, so the reaction rate is increased by the addition of base.¹⁵¹⁹ This was demonstrated by Lapworth in 1903, and consequently this was one of the first organic mechanisms to be known.¹⁵²⁰ This method is especially useful for aromatic aldehydes, since it avoids competition from the benzoin condensation. If desired, it is possible to hydrolyze the cyanohydrin in situ to the corresponding α-hydroxy acid. This reaction is important in the Kiliani–Fischer method of extending the carbon chain of a sugar.

A particularly useful variation of this reaction uses cyanide rather than HCN. α-Amino nitriles¹⁵²¹ can be prepared in one step by the treatment of an aldehyde or ketone with NaCN and NH₄Cl. This is called the Strecker synthesis;¹⁵²² and it is a special case of the Mannich reaction (16-19). Since the CN group is easily hydrolyzed to the acid, this is a convenient method for the preparation of α-amino acids. The reaction has also been carried out with NH₃ + HCN and with NH₄CN. Salts of primary and secondary amines can be used instead of NH₄⁺ to obtain N-substituted and N,N-disubstituted α-amino nitriles. Brnsted acids can also be used.¹⁵²³ Unlike Reaction 16-52, the Strecker synthesis is useful for aromatic as well as aliphatic ketones. As in Reaction 16-52, the Me₃SiCN method has been used; 76 is converted to the product with ammonia or an amine.¹⁵²⁴ The effect of pressure on the Strecker synthesis has been studied.¹⁵²⁵ A catalyst-free multi-component Strecker reaction is known.¹⁵²⁶ There is also an In mediated Strecker reaction in aq media.¹⁵²⁷ Enantioselective Strecker syntheses are possible using chiral ammonium salts¹⁵²⁸ and other organocatalysts,¹⁵²⁹ chiral acids,¹⁵³⁰ or chiral metal complexes.¹⁵³¹ There is a radical version of the Strecker synthesis.¹⁵³²

OS I, 336; II, 7, 29, 387; III, 436; IV, 58, 506; VI, 307; VII, 20, 381, 517, 521. For the reverse reaction, see OS III, 101. For the Strecker synthesis, see OS I, 21, 355; III, 66, 84, 88, 275; IV, 274; V, 437; VI, 334.

16-53 The Addition of HCN to C=N and CN Bonds

N-Hydro-C-cyano-addition

Hydrogen cyanide adds to imines, Schiff bases, hydrazones, oximes, and similar compounds. Cyanide can be added to iminium ions to give α-cyano amines (89). As in Reaction 16-50, the addition to imines has been carried out enantioselectively.¹⁵³³ Chiral ammonium salts have been used with HCN.¹⁵³⁴ Trimethylsilyl cyanide (TMSCN) reacts with N-tosyl

imines in the presence of BF₃·OEt₂ to give the α-cyano N-tosyl amine.¹⁵³⁵ In the presence of a chiral Zr¹⁵³⁶ or Al¹⁵³⁷ catalyst, Bu₃SnCN reacts with imines to give α-cyanoamines enantioselectively. The imine can be formed in situ by reaction of an aldehyde or ketone with an amine, in the present of TMSCN and a suitable promoter.¹⁵³⁸ The reaction of an imine and TMSCN gives the cyano amine with good enantioselectivity using a chiral Sc catalyst.¹⁵³⁹Titanium catalysts have been used in the presence of a chiral Schiff base.¹⁵⁴⁰

The addition of KCN to triisopropylbenzenesulfonyl hydrazones (90) provides an indirect method for achieving the conversion RR′CO → RR′CHCN.¹⁵⁴¹ The reaction is successful for hydrazones of aliphatic aldehydes and ketones.

Hydrogen cyanide can also be added to the CN bond to give iminonitriles or α-aminomalononitriles.¹⁵⁴²

The acylcyanation of imines is known, and enantioselectivity is achieved with a suitable catalyst.¹⁵⁴³

OS V, 344. See also, OS V, 269.

16-54 The Prins Reaction

The addition of an alkene to formaldehyde in the presence of an acid¹⁵⁴⁴ catalyst is called the Prins reaction.¹⁵⁴⁵ Three main products are possible; which one predominates depends on the alkene and the conditions. When the product is the 1,3-diol or the dioxane,¹⁵⁴⁶ the reaction involves addition to the C=C as well as to the C=O. The mechanism is one of electrophilic attack on both double bonds. The acid first protonates the C=O, and the resulting carbocation is attacked by the C=C to give 91. The cation product 91 can

undergo loss of H⁺ to give the alkene or add water to give the diol.¹⁵⁴⁷ It has been proposed that 91 is stabilized by neighboring-group attraction, with either the oxygen¹⁵⁴⁸ or a carbon¹⁵⁴⁹ stabilizing the charge (92 and 94, respectively). This stabilization is postulated to explain the fact that with 2-butenes¹⁵⁵⁰ and with cyclohexenes the addition is anti. A backside attack of H₂O on the three- or four-membered ring would account for it. Other products are obtained too, which can be explained on the basis of 92 or 93.^1548,1549 Additional evidence for the intermediacy of 92 is the finding that oxetanes (94) subjected to the reaction conditions, which would protonate 94 to give 92, give essentially the same product ratios as the corresponding alkenes.¹⁵⁵¹ An argument against the intermediacy of 92 and 93 is that not all alkenes show the anti-stereoselectivity mentioned above. Indeed, the stereochemical results are often quite complex, with syn, anti, and nonstereoselective addition reported, depending on the nature of the reactants and the reaction conditions.¹⁵⁵² Since addition to the C=C bond is electrophilic, the reactivity of the alkene increases with alkyl substitution and Markovnikov's rule is followed. The dioxane product may arise from a reaction between the 1,3-diol and formaldehyde¹⁵⁵³ (16-5) or between 92 and formaldehyde. Racemization may occur in the Prins cyclization reaction by 2-oxonia-Cope rearrangements (see 18-32) by way of a (Z)-oxocarbenium ion intermediate.¹⁵⁵⁴

Iodine can promote the Prins reaction.¹⁵⁵⁵ Lewis acids (e.g., SnCl₄) also catalyze the reaction, in which case the species that adds to the alkenes is H₂C⁺–O–SnCl₄.¹⁵⁵⁶ The reaction can also be catalyzed by peroxides, in which case the mechanism is probably a free radical one. Other transition metal complexes can be used to form homoallylic alcohols. A typical example is the reaction of methylenecyclohexane with an aryl aldehyde to give 95.¹⁵⁵⁷

Samarium iodide promotes this addition reaction.¹⁵⁵⁸ In a related reaction, simple alkene units add to esters in the presence of sodium and liquid ammonia to give an alcohol.¹⁵⁵⁹ Dienes react with alcohols in the presence of a transition metal compound, to give alkenyl alcohols.¹⁵⁶⁰ Allenes also add to aldehydes.¹⁵⁶¹ Enynes undergo Prins cyclization with Au catalysts.¹⁵⁶²

Structural variations in the alkene lead to different products. Homoallylic alcohols react with aldehydes in the presence of Montmorillonite-KSF clay to give 4-hydroxytetrahydropyrans.¹⁵⁶³ A variation of this reaction converts an aryl aldehyde and a homoallylic alcohol to a 4-chlorotetrahydropyran in the presence of InCl₃.¹⁵⁶⁴ Homoallylic alcohols, protected as –O(CHMeOAc) react with BF₃·OEt₂ and acetic acid to give 4-acetoxytetrahydropyrans or with SnBr₄ to give 4-bromotetrahydropyrans.¹⁵⁶⁵ Homoallylic alcohols with a vinyl silane moiety react with InCl₃ and an aldehyde to give a dihydropyran.¹⁵⁶⁶

A closely related reaction has been performed with activated aldehydes or ketones; without a catalyst (e.g., chloral and acetoacetic ester), but with heat.¹⁵⁶⁷ The product in these cases is a β-hydroxy alkene, and the mechanism is pericyclic:¹⁵⁶⁸

This reaction is reversible and suitable β-hydroxy alkenes can be cleaved by heat (17-32). There is evidence that the cleavage reaction occurs by a cyclic mechanism (see 17-32), and, by the principle of microscopic reversibility, the addition mechanism should be cyclic too.¹⁵⁶⁹ Note that this reaction is an oxygen analogue of the ene synthesis (15-23). This reaction can also be done with unactivated aldehydes¹⁵⁷⁰ and ketones¹⁵⁷¹ if Lewis acid catalysts [e.g., dimethylaluminum chloride (Me₂AlCl) or ethylaluminum dichloride (EtAlCl₂)] are used.¹⁵⁷² Lewis acid catalysts also increase rates with activated aldehydes.¹⁵⁷³ The use of optically active catalysts has given optically active products with a high % ee.¹⁵⁷⁴

In a related reaction, alkenes can be added to aldehydes and ketones to give reduced alcohols (96). This has been accomplished by several methods,¹⁵⁷⁵ including treatment with SmI₂¹⁵⁷⁶ or Zn and Me₃SiCl,¹⁵⁷⁷ and by electrochemical¹⁵⁷⁸ and photochemical¹⁵⁷⁹ methods. Most of these methods have been used for intramolecular addition and most or all involve free radical intermediates.

There is an aza-Prins reaction, promoted by TiI₄ and I₂.¹⁵⁸⁰

OS IV, 786. See also, OS VII, 102.

16-55 The Benzoin Condensation

Benzoin aldehyde condensation

When certain aldehydes are treated with cyanide ion, benzoins (97) are produced in a reaction called the benzoin condensation. The condensation can be regarded as involving the addition of one molecule of aldehyde to the C=O group of another. The reaction only occurs with aromatic aldehydes, but not all of them,¹⁵⁸¹ and for glyoxals (RCOCHO). The two molecules of aldehyde obviously perform different functions. The one that no longer has a C–H bond in the product is called the donor, because it has “donated” its hydrogen to the oxygen of the other molecule, the acceptor. Some aldehydes can perform only one of these functions, and hence cannot be self-condensed, though they can often be condensed with a different aldehyde. For example, p-dimethylaminobenzaldehyde is not an acceptor, but only a donor. Thus it cannot condense with itself. It can condense with benzaldehyde, which can perform both functions, but is a better acceptor than it is a donor. N-alkyl-3-methylimidazolium salts catalyze the reaction,¹⁵⁸² as does an imidazole-based solid supported catalyst.¹⁵⁸³

The following is the accepted mechanism¹⁵⁸⁴ for this reversible reaction, which was originally proposed by Lapworth in 1903:¹⁵⁸⁵

The key step, the loss of the aldehyde proton, can take place because the acidity of this C–H bond is increased by the electron-withdrawing power of the CN group. Thus, cyanide is a highly specific catalyst for this reaction, because, almost uniquely, it can perform three functions: (1) It acts as a nucleophile; (2) its electron-withdrawing ability permits loss of the aldehyde proton; and (3) having done this, it then acts as a leaving group. Certain thiazolium salts can also catalyze the reaction.¹⁵⁸⁶ In this case, aliphatic aldehydes can also be used¹⁵⁸⁷ (the products are called acyloins), and mixtures of aliphatic and aromatic aldehydes give mixed α-hydroxy ketones.¹⁵⁸⁸ The reaction has also been carried out without cyanide, by using the benzoylated cyanohydrin as one of the components in a phase-transfer catalyzed process. By this means, products can be obtained from aldehydes that normally fail to self-condense.¹⁵⁸⁹ The condensation has also been done with excellent enantioselectivity using benzoyl formate decarboxylase.¹⁵⁹⁰ Enantiopure triazolium salts have been evaluated as catalysts in the enantioselective benzoin condensation.¹⁵⁹¹N-Heterocyclic carbene catalysts have been used for asymmetric induction.¹⁵⁹²

A “mixed”-benzoin condensation has been accomplished by using aryl silyl ketones [ArC(=O)SiMe₂Ph] and aldehydes with a La catalyst.¹⁵⁹³ The reaction of acylsilanes and aldehydes, catalyzed by metal cyanides, is known as the silyl-benzoin reaction.¹⁵⁹⁴

OS I, 94; VII, 95.

16-56 Addition of Radicals to C=O, C=S, C=N Compounds

Radical cyclization is not limited to reaction with a C=C unit (see 15-29 and 15-30), and reactions with both C=N and C=O moieties are known. Reaction of MeON=CH(CH₂)₃CHO with Bu₃SnH and AIBN, for example, led to trans-2-(methoxyamino)cyclopentanol in good yield.¹⁵⁹⁵ Conjugated ketones add to aldehyde via the β carbon under radical conditions (2 molar equivalents of Bu₃SnH and 0.1 equiv of CuCl) to give a β-hydroxy ketone.¹⁵⁹⁶

In a related reaction, diazo compounds add to ketenes to give allenes.¹⁵⁹⁷

Addition of radical to the C=N unit of R–C=N–SPh¹⁵⁹⁸ or R–C=N–OBz¹⁵⁹⁹ led to cyclic imines. Radical addition to simple imines leads to aminocycloalkenes.¹⁶⁰⁰ Radicals also add to the carbonyl unit of phenylthio esters to give cyclic ketones.¹⁶⁰¹ Carbon-centered radicals add to imines.¹⁶⁰² The reaction of an alkyl halide with BEt₃ in aq methanol, for example, gives the imine addition product, an alkylated amine.¹⁶⁰³

Secondary alkyl iodides add to O-alkyl oximes in the presence of BEt₃ and AIBN. This methodology was used to convert MeO₂C-CH=NOBn to MeO₂C–CH(R)NOBn.¹⁶⁰⁴ Benzylic halides add to imines under photochemical conditions, and in the presence of 1-benzyl-1,4-dihydronicotinamide¹⁶⁰⁵ or with BEt₃ in aq methanol.¹⁶⁰⁶ Tertiary alkyl iodides add to oxime ethers using BF₃·OEt₂ in the presence of BEt₃/O₂.¹⁶⁰⁷ O-Trityl oximes of 5- and 6-iodoaldehydes undergo radical cyclization to give oximes¹⁶⁰⁸ (also see Reaction 15-30). Enantioselective radical addition reactions to N-benzoyl hydrazones used chiral ammonium salts.¹⁶⁰⁹ The Mn mediated reaction of allyl iodides with chiral N-acylhydrazones leads to chiral hydrazine derivatives.¹⁶¹⁰

N,N-Dimethylaniline reacts with aldehydes under photochemical conditions to give acyl addition via the carbon atom of one of the methyl groups.¹⁶¹¹ The reaction of PhNMe₂ and benzaldehyde, for example, gave PhN(Me)CH₂CH(OH)Ph upon photolysis.

16.B.ii. Acyl Substitution Reactions

A. O,N, and S Nucleophiles

16-57 Hydrolysis of Acyl Halides

Hydroxy-de-halogenation

Acyl halides are so reactive that hydrolysis is easily carried out.¹⁶¹² In fact, most simple acyl halides must be stored under anhydrous conditions or they may react with water in the air. Consequently, water is usually a strong enough nucleophile for the reaction, but in unreactive systems hydroxide ion may be required. The reaction is seldom synthetically useful, because acyl halides are normally prepared from acids. The reactivity order is F < Cl < Br < I.¹⁶¹³ If a carboxylic acid is used as the nucleophile, an exchange may take place (see Reaction 16-79). The mechanism¹⁶¹³ of hydrolysis can be either S_N1 or tetrahedral, the former occurring in highly polar solvents and in the absence of strong nucleophiles.¹⁶¹⁴ There is also evidence for the S_N2 mechanism in some cases.¹⁶¹⁵

Hydrolysis of acyl halides is not usually catalyzed by acids, except for acyl fluorides, where hydrogen bonding can assist in the removal of F.¹⁶¹⁶ There are several methods available for the hydrolysis of acyl fluorides.¹⁶¹⁷

OS II, 74.

16-58 Hydrolysis of Anhydrides

Hydroxy-de-acyloxy-substitution

Anhydrides are somewhat more difficult to hydrolyze than acyl halides, but here too water is usually a strong enough nucleophile. The mechanism is usually tetrahedral.¹⁶¹⁸ The S_N1 mechanism only occurs with acid catalysis and seldom even then.¹⁶¹⁹ Anhydride hydrolysis can also be catalyzed by bases. Of course, hydroxide ion attacks more readily than water, but other bases can also catalyze the reaction. This phenomenon, called nucleophilic catalysis(Sec. 16.A.i, category 4), is actually the result of two successive tetrahedral mechanisms. For example, pyridine catalyzes the hydrolysis of acetic anhydride in this manner.¹⁶²⁰

Many other nucleophiles similarly catalyze the reaction.

OS I, 408; II, 140, 368, 382; IV, 766; V, 8, 813.

16-59 Hydrolysis of Carboxylic Esters

Hydroxy-de-alkoxylation

Ester hydrolysis is usually catalyzed by acids or bases. Since OR is a much poorer leaving group than halide or OCOR, water alone does not hydrolyze most esters. When bases catalyze the reaction, the attacking species is the more powerful nucleophile ⁻OH. This reaction is called saponification and gives the salt of the acid. Acids catalyze the reaction by making the carbonyl carbon more positive, and therefore more susceptible to attack by the nucleophile. Both reactions are equilibrium reactions, so there must be a way to shift the equilibrium to the right for this to be useful. Since formation of the salt does just this, ester hydrolysis is almost always done for preparative purposes in basic solution, unless the compound is base sensitive. Even in the case of 98, however, selective base hydrolysis of the ethyl ester gave an 80% yield of the acid–dimethyl ester (99).¹⁶²¹

Ester hydrolysis can also be catalyzed¹⁶²² by metal ions, by cyclodextrins,¹⁶²³ by enzymes,¹⁶²⁴ and by nucleophiles.¹⁶²¹ Other reagents used to cleave carboxylic esters include Dowex-50,¹⁶²⁵ Me₃SiI,¹⁶²⁶ and InCl₃ on moist silica gel using microwave irradiation.¹⁶²⁷ Cleavage of phenolic esters is usually faster than carboxylic esters derived from aliphatic acids. The reagent Sm/I₂ at -78 °C has been used,¹⁶²⁸ ammonium acetate in aq methanol,¹⁶²⁹ Amberlyst 15 in methanol,¹⁶³⁰ and phenolic esters have been selectively hydrolyzed in the presence of alkyl esters on alumina with microwave irradiation.¹⁶³¹ Allylic esters were cleaved with DMSO–I₂.¹⁶³² Thiophenol with K₂CO₃ in NMP quantitatively converted methyl benzoate to benzoic acid.¹⁶³³ Allylic esters were cleaved with 2% Me₃SiOTf in dichloromethane,¹⁶³⁴ with CeCl₃·7 H₂O–NaI,¹⁶³⁵ and with NaHSO₄·silica gel.¹⁶³⁶ Lactones also undergo the reaction¹⁶³⁷ (but if the lactone is five- or six-membered, the hydroxy acid often spontaneously re-forms the lactone) and thiol esters (RCOSR′) give thiols R′SH. Typical reagents for this latter transformation include NaSMe in methanol,¹⁶³⁸ borohydride exchange resin–Pd(OAc)₂ for reductive cleavage of thiol esters to thiols,¹⁶³⁹ and TiCl₄/Zn for the conversion of phenylthioacetates to thiophenols.¹⁶⁴⁰ Sterically hindered esters are hydrolyzed with difficulty (Sec. 10.G.i), but reaction of 2 equiv of t-BuOK with 1 equiv of water is effective.¹⁶⁴¹ Hindered esters can also be cleaved by sequential treatment with zinc bromide and then water,¹⁶⁴² with silica gel in refluxing toluene,¹⁶⁴³ and on alumina when irradiated with microwaves.¹⁶⁴⁴ For esters insoluble in water the rate of two-phase ester saponification can be greatly increased by the application of ultrasound,¹⁶⁴⁵ and phase-transfer techniques have been applied.¹⁶⁴⁶

Enzymatic hydrolysis of diesters with esterase has been shown to give the hydroxy-ester,¹⁶⁴⁷ and selective hydrolysis of dimethyl succinate to monomethyl succinic acid was accomplished with aq NaOH in THF.¹⁶⁴⁸ Hydrolysis of vinyl esters leads to ketones, and the reaction of C-substituted vinyl acetates with an esterase derived from Marchantia polymorpha gave substituted ketones with high enantioselectivity.¹⁶⁴⁹ Scandium triflate was shown to hydrolyze α-acetoxy ketones to α-hydroxy ketones.¹⁶⁵⁰

Ingold¹⁶⁵¹ has classified the acid- and base-catalyzed hydrolyses of esters (and the formation of esters, since these are reversible reactions and thus have the same mechanisms) into eight possible mechanisms (Table 16.3),^1651,1652depending on the following criteria: (1) acid- or base catalyzed, (2) unimolecular or bimolecular, and (3) acyl cleavage or alkyl cleavage.¹⁶⁵³ All eight of these are S_N1, S_N2, or tetrahedral mechanisms. The acid-catalyzed mechanisms are shown with reversible arrows. They are not only reversible, but also symmetrical; that is, the mechanisms for ester formation are exactly the same as for hydrolysis, except that H replaces R. Internal proton transfers, such as shown for B and C, may not actually be direct but may take place through the solvent. There is much physical evidence to show that esters are initially protonated on the carbonyl and not on the alkyl oxygen (Chap 8, Ref. 17). Nevertheless the A_AC1 mechanism is shown as proceeding through the ether-protonated intermediate A, since it is difficult to envision OR′ as a leaving group here. It is of course possible for a reaction to proceed through an intermediate even if only a tiny concentration is present. The designations A_AC1, and so on, are those of Ingold. The A_AC2 and A_AC1 mechanisms are also called A2 and A1, respectively. Note that the A_AC1 mechanism is actually the same as the S_N1cA mechanism for this type of substrate and that A_AL2 is analogous to S_N2cA. Some authors use A1 and A2 to refer to all types of nucleophilic substitution in which the leaving group first acquires a proton. The base-catalyzed reactions are not shown with reversible arrows, since they are reversible only in theory and not in practice. Hydrolyses taking place under neutral conditions are classified as B mechanisms. Molecular dynamics has shown that “the rate of hydrolysis of methyl formate in pure water is consistent with mechanisms involving cooperative catalysis by autoionization-generated hydroxide and hydronium, a process known to have an activation free energy of 23.8 kcal mol^–1 (99.6 kJ mol^–1).”¹⁶⁵⁴

Table 16.3 Classification of the Eight Mechanisms for Ester Hydrolysis and Formation.

Name

Ingold

IUPAC^a

Type

A_AC1

A_h+D_N+A_N+D_h

S_N1

A_AC2

A_h+A_N+A_hD_h+D_h

Tetrahedral

A_AL1

A_h+D_N+A_N+D_h

S_N1

A_AL2

A_h+A_ND_N+D_h

S_N2

B_AC1

D_N+A_N+A_xhD_h

S_N1

B_AC2

A_N+D_N+A_xhD_h

Tetrahedral

B_AL1

D_N+A_N+A_xhD_h

S_N1

B_AL2

A_ND_N

S_N2

Adapted material from Structure and Mechanism in Organic Chemistry, 2d ed., Cornell University Press, Ithaca, NY, 1969, pp. 1129–1131, edited by Ingold, C.K. Copyright ©1969 by Cornell University. Used by permission of the publisher, Cornell University Press.

a. See Ref. 1652

Of the eight mechanisms, seven have actually been observed in the hydrolysis of carboxylic esters. The one that has not been observed is the B_AC1 mechanism.¹⁶⁵⁵ The most common mechanisms are the B_AC2 for basic catalysis and the A_AC2¹⁶⁵⁶ for acid catalysis, that is, the two tetrahedral mechanisms. Both involve acyl–oxygen cleavage. The evidence is (1) hydrolysis with H₂¹⁸O results in the ¹⁸O appearing in the acid and not in the alcohol;¹⁶⁵⁷ (2) esters with chiral R′ groups give alcohols with retention of configuration;¹⁶⁵⁸ (3) allylic R′ gives no allylic rearrangement;¹⁶⁵⁹ (4) neopentyl R′ gives no rearrangement¹⁶⁶⁰; all these facts indicate that the O–R′ bond is not broken. It has been concluded that two molecules of water are required in the A_AC2 mechanism, as shown:

If this were so, the protonated derivatives B and C would not appear at all. This conclusion stems from a value of w (see Sec. 8.C) of ~ 5, indicating that water acts as a proton donor here as well as a nucleophile.¹⁶⁶¹ Termolecular processes are rare, but in this case the two water molecules are already connected by a hydrogen bond. (A similar mechanism, called B_AC3, also involving two molecules of water, has been found for esters that hydrolyze without a catalyst.¹⁶⁶² Such esters are mostly those containing halogen atoms in the R group.)

The other mechanism involving acyl cleavage is the A_AC1 mechanism. This is rare, being found only where R is very bulky, so that bimolecular attack is sterically hindered, and only in ionizing solvents. The mechanism has been demonstrated for esters of 2,4,6-trimethylbenzoic acid (mesitoic acid). This acid depresses the freezing point of sulfuric acid four times as much as would be predicted from its molecular weight, which is evidence for the equilibrium

In a comparable solution of benzoic acid, the freezing point is depressed only twice the predicted amount, indicating only a normal acid–base reaction. Further, a sulfuric acid solution of methyl mesitoate when poured into water gave mesitoic acid, while a similar solution of methyl benzoate similarly treated did not.¹⁶⁶³ The A_AC1 mechanism is also found when acetates of phenols or of primary alcohols are hydrolyzed in concentrated (>90%) H₂SO₄ (the mechanism under the more usual dilute acid conditions is the normal A_AC2).¹⁶⁶⁴

The mechanisms involving alkyl–oxygen cleavage are ordinary S_N1 and S_N2 mechanisms in which OCOR (an acyloxy group) or its conjugate acid is the leaving group. Two of the three mechanisms, the B_AL1 and A_AL1 mechanisms, occur most readily when R′ comes off as a stable carbocation (i.e., when R′ is tertiary alkyl, allylic, benzylic, etc.). For acid catalysis, most esters with this type of alkyl group (especially tertiary alkyl) cleave by this mechanism, but even for these substrates, the B_AL1 mechanism occurs only in neutral or weakly basic solution, where the rate of attack by hydroxide is so slowed that the normally slow (by comparison) unimolecular cleavage takes over. These two mechanisms have been established by kinetic studies, ¹⁸O labeling, and isomerization of R′.¹⁶⁶⁵ Secondary and benzylic acetates hydrolyze by the A_AC2 mechanism in dilute H₂SO₄, but in concentrated acid the mechanism changes to A_AL1.¹⁶⁶⁵ Despite its designation, the B_AL1 mechanism is actually uncatalyzed (as is the unknown B_AC1 mechanism).

The two remaining mechanisms, B_AL2 and A_AL2, are very rare. The B_AL2 mechanism because it requires hydroxide ion to attack an alkyl carbon when an acyl carbon is also available,¹⁶⁶⁶ and the A_AL2 because it requires water to be a nucleophile in an S_N2 process. Both have been observed, however. The B_AL2 has been seen in the hydrolysis of β-lactones under neutral conditions¹⁶⁶⁷ (because cleavage of the C–O bond in the transition state opens the four-membered ring and relieves strain), the alkaline hydrolysis of methyl 2,4,6-tri-tert-butyl benzoate,¹⁶⁶⁸ and in the unusual reaction:¹⁶⁶⁹

When it does occur, the B_AL2 mechanism is easy to detect, since it is the only one of the base-catalyzed mechanisms that requires inversion at R′. However, in the last example given, the mechanism is evident from the nature of the product, since the ether could have been formed in no other way. The A_AL2 mechanism has been reported in the acid cleavage of γ-lactones.¹⁶⁷⁰

To sum up the acid-catalysis mechanisms, A_AC2 and A_AL1 are the common mechanisms, the latter for R′ that give stable carbocations, the former for practically all the rest. The A_AC1 mechanism is rare, being found mostly with strong acids and sterically hindered R. The A_AL2 mechanism is even rarer. For basic catalysis, B_AC2 is almost universal; B_AL1 occurs only with R′ that give stable carbocations and then only in weakly basic or neutral solutions; B_AL2 is very rare; and B_AC1 has never been observed.

The above results pertain to reactions in solution. In the gas phase,¹⁶⁷¹ reactions can take a different course, as illustrated by the reaction of carboxylic esters with MeO⁻, which in the gas phase was shown to take place only by the B_AL2 mechanism,¹⁶⁷² even with aryl esters,¹⁶⁷³ where this means that an S_N2 mechanism takes place at an aryl substrate. However, when the gas-phase reaction of aryl esters was carried out with MeO⁻ ions, each of which was solvated with a single molecule of MeOH or H₂O, the B_AC2 mechanism was observed.¹⁶⁷²

In the special case of alkaline hydrolysis of N-substituted aryl carbamates, there is another mechanism¹⁶⁷⁴ involving elimination–addition:¹⁶⁷⁵

This mechanism does not apply to unsubstituted or N,N-disubstituted aryl carbamates, which hydrolyze by the normal mechanisms. Carboxylic esters substituted in the α position by an electron-withdrawing group (e.g., CN or CO₂Et) can also hydrolyze by a similar mechanism involving a ketene intermediate.¹⁶⁷⁶ These elimination–addition mechanisms usually are referred to as E1cB mechanisms, because that is the name given to the elimination portion of the mechanism (Sec. 17.A.iii).

The acid-catalyzed hydrolysis of enol esters (RCOOCR′=CR) can take place either by the normal A_AC2 mechanism or by a mechanism involving initial protonation on the double-bond carbon, similar to the mechanism for the hydrolysis of enol ethers given in Reaction 10-6,¹⁶⁷⁷ depending on reaction conditions.¹⁶⁷⁸ In either case, the products are the carboxylic acid (RCO₂H) and the aldehyde or ketone (R–CHCOR′).

OS I, 351, 360, 366, 379, 391, 418, 523; II, 1, 5, 53, 93, 194, 214, 258, 299, 416, 422, 474, 531, 549; III, 3, 33, 101, 209, 213, 234, 267, 272, 281, 300, 495, 510, 526, 531, 615, 637, 652, 705, 737, 774, 785, 809 (but see OS V, 1050), 833, 835; IV, 15, 55, 169, 317, 417, 444, 532, 549, 555, 582, 590, 608, 616, 628, 630, 633, 635, 804; V, 8, 445, 509, 687, 762, 887, 985, 1031; VI, 75, 121, 560, 690, 824, 913, 1024; VII, 4, 190, 210, 297, 319, 323, 356, 411; VIII, 43, 141, 219, 247, 258, 263, 298, 486, 516, 527. Ester hydrolyses with concomitant decarboxylation are listed at Reaction 12-40.

16-60 Hydrolysis of Amides

Hydroxy-de-amination

Unsubstituted amides (RCONH₂) can be hydrolyzed with either acidic or basic catalysis, and the products are, respectively, the free acid and the ammonium ion or the salt of the acid and ammonia. N-Substituted (RCONHR′) and N,N-disubstituted (RCONR₂′) amides can be hydrolyzed analogously, and the product is the primary or secondary amine, respectively (or their salts), rather than ammonia. Twisting of the amide bond leads to an acceleration of water-promoted hydrolysis reactions.¹⁶⁷⁹ Lactams, imides, cyclic imides, hydrazides, and so on, also undergo the reaction.

Water alone is not sufficient to hydrolyze most amides,¹⁶⁸⁰ since NH₂ is even a poorer leaving group than OR.¹⁶⁸¹ Prolonged heating is often required, even with acidic or basic catalysts.¹⁶⁸² Treatment of primary amides with phthalic anhydride at 250 °C and 4 atm gives the carboxylic acid and phthalimide.¹⁶⁸³ Hydrolysis of carbamates (RNHCO₂R) to the corresponding amine can be categorized in this section. Although the product is an amine and the carboxyl unit fragments, this reaction is simply a variation of amide hydrolysis. Strong acids [e.g., trifluoroacetic acid (in dichloromethane)] are usually employed.¹⁶⁸⁴ Treatment of N-Boc derivatives (RNHCO₂t-Bu) with AlCl₃¹⁶⁸⁵or with aq sodium tert-butoxide¹⁶⁸⁶ gave the amine. The byproducts of this reaction are typically carbon dioxide and isobutylene.

In difficult cases, nitrous acid, NOCl, N₂O₄,¹⁶⁸⁷ or a similar compound can be used (unsubstituted amides only¹⁶⁸⁸). These reactions involve a diazonium ion (see Reaction 13-19) and are much faster than ordinary hydrolysis. The benzamide–nitrous acid reaction took place 2.5 × 10⁷ times faster than ordinary hydrolysis, for example.¹⁶⁸⁹ Another procedure for difficult cases involves treatment with aq sodium peroxide.¹⁶⁹⁰ In still another method, the amide is treated with water and t-BuOK at room temperature.¹⁶⁹¹ A kinetic study has been done on the alkaline hydrolyses of N-trifluoroacetyl aniline derivatives.¹⁶⁹² Amide hydrolysis can also be catalyzed by nucleophiles (see Sec. 16.A.i, category 4).

The same framework of eight possible mechanisms discussed for ester hydrolysis in Reaction 16-59 can also be applied to amide hydrolysis.¹⁶⁹³ Both the acid- and base-catalyzed hydrolyses are essentially irreversible, since salts are formed in both cases. For basic catalysis¹⁶⁹⁴ the mechanism is B_AC2. There is much evidence for this mechanism, similar to that discussed for ester hydrolysis. Molecular orbital studies on the mechanism of amide hydrolysis suggest a highly tetrahedral transition state.¹⁶⁹⁵ In certain cases, kinetic studies have shown that the reaction is second order in ⁻OH indicating that 100 can lose a proton to give 101.¹⁶⁹⁶ Depending on the nature of R′, 101 can cleave directly to give the two negative ions (path a) or become N-protonated prior to or during the act of cleavage (path b), in which case the products are obtained directly and a final proton transfer is not necessary.¹⁶⁹⁷ Studies of the effect, on the rate of hydrolysis and on the ratio k_-1/k₂, of substituents on the aromatic rings in a series of amides (CH₃CONHAr) led to the conclusion that path a is taken when Ar contains

electron-withdrawing substituents and path b when electron-donating groups are present.¹⁶⁹⁸ The presence of electron-withdrawing groups helps stabilize the negative charge on the nitrogen, so that NR₂^′− can be a leaving group (path a). Otherwise, the C–N bond does not cleave until the nitrogen is protonated (either prior to or in the act of cleavage), so that the leaving group, even in the base-catalyzed reaction, is not NR₂^′− but the conjugate NHR₂^′ (path b). It is known that formation of 100 is the rate-determining step in the B_AC2 mechanism, but only at high-base concentrations. At lower concentrations of base, the cleavage of 100 or 101 becomes rate determining.¹⁶⁹⁹

For acid catalysis, matters are less clear. The reaction is generally second order, and it is known that amides are primarily protonated on the oxygen. Because of these facts, it has been generally agreed that most acid-catalyzed amide hydrolysis takes place by the A_AC2 mechanism. Further evidence for this mechanism is that a small but detectable amount of ¹⁸O exchange (see Sec. 16.A.i, category 3) has been found in the acid-catalyzed hydrolysis of benzamide¹⁷⁰⁰ (¹⁸O exchange has also been detected for the base-catalyzed process,¹⁷⁰¹ in accord with the B_AC2 mechanism). Kinetic data have shown that three molecules of water are involved in the rate-determining step,¹⁷⁰²suggesting that, as in the A_AL2 mechanism for ester hydrolysis (Reaction 16-59), additional water molecules take part in a process, such as:

The four mechanisms involving alkyl–N cleavage (the AL mechanisms) do not apply to this reaction. They are not possible for unsubstituted amides, since the only N–C bond is the acyl bond. They are possible for N-substituted and N,N-disubstituted amides, but in these cases they give entirely different products and are not amide hydrolyses at all.

The reaction shown involves attack by the base on an N-alkyl group to give an alcohol. While rare, it has been observed for various N-tert-butylamides in 98% sulfuric acid, where the mechanism was A_AL1,¹⁷⁰³ and for certain amides containing an azo group, where a B_AL1 mechanism was postulated.¹⁷⁰⁴ Of the two first-order acyl cleavage mechanisms, only the A_AC1 has been observed, in concentrated sulfuric acid solutions.¹⁷⁰⁵ Of course, the diazotization of unsubstituted amides might be expected to follow this mechanism, and there is evidence that this is true.¹⁶⁸⁹

OS I, 14, 111, 194, 201, 286; II, 19, 25, 28, 49, 76, 208, 330, 374, 384, 457, 462, 491, 503, 519, 612; III, 66, 88, 154, 256, 410, 456, 586, 591, 661, 735, 768, 813; IV, 39, 42, 55, 58, 420, 441, 496, 664; V, 27, 96, 341, 471, 612, 627; VI, 56, 252, 507, 951, 967; VII, 4, 287; VIII, 26, 204, 241, 339, 451.

The oxidation of aldehydes to carboxylic acids can proceed by a nucleophilic mechanism, but more often it does not. The reaction is considered in Reaction 19-23. Basic cleavage of β-keto esters and the haloform reaction could be considered at this point, but they are also electrophilic substitutions and are treated in Reactions 12-43 and 12-44.

B. Attack by OR at an Acyl Carbon

16-61 Alcoholysis of Acyl Halides

Alkoxy-de-halogenation

The reaction between acyl halides and alcohols or phenols is the best general method for the preparation of carboxylic esters. It is believed to proceed by a S_N2 mechanism.¹⁷⁰⁶ As with Reaction 16-57, however, the mechanism can be S_N1 or tetrahedral.¹⁶¹³ Lewis acids (e.g., lithium perchlorate) can be used.¹⁷⁰⁷ The reaction is of wide scope, and many functional groups do not interfere. A base is frequently added to combine with the HX formed. When aq alkali is used, this is called the Schotten–Baumann procedure, but pyridine is also frequently used. Indeed, pyridine catalyzes the reaction by the nucleophilic catalysis route (see Reaction 16-58). Both R and R′ may be primary, secondary, or tertiary alkyl or aryl. Enol esters can also be prepared by this method, though C-acylation competes in these cases. In difficult cases, especially with hindered acids or tertiary R′, the alkoxide can be used instead of the alcohol.¹⁷⁰⁸ Activated alumina has also been used as a catalyst, for tertiary R′.¹⁷⁰⁹ Thallium salts of phenols give very high yields of phenolic esters,¹⁷¹⁰ and BiOCl is very effective for the preparation of phenolic acetates.¹⁷¹¹ Phase-transfer catalysis has been used for hindered phenols.¹⁷¹² Zinc has been used to couple alcohols and acyl chlorides,¹⁷¹³ Zr compounds were used,¹⁷¹⁴ and catalytic Cu(acac)₂ and benzoyl chloride was used to prepare the monobenzoate of ethylene glycol.¹⁷¹⁵ Selective acylation is possible in some cases.¹⁷¹⁶

Acyl halides react with thiols, in the presence of Zn, to give the corresponding thio-ester.¹⁷¹⁷ The reaction of acid chlorides or anhydrides (see 16-62) with diphenyldiselenide, in the presence of Sm/CoCl₂¹⁷¹⁸ or Sm/CrCl₃¹⁷¹⁹ gave the corresponding seleno ester (PhSeCOMe).

Acyl halides can also be converted to carboxylic acids by using ethers instead of alcohols, as shown, in MeCN in the presence of certain catalysts [e.g., cobalt(II) chloride].¹⁷²⁰ A variation of this reaction has been reported that uses acetic anhydride (also see 16-62).¹⁷²¹

This is a method for the cleavage of ethers (see also, Reaction 10-49).

OS I, 12; III, 142, 144, 167, 187, 623, 714; IV, 84, 263, 478, 479, 608, 616, 788; V, 1, 166, 168, 171; VI, 199, 259, 312, 824; VII, 190; VIII, 257, 516.

16-62 Alcoholysis of Anhydrides

Alkoxy-de-acyloxy-substitution

The scope of this reaction is similar to that of 16-61. Anhydrides are somewhat less reactive than acyl halides, but they are often used to prepare carboxylic esters. Acids,¹⁷²² Lewis acids,¹⁷²³ and bases (e.g., pyridine) are often used as catalysts.¹⁷²⁴ Acetic anhydride and NiCl₂ with microwave irradiation converts benzylic alcohols to the corresponding acetate.¹⁷²⁵ The monoacetates of 1,2-diols have been prepared using CeCl₃ as a catalyst.¹⁷²⁶ Pyridine is a nucleophilic-type catalyst (see Reaction 16-58), but DMAP is superior and can be used in cases where pyridine fails.¹⁷²⁷ Nonaromatic amidine derivatives have been used to catalyze the reaction with acetic anhydride.¹⁷²⁸ Formic anhydride is not a stable compound but esters of formic acid can be prepared by treating alcohols¹⁷²⁹ or phenols¹⁷³⁰ with acetic-formic anhydride. Cyclic anhydrides give monoesterified dicarboxylic acids (e.g., 102).¹⁷³¹ The asymmetric alcoholysis of cyclic anhydrides has been reviewed.¹⁷³²

Alcohols can also be acylated by mixed organic–inorganic anhydrides [e.g., acetic-phosphoric anhydride, MeCOOPO(OH)₂,¹⁷³³ see Reaction 16-68]. Thioesters of the type ArS(C=O)Me have been prepared by simple reaction of thiols and anhydrides in the presence of potassium carbonate,¹⁷³⁴ and from diphenyl disulfide and PBu₃, followed by treatment with acetic anhydride.¹⁷³⁵

OS I, 285, 418; II, 69, 124; III, 11, 127, 141, 169, 237, 281, 428, 432, 690, 833; IV, 15, 242, 304; V, 8, 459, 591, 887; VI, 121, 245, 560, 692; 486; VIII, 141, 258.

16-63 Esterification of Carboxylic Acids

Alkoxy-de-hydroxylation

The acid-catalyzed esterification of carboxylic acids with alcohols¹⁷³⁶ is the reverse of Reaction 16-60 and can be accomplished only if a means is available to drive the equilibrium to the right.¹⁷³⁷ There are many ways of doing this, among which are (1) addition of an excess of one of the reactants, usually the alcohol; (2) removal of the ester or the water by distillation; (3) removal of water by azeotropic distillation; and (4) removal of water by use of a dehydrating agent, silica gel,¹⁷³⁸ or a molecular sieve. When R′ is methyl, the most common way of driving the equilibrium is by adding excess MeOH; when R′ is ethyl or larger, it is preferable to remove water by azeotropic distillation.¹⁷³⁹ The most common catalysts are H₂SO₄ and TsOH, but some reactive carboxylic acids (e.g., formic,¹⁷⁴⁰ trifluoroacetic¹⁷⁴¹) do not require a catalyst. Ammonium salts have been used to initiate esterification,¹⁷⁴² and boric acid has been used to esterify α-hydroxy acids.¹⁷⁴³ The R′ group may be primary or secondary alkyl groups other than methyl or ethyl, but tertiary alcohols usually give carbocations and elimination. Phenols can sometimes be used to prepare phenolic esters, but yields are generally very low. Selective esterification of an aliphatic carboxylic acid in the presence of an aromatic acid was accomplished with NaHSO₄·SiO₂ and methanol.¹⁷⁴⁴Diphenylammonium triflate was useful for direct esterification of carboxylic acids with longer chain aliphatic alcohols.¹⁷⁴⁵ Photoirradiation of carboxylic acid with CBr₄¹⁷⁴⁶ or CCl₄¹⁷⁴⁷ in methanol was shown to give the methyl ester, with high selectivity for nonconjugated acids in the case of CBr₄. Esterification has been accomplished in ionic liquids.¹⁷⁴⁸ A solid-state esterification was reported on P₂O₅/SiO₂.¹⁷⁴⁹ Diols are converted to the monoacetate by heating with acetic acid on a zeolite.¹⁷⁵⁰ Vinyl acetate and iodine has been used for the acetylation of alcohols.¹⁷⁵¹

O-Alkylisoureas react with conjugated carboxylic acids to give the corresponding ester with microwave irradiation,¹⁷⁵² and a polymer-bound O-alkylurea has been used as well.¹⁷⁵³ Transition metal compounds of Ti,¹⁷⁵⁴ or Co¹⁷⁵⁵catalyze esterification. Allylic sulfonium salts react with carboxylic acids to give allylic esters in the presence of CuBr.¹⁷⁵⁶ Triphenylphosphine dibromide is a useful esterification reagent.¹⁷⁵⁷ Phenols can be esterified using amide acetals.¹⁷⁵⁸

The reaction of a carboxylic acid with an alcohol, in the presence of triphenylphosphine and DEAD gives the corresponding ester. This reaction is known as the Mitsunobu reaction (see 10-17). A variation of this esterification reaction used azopyridines to mediate formation of the ester.¹⁷⁵⁹

Both γ- and δ-hydroxy acids (e.g., 103) are easily converted to a lactone by treatment with acids, or often simply on standing, but larger and smaller lactone rings cannot be made in this manner, because polyester formation occurs more readily.¹⁷⁶⁰ Often the conversion of one group to a hydroxyl group, gives the lactone directly, since the hydroxy acid cyclizes too rapidly for isolation. Such groups include keto or halogen that are γ or δ to a carbotyl group. β-Substituted β-hydroxy acids can be converted to β-lactones by treatment with benzenesulfonyl chloride in pyridine at 0–5 °C.¹⁷⁶¹ ε-Lactones (seven-membered rings) have been made by cyclization of ε-hydroxy acids at high dilution.¹⁷⁶² Macrocyclic lactones¹⁷⁶³ can be prepared indirectly in very good yields by conversion of the hydroxy acids to 2-pyridinethiol esters and adding these to refluxing xylene.¹⁷⁶⁴

A closely related method, which often gives higher yields of a macrocyclic lactone, involves treatment of the hydroxy acids with 1-methyl- or 1-phenyl-2-halopyridinium salts, especially 1-methyl-2-chloropyridinium iodide (Mukaiyama's reagent).¹⁷⁶⁵ A macrocyclization technique has been developed based on formation of a mixed anhydride. The Yamaguchi protocol¹⁷⁶⁶ reacts a seco acid (the hydroxy acid precursor of a macrocyclic lactone) with 2,4,6-trichlorobenzoyl chloride. The resulting mixed anhydride is heated with DMAP in toluene.

Esterification is catalyzed by acids (not bases) in ways that were presented in Table 16.3 in Reaction 16-59.¹⁶⁵³ The mechanisms are usually A_AC2, but A_AC1 and A_AL1 have also been observed.¹⁷⁶⁷ Certain acids (e.g., 2,6-di-ortho-substituted benzoic acids), cannot be esterified by the A_AC2 mechanism because of steric hindrance (Sec. 10.G.i, category 1). In such cases, esterification can be accomplished by dissolving the acid in 100% H₂SO₄ (forming the ion RCO⁺) and pouring the solution into the alcohol (A_AC1 mechanism). The reluctance of hindered acids to undergo the normal A_AC2 mechanism can sometimes be put to advantage when, in a molecule containing two CO₂H groups, only the less hindered one is esterified. The A_AC1 pathway cannot be applied to unhindered carboxylic acids.

Another way to esterify a carboxylic acid is to treat it with an alcohol in the presence of a dehydrating agent.¹⁷³⁸ One of these is dicyclohexylcarbodiimide (DCC), which is converted in the process to dicyclohexylurea (DHU). The mechanism¹⁷⁶⁸ has much in common with the nucleophilic catalysis mechanism; the acid is converted to a compound with a better leaving group, but the conversion is not by a tetrahedral mechanism (as it is in nucleophilic catalysis), since the C–O bond remains intact during this step:

Evidence for this mechanism was the preparation of O-acylureas similar to 104 and the finding that when catalyzed by acids they react with alcohols to give esters.¹⁷⁶⁹ Hindered tertiary alcohols can be coupled via DCC to give the hindered ester.¹⁷⁷⁰ A polymer-bound carbodiimide has been used to prepare macrocyclic lactones.¹⁷⁷¹ In at least one case, the reaction of HOOCCH₂CN with DCC and tert-butanol gave the tert-butyl ester via a ketene intermediate.¹⁷⁷²

There are limitations to the use of DCC; yields are variable and N-acylureas are side products. Many other dehydrating agents¹⁷⁷³ have been used, including DCC and an aminopyridine,¹⁷⁷⁴ chlorosilanes,¹⁷⁷⁵ and N,N'-carbonyldiimidazole (105).¹⁷⁷⁶ In the latter case, imidazolides (106) are intermediates that react with alcohols.

It is known that the Lewis acid BF₃ promotes the esterification by converting the acid to RCO⁺⁻BF₃^–OH, so the reaction proceeds by an A_AC1 type of mechanism. The use of BF₃–etherate is simple and gives high yields.¹⁷⁷⁷ Other Lewis acids can be used.¹⁷⁷⁸

Carboxylic esters can also be prepared by treating carboxylic acids with tert-butyl ethers and acid catalysts:¹⁷⁷⁹

Carboxylic esters can be formed from the carboxylate anion and a suitable alkylating agent (Reaction 10-26).

Thioesters of the type RSC(=S)R′ (a dithiocarboxylic ester) and RSC(C=O)R′ (a thiocarboxylic ester) can be generated by reaction of carboxylic acids with thiols. In one example, phosphorous pentasulfide was used in conjunction with a thiol to make dithiocarboxylic esters¹⁷⁸⁰ or thiocarboxylic esters.¹⁷⁸¹ Thiocarboxylic esters were prepared from thiols and triflic acid.¹⁷⁸²

OS I, 42, 138, 237, 241, 246, 254, 261, 451; II, 260, 264, 276, 292, 365, 414, 526; III, 46, 203, 237, 381, 413, 526, 531, 610; IV, 169, 178, 302, 329, 390, 398, 427, 506, 532, 635, 677; V, 80, 762, 946; VI, 471, 797; VII, 93, 99, 210, 319, 356, 386, 470; VIII, 141, 251, 597; IX, 24, 58; 75, 116; 75, 129. Also see, OS III, 536, 742.

16-64 Transesterification

Alkoxy-de-alkoxylation

Transesterification¹⁷⁸³ is catalyzed¹⁷⁸⁴ by acids¹⁷⁸⁵ or bases,¹⁷⁸⁶ or done under neutral conditions.¹⁷⁸⁷ It is an equilibrium reaction that must be shifted in the desired direction.¹⁷⁸⁸ In many cases, low-boiling esters can be converted to higher-boiling ones by the distillation of the lower-boiling alcohol as fast as it is formed. Reagents used to catalyze¹⁷⁸⁹ transesterification include various Lewis acids.¹⁷⁹⁰ Zwitterionic salts have been used as organocatalysts for this reaction.¹⁷⁹¹ A polymer-bound siloxane has been used to induce transesterification.¹⁷⁹² Vinyl acetate has been used for transesterification, usually with a coreagent or metal mediator.¹⁷⁹³ This reaction has been used as a method for the acylation of a primary OH in the presence of a secondary OH.¹⁷⁹⁴ Regioselectivity has also been accomplished by using enzymes (lipases) as catalysts.¹⁷⁹⁵ Lactones, (e.g., 107) are easily opened by treatment with alcohols¹⁷⁹⁶to give open-chain hydroxy esters.

Transesterification has been carried out with phase-transfer catalysts, without an added solvent.¹⁷⁹⁷ Nonionic superbases (see Sec. 8.A.i) of the type P(RNCH₂CH₂)₃N catalyze the transesterification of carboxylic acid esters at 25 °C.¹⁷⁹⁸ Silyl esters (R′CO₂SiR₃) have been converted to alkyl esters (R′CO₂R) via reaction with alkyl halides and tetrabutylammonium fluoride.¹⁷⁹⁹ Thioesters are converted to phenolic esters by treatment with triphosgene–pyridine and then phenol.¹⁸⁰⁰

Transesterification occurs by mechanisms¹⁸⁰¹ that are identical with those of ester hydrolysis, except that ROH replaces HOH (by the acyl–oxygen fission mechanisms). When alkyl fission takes place, the products are the acid and the ether:

Therefore, transesterification reactions frequently fail when R′ is tertiary, since this type of substrate most often reacts by alkyl–oxygen cleavage. In such cases, the reaction is of the Williamson type with OCOR as the leaving group (see Reaction 10-10).

With enol esters (e.g., 108), reaction with an alcohol gives an ester and the enol of a ketone, which readily tautomerizes to the ketone as shown. Hence, enol esters are good acylating agents for alcohols.¹⁸⁰² This transformation has been accomplished in ionic liquid media,¹⁸⁰³ and there is a PdCl₂/CuCl₂ mediated version.¹⁸⁰⁴ Isopropenyl acetate can also be used to convert other ketones to the corresponding enol acetates in an exchange reaction:¹⁸⁰⁵

Enol esters can also be prepared in the opposite type of exchange reaction, catalyzed by mercuric acetate¹⁸⁰⁶ or Pd(II) chloride,¹⁸⁰⁷ for example,

A closely related reaction is equilibration of a dicarboxylic acid and its diester to produce monoesters: The reaction of a carboxylic acid with ethyl acetate, in the presence of NaHSO₄·SiO₂, was shown to give the corresponding ethyl ester.¹⁸⁰⁸ Iodine catalyzes the transesterification of β-keto esters.¹⁸⁰⁹

OS II, 5, 122, 360; III, 123, 146, 165, 231, 281, 581, 605; IV, 10, 549, 630, 977; V, 155, 545, 863; VI, 278; VII, 4, 164, 411; VIII, 155, 201, 235, 263, 350, 444, 528. See also, OS VII, 87; VIII, 71.

16-65 Alcoholysis of Amides

Alkoxy-de-amidation

Alcoholysis of amides is possible,¹⁸¹⁰ although it is usually difficult. It has been most common with the imidazolide type of amides (e.g., 100). For other amides, an activating agent is usually necessary before the alcohol will replace the NR₂ unit. N, N-Dimethylformamide, however, reacted with primary alcohols in the presence of 2,4,6-trichloro-1,3,5-pyrazine (cyanuric acid) to give the corresponding formate ester.¹⁸¹¹ Treatment of an amide with triflic anhydride (CF₃SO₂OSO₂CF₃) in the presence of pyridine, and then with an excess of alcohol, leads to the ester,¹⁸¹² as does treatment with Me₂NCH(OMe)₂ followed by the alcohol.¹⁸¹³ Trimethyloxonium tetrafluoroborate converted primary amides to methyl esters.¹⁸¹⁴ The reaction of acetanilide derivatives with sodium nitrite in the presence of acetic anhydride–acetic acid leads to phenolic acetates.¹⁸¹⁵ Acyl hydrazides (RCONHNH₂) were converted to esters by reaction with alcohols and various reagents,¹⁸¹⁶ and methoxyamides (RCONHOMe) were converted to esters with TiCl₄/ROH.¹⁸¹⁷ The reaction of an oxazolidinone amide (109) with methanol and 10% MgBr₂ gave the corresponding methyl ester.¹⁸¹⁸

C. Attack by OCOR at an Acyl Carbon

16-66 Acylation of Carboxylic Acids with Acyl Halides

Acyloxy-de-halogenation

Unsymmetrical, as well as symmetrical, anhydrides are often prepared by the treatment of an acyl halide with a carboxylic acid salt. If a metallic salt is used, Na⁺, K⁺, or Ag⁺ are the most common cations, but more often pyridine or another tertiary amine is added to the free acid. The resulting salt is subsequently treated with the acyl halide. Zinc–DMF has been used to mediate the synthesis of symmetrical anhydrides from acid chlorides.¹⁸¹⁹ Cobalt(II) chloride (CoCl₂) has been used as a catalyst.¹⁸²⁰ Mixed formic anhydrides are prepared from sodium formate and an aryl halide, by use of a solid-phase copolymer of pyridine-1-oxide.¹⁸²¹ Symmetrical anhydrides can be prepared by reaction of the acyl halide with aq NaOH or NaHCO₃ under phase-transfer conditions,¹⁸²² or with sodium bicarbonate with ultrasound.¹⁸²³

OS III, 28, 422, 488; IV, 285; VI, 8, 910; VIII, 132. See also, OS VI, 418.

16-67 Acylation of Carboxylic Acids with Carboxylic Acids

Acyloxy-de-hydroxylation

Anhydrides can be formed from two molecules of an ordinary carboxylic acid only if a dehydrating agent is present so that the equilibrium can be driven to the right. Common dehydrating agents¹⁸²⁴ are acetic anhydride, trifluoroacetic anhydride, dicyclohexylcarbodiimide,¹⁸²⁵ and P₂O₅. Triphenylphosphine/CCl₃CN with triethylamine has also been used with benzoic acid derivatives.¹⁸²⁶ The method is very poor for the formation of mixed anhydrides, which in any case generally undergo disproportionation to the two simple anhydrides when they are heated. However, simple heating of dicarboxylic acids does give cyclic anhydrides, provided that the ring formed contains five, six, or seven members, for example:

Malonic acid and its derivatives, which would give four-membered cyclic anhydrides, do not give this reaction when heated, but undergo decarboxylation (12-40) instead.

Carboxylic acids exchange with amides and esters; these methods are sometimes used to prepare anhydrides if the equilibrium can be shifted. Enolic esters are especially good for this purpose, because the equilibrium is shifted by formation of the ketone.

The combination of KF with 2-acetoxypropene under microwave conditions was effective.¹⁸²⁷ Carboxylic acids also exchange with anhydrides; indeed, this is how acetic anhydride acts as a dehydrating agent in this reaction.

Anhydrides can be formed from certain carboxylic acid salts (e.g., by treatment of trimethylammonium carboxylates with phosgene):¹⁸²⁸

or of thallium(I) carboxylates with thionyl chloride,¹⁷¹⁰ or of sodium carboxylates with CCl₄ and a catalyst (e.g., CuCl or FeCl₂).¹⁸²⁹

OS I, 91, 410; II, 194, 368, 560; III, 164, 449; IV, 242, 630, 790; V, 8, 822; IX, 151. Also see, OS VI, 757; VII, 506.

16-68 Preparation of Mixed Organic–Inorganic Anhydrides

Nitrooxy-de-acyloxy-substitution

Mixed organic–inorganic anhydrides are seldom isolated, although they are often intermediates when acylation is carried out with acid derivatives catalyzed by inorganic acids. Sulfuric, perchloric, phosphoric, and other acids form similar anhydrides, most of which are unstable or not easily obtained because the equilibrium lies in the wrong direction. These intermediates are formed from amides, carboxylic acids, and esters, as well as anhydrides. Organic anhydrides of phosphoric acid are more stable than most others and, for example, RCOOPO(OH)₂ can be prepared in the form of its salts.¹⁸³⁰ Mixed anhydrides of carboxylic and sulfonic acids (RCOOSO₂R′) are obtained in high yields by treatment of sulfonic acids with acyl halides or (less preferred) anhydrides.¹⁸³¹

OS I, 495; VI, 207; VII, 81.

16-69 Attack by SH or SR at an Acyl Carbon¹⁸³²

Thiol acids and thiol esters¹⁸³³ can be prepared in this manner, which is analogous to Reaction 16-57 and 16-64. Anhydrides¹⁸³⁴ and aryl esters (RCOOAr)¹⁸³⁵ are also used as substrates, but the reagents in these cases are usually HS⁻ and RS⁻. Thiol esters can also be prepared by treatment of carboxylic acids with P₄S₁₀–Ph₃SbO,¹⁸³⁶ or with a thiol (RSH) and either polyphosphate ester or phenyl dichlorophosphate (PhOPOCl₂).¹⁸³⁷ Carboxylic acids are converted to thioacids with Lawesson's reagent (structure 18 in Reaction 16-11).¹⁸³⁸ Esters RCOOR′ can be converted to thiol esters (RCOSR²) by treatment with trimethylsilyl sulfides (Me₃SiSR²) and AlCl₃.¹⁸³⁹

Alcohols, when treated with a thiol acid and zinc iodide, give thiol esters (R′COSR)¹⁸⁴⁰

OS III, 116, 599; IV, 924, 928; VII, 81; VIII, 71.

16-70 Transamidation

Alkylamino-de-amidation

It is sometimes necessary to replace one amide group with another, particularly when the group attached to nitrogen functions as a protecting group¹⁸⁴¹N-Benzyl amides can be converted to the corresponding N-allyl amide with allylamine and Ti catalysts.¹⁸⁴² Reaction of N-Boc 2-phenylethylamine with Ti(OiPr)₄ and benzyl alcohol, for example, gives the N-Cbz derivative.¹⁸⁴³N-Carbamoyl amines were converted to N-acetyl amines with acetic anhydride, Bu₃SnH, and a Pd catalyst¹⁸⁴⁴ Triethylaluminum converts methyl carbamates (ArNHCO₂Me) to the corresponding propanamide.¹⁸⁴⁵

A related process reacts acetamide with amines and aluminum chloride to give the N-acetyl amine.¹⁸⁴⁶ Another related process converted imides to O-benzyloxy amides by the Sm catalyzed reaction with O-benzylhydroxylamine.¹⁸⁴⁷

Thioamides can be prepared from amide by reaction with an appropriate sulfur reagent. The reaction of N,N-dimethylacetamide under microwave irradiation, with the polymer-bound reagent 110 (where = polymeric backbone) gave 111.¹⁸⁴⁸ Reaction of the thioamide with Bi(NO₃)₃·5 H₂O regenerates the amide.¹⁸⁴⁹ Other methods are known to convert a thioamide to an amide.¹⁸⁵⁰ Selenoamides [RC(=Se)NR′₂] have also been prepared from amides.¹⁸⁵¹

D. Attack by Halogen

16-71 The Conversion of Carboxylic Acids to Halides

Halo-de-oxido,oxo-tersubstitution

In certain cases, carboxyl groups can be replaced by halide. Acrylic acid derivatives (ArCH=CHCOOH), for example, react with 3 molar equivalents of Oxone in the presence of NaBr to give a vinyl bromide (ArCH=CHBr).¹⁸⁵²Diphosphorus tetraiodide/tetraethylammonium bromide (TEAB) readily converts conjugated acids to vinyl bromides.¹⁸⁵³ In other cases, conjugated acids, (e.g., 112), have been converted to the bromide by reaction with (NBS, Reaction 14-3) and LiOAc.¹⁸⁵⁴

E. Attack by Nitrogen at an Acyl Carbon¹⁸⁵⁵

16-72 Acylation of Amines by Acyl Halides

Amino-de-halogenation

The treatment of acyl halides with ammonia or amines is a very general reaction for the preparation of amides.¹⁸⁵⁶ The reaction is exothermic and must be carefully controlled, usually by cooling or dilution. Ammonia gives unsubstituted amides, primary amines give N-substituted amides,¹⁸⁵⁷ and secondary amines give N,N-disubstituted amides. Arylamines can be similarly acylated. Hydroxamic acids have been prepared by this route.¹⁸⁵⁸ In some cases, aq alkali is added to combine with the liberated HCl. This is called the Schotten–Baumann procedure, as in Reaction 16-61. Activated Zn can be used to increase the rate of amide formation when hindered amines and/or acid chlorides are used.¹⁸⁵⁹ A solvent-free reaction was reported using DABCO and methanol.¹⁸⁶⁰ Metal-mediated reactions using In,¹⁸⁶¹ Sm,¹⁸⁶² or a BiOCl mediated reaction¹⁸⁶³ have been reported. A variation of this basic reaction uses DMF with acyl halides to give N,N-dimethylamides.¹⁸⁶⁴ Formic acid and iodine react with amines to give the formamide.¹⁸⁶⁵

Hydrazine and hydroxylamine also react with acyl halides to give, respectively, hydrazides (RCONHNH₂)¹⁸⁶⁶ and hydroxamic acids (RCONHOH).¹⁸⁶⁷ When phosgene is the acyl halide, both aliphatic and aromatic primary amines give chloroformamides (ClCONHR) that lose HCl to give isocyanates (RNCO).¹⁸⁶⁸ This is one of the most common methods for the preparation of isocyanates.¹⁸⁶⁹ Similar

treatment with thiophosgene¹⁸⁷⁰ gives isothiocyanates. A safer substitute for phosgene in this reaction is trichloromethyl chloroformate (CCl₃OCOCl).¹⁸⁷¹ When chloroformates (ROCOCl) are treated with primary amines, carbamates (ROCONHR′) are obtained.¹⁸⁷² An example of this reaction is the use of benzyl chloroformate to protect the amino group of amino acids and peptides.

The PhCH₂OCO group in 113 has been called the carbobenzoxy group,¹⁸⁷³ and is often abbreviated Cbz or Z, but it is really a benzyl carbamate. Another important group similarly used is Boc, which is a tert-butyl carbamate. In this case, the chloride (Me₃COCOCl) is unstable, so the anhydride [(Me₃COCO)₂O] is used instead, in an example of Reaction 16-73. Amino groups in general are often protected by conversion to amides.¹⁸⁷⁴ The reactions proceed by the tetrahedral mechanism.¹⁸⁷⁵

An interesting variation of this transformation reacts carbamoyl chlorides with organocuprates to give the corresponding amide.¹⁸⁷⁶

OS I, 99, 165; II, 76, 208, 278, 328, 453; III, 167, 375, 415, 488, 490, 613; IV, 339, 411, 521, 620, 780; V, 201, 336; VI, 382, 715; VII, 56, 287, 307; VIII, 16, 339; IX, 559; 81, 254. See also, OS VII, 302.

16-73 Acylation of Amines by Anhydrides

Amino-de-acyloxy-substitution

This reaction, similar in scope and mechanism¹⁸⁷⁷ to Reaction 16-72, can be carried out with ammonia or primary or secondary amines.¹⁸⁷⁸ Note that there is a report where a tertiary amine (an N-alkylpyrrolidine) reacted with acetic anhydride at 120 °C, in the presence of a BF₃·etherate catalyst, to give N-acetylpyrrolidine (an acylative dealkylation).¹⁸⁷⁹ Amino acids can be N-acylated using acetic anhydride and ultrasound.¹⁸⁸⁰ However, ammonia and primary amines can also give imides, in which two acyl groups are attached to the nitrogen. The conversion of cyclic anhydrides to cyclic imides is generally facile,¹⁸⁸¹ although elevated temperatures are occasionally required to generate the imide.¹⁸⁸² Microwave irradiation of formamide and a cyclic anhydride generates the cyclic imide.¹⁸⁸³ Cyclic imides have also been formed in ionic liquids.¹⁸⁸⁴ Cyclic imides were also formed by microwave irradiation of a polymer-bound phthalate after initial reaction with an amine.¹⁸⁸⁵

The second step for imide formation, which is much slower than the first, is the attack of the amide nitrogen on the carboxylic carbon. Unsubstituted and N-substituted amides have been used instead of ammonia. Since the other product of this reaction is RCOOH, this is a way of “hydrolyzing” such amides in the absence of water.¹⁸⁸⁶

Even though formic anhydride is not a stable compound (see Reaction 11-17), amines can be formylated with the mixed anhydride of acetic and formic acids (HCOOCOMe)¹⁸⁸⁷ or with a mixture of formic acid and acetic anhydride. Acetamides are not formed with these reagents. Secondary amines can be acylated in the presence of a primary amine by conversion to their salts and addition of 18-crown-6.¹⁸⁸⁸ The crown ether complexes the primary ammonium salt, preventing its acylation, while the secondary ammonium salts, which do not fit easily into the cavity, are free to be acylated. Dimethyl carbonate can be used to prepare methyl carbamates in a related procedure.¹⁸⁸⁹N-Acetylsulfonamides were prepared from acetic anhydride and a primary sulfonamide, catalyzed by Montmorillonite K10–FeO¹⁸⁹⁰ or sulfuric acid.¹⁸⁹¹

There are acylating reagents other than anhydrides of course. The reaction with acyl halides is discussed in Reaction 16-72. There are a few specialized reagents. Kinetic resolution of racemic amines was accomplished using (1S,2S)-N-acetyl-1,2- bis(trifluoromethanesulfonamido)cyclohexane.¹⁸⁹²

OS I, 457; II, 11; III, 151, 456, 661, 813; IV, 5, 42, 106, 657; V, 27, 373, 650, 944, 973; VI, 1; VII, 4, 70; VIII, 132; 76, 123.

16-74 Acylation of Amines by Carboxylic Acids

Amino-de-hydroxylation

When carboxylic acids are treated with ammonia or amines, salts are obtained. The salts of ammonia or primary or secondary amines can be pyrolyzed to give amides,¹⁸⁹³ but the method is less convenient than Reaction 16-72, 16-73, and 16-75 and is seldom of preparative value.¹⁸⁹⁴ Heating in the presence of a base (e.g., hexamethyldisilazide) makes the amide-forming process more efficient.¹⁸⁹⁵ Boronic acids catalyze the direct conversion of carboxylic acid and amine to amides.¹⁸⁹⁶ Polymer-bound reagents have also been used.¹⁸⁹⁷ Triphenylphosphine/trichloroisocyanuric acid converts acids and amides to the amide.¹⁸⁹⁸ The Burgess reagent (Et₃N^+–SO₂N–CO₂Me; see Reaction 17-29) activates carboxylic acids for amide formation.¹⁸⁹⁹ The reaction of a carboxylic acid and imidazole under microwave irradiation gives the amide.¹⁹⁰⁰ Microwave irradiation of a secondary amine, formic acid, 2-chloro-4,6-dimethoxy[1,3,5]triazine, and a catalytic amount of DMAP leads to the formamide.¹⁹⁰¹ Ammonium bicarbonate and formamide converts acids to amides with microwave irradiation.¹⁹⁰² Formamides are produced from formic acid and anion nitriles in the presence of ZnO.¹⁹⁰³

Lactams are readily produced from γ- or δ-amino acids,¹⁹⁰⁴ for example,

This lactonization process can be promoted by enzymes (e.g., pancreatic porcine lipase).¹⁹⁰⁵ Reduction of ω-azide carboxylic acids leads to macrocyclic lactams.¹⁹⁰⁶

Although treatment of carboxylic acids with amines does not directly give amides, the reaction can be made to proceed in good yield at room temperature or slightly above by the use of coupling agents,¹⁹⁰⁷ the most important of which is dicyclohexylcarbodiimide. This reagent is very convenient and is used¹⁹⁰⁸ a great deal in peptide synthesis.¹⁹⁰⁹ A polymer-supported carbodiimide has been used.¹⁹¹⁰ The mechanism is probably the same as in Reaction 16-63 up to the formation of 114. This intermediate is then attacked by another molecule of RCOO⁻ to give the anhydride (RCO)₂O, which is the actual species that reacts with the amine:

The anhydride has been isolated from the reaction mixture and then used to acylate an amine.¹⁹¹¹

The synthetically important Weinreb amides [RCON(Me)OMe, see Reaction 16-82] can be prepared from the carboxylic acid and MeO(Me)NH·HCl in the presence of tributylphosphine and 2-pyridine-N-oxide disulfide.¹⁹¹² Di(2-pyridyl)carbonate has been used in a related reaction that generates amides directly.¹⁹¹³ Other promoting agents¹⁹¹⁴ are ArB(OH)₂ reagents,¹⁹¹⁵N,N'-carbonyldiimidazole (115, in Reaction 16-63),¹⁹¹⁶ POCl₃,¹⁹¹⁷ TiCl₄,¹⁹¹⁸ molecular sieves,¹⁹¹⁹Lawesson's reagent (Reaction 16-11),¹⁹²⁰ and (MeO)₂POCl.¹⁹²¹ Certain dicarboxylic acids form amides simply on treatment with primary aromatic amines. In these cases, the cyclic anhydride is an intermediate and is the species actually attacked by the amine.¹⁹²² Carboxylic acids can also be converted to amides by heating with amides of carboxylic acids (exchange),¹⁹²³ sulfonic acids, or phosphoric acids, for example,¹⁹²⁴

or by treatment with trisalkylaminoboranes [B(NHR′)₃], with trisdialkylaminoboranes [B(NR₂')₃],¹⁹²⁵

or with bis(diorganoamino)magnesium reagents [(R₂N)₂Mg].¹⁹²⁶ The reaction of thiocarboxylic acids and azides, in the presence of triphenylphosphine, gives the corresponding amide.¹⁹²⁷

An important technique, discovered by R.B. Merrifield¹⁹²⁸ and since used for the synthesis of many peptides,¹⁹²⁹ is called solid-phase synthesis or polymer-supported synthesis.¹⁹³⁰ The reactions used are the same as in ordinary synthesis, but one of the reactants is anchored onto a solid polymer. For example, if it is desired to couple two amino acids (to form a dipeptide), the polymer selected might be polystyrene with CH₂Cl side chains. One of the amino acids, protected by (Boc), would then be coupled to the side chains. It is not necessary that all the side chains be converted, but a random selection will be converted. The Boc group is then removed by hydrolysis with trifluoroacetic acid in CH₂Cl₂ and the second amino acid is coupled to the first, using DCC or some other coupling agent. The second Boc group is removed, resulting in a dipeptide that is still anchored to the polymer. If this dipeptide is the desired product, it can be cleaved from the polymer by various methods,¹⁹³¹ one of which is treatment with HF. If a longer peptide is wanted, additional amino acids can be added by repeating the requisite steps.

The basic advantage of the polymer support techniques is that the polymer (including all chains attached to it) is easily separated from all other reagents, because it is insoluble in the solvents used. Excess reagents, other reaction products (e.g., dicyclohexylurea), side products, and the solvents themselves are quickly washed away. Purification of the polymeric species is rapid and complete. The process can even be automated,¹⁹³² to the extent that six or more amino acids can be added to a peptide chain in one day. Commercial automated peptide synthesizers are now available.¹⁹³³

Although the solid-phase technique was first developed for the synthesis of peptide chains and has seen considerable use for this purpose, it has also been used to synthesize chains of polysaccharides and polynucleotides; in the latter case, solid-phase synthesis has almost completely replaced synthesis in solution.¹⁹³⁴ The technique has been applied less often to reactions in which only two molecules are brought together (nonrepetitive syntheses), but many examples have been reported.¹⁹³⁵ Combinatorial chemistry in some ways can be viewed as an extension of the Merrifield synthesis, particularly when applied to peptide synthesis, and continues as an important part of modern organic chemistry.¹⁹³⁶

OS I, 3, 82, 111, 172, 327; II, 65, 562; III, 95, 328, 475, 590, 646, 656, 768; IV, 6, 62, 513; V, 670, 1070; VIII, 241; 81, 262. Also see, OS III, 360; VI, 263; VIII, 68.

16-75 Acylation of Amines by Carboxylic Esters

Amino-de-alkoxylation

The conversion of carboxylic esters to amides is a useful reaction, and unsubstituted, N-substituted, and N,N-disubstituted amides can be prepared this way from the appropriate amine¹⁹³⁷ or ammonia.¹⁹³⁸ Both R and R′ can be alkyl or aryl, but an especially good leaving group is p-nitrophenyl. Ethyl trifluoroacetate was found to react selectively with primary amines to form the corresponding trifluoroacetyl amide.¹⁹³⁹ Many simple esters (R = Me, Et, etc.) are not very reactive, and strongly basic catalysis has been used in such cases,¹⁹⁴⁰ but catalysis by cyanide ion¹⁹⁴¹ MgBr₂,¹⁹⁴² InI₃,¹⁹⁴³ and acceleration by high pressure¹⁹⁴⁴ have been reported. Methyl esters¹⁹⁴⁵ and ethyl esters¹⁹⁴⁶have been converted to the corresponding amide under microwave irradiation. Lithium amides have been used to convert esters to amides as well.¹⁹⁴⁷ β-Keto esters undergo the reaction especially easily.¹⁹⁴⁸ Aniline was treated with n-butyllithium to form the lithium amide, which reacted with an ester to give the amide.¹⁹⁴⁹ An enzyme-mediated amidation is known using amino cyclase I.¹⁹⁵⁰ The reaction of dimethyl carbonate and an amine is an effective way to prepare methyl carbamates.¹⁹⁵¹

Lactones give lactams when treated with ammonia or primary amines. Lactams are also produced from γ- and δ-amino esters in an internal example of this reaction. Lactonization has been accomplished in ionic liquids.¹⁹⁵²

As in Reaction 16-72, hydrazides and hydroxamic acids can be prepared from carboxylic esters, with hydrazine and hydroxylamine,¹⁹⁵³ respectively. Both hydrazine and hydroxylamine react more rapidly than ammonia or primary amines (the alpha effect, Sec. 10.G.ii). Imidates [RC(=NH)OR′] give amidines [RC(=NH)NH₂]. Isopropenyl formate is a useful compound for the formylation of primary and secondary amines.¹⁹⁵⁴

Although more studies have been devoted to the mechanism of the acylation of amines with carboxylic esters than with other reagents, the mechanistic details are not yet entirely clear.¹⁹⁵⁵ In its broad outlines, the mechanism appears to be essentially B_AC2.¹⁹⁵⁶ Under normal basic conditions, the reaction is general base catalyzed,¹⁹⁵⁷ indicating that a proton is being transferred in the rate-determining step and that two molecules

of amine are involved.¹⁹⁵⁸ Alternatively, another base (e.g., H₂O or OH⁻) can substitute for the second molecule of amine. With some substrates and under some conditions, especially at low pH, the breakdown of 116 can become rate determining.¹⁹⁵⁹ The reaction also takes place under acidic conditions and is general acid catalyzed, so that breakdown of 116 is rate determining and proceeds as follows:¹⁹⁶⁰

Here HA may be R²NH₃⁺ or another acid. Intermediate 116 may or may not be further protonated on the nitrogen. Even under basic conditions, a proton donor may be necessary to assist leaving-group removal. Evidence for this is that the rate is lower with NR₂⁻ in liquid ammonia than with NHR₂ in water, apparently owing to the lack of acids to protonate the leaving oxygen.¹⁹⁶¹

In the special case of β-lactones, where small-angle strain is an important factor, alkyl–oxygen cleavage is observed (B_AL2 mechanism, as in the similar case of hydrolysis of β-lactones, Reaction 16-59), and the product is not an amide but a β-amino acid (β-alanine).

A similar result has been found for certain sterically hindered esters.¹⁹⁶² This reaction is similar to 10-31, with OCOR as the leaving group. Other lactones have been opened to ω-hydroxy amides with Dibal-H:BnNH₂.¹⁹⁶³

OS I, 153, 179; II, 67, 85; III, 10, 96, 108, 404, 440, 516, 536, 751, 765; IV, 80, 357, 441, 486, 532, 566, 819; V, 168, 301, 645; VI, 203, 492, 620, 936; VII, 4, 30, 41, 411; VIII, 26, 204, 528. Also see, OS I, 5; V, 582; VII, 75.

16-76 Acylation of Amines by Amides

Alkylamino-de-amination

This is an exchange reaction and is usually carried out with the salt of the amine.¹⁹⁶⁴ The leaving group is usually NH₂ rather than NHR or NR₂ and primary amines (in the form of their salts) are the most common reagents. Boron trifluoride can be added to complex with the leaving ammonia. Neutral amines also react in some cases to give the new amide.¹⁹⁶⁵ The reaction is catalyzed by Al(III).¹⁹⁶⁶ The reaction is often used to convert urea to substituted ureas: NH₂CONH₂ + RNH₃⁺ → NH₂CONHR + NH₄⁺.¹⁹⁶⁷ An N-aryl group of a urea can be converted to a N,N-dialkyl group by heating the urea with the amine in an autoclave.¹⁹⁶⁸N-alkyl substituted amides (alkyl = R) are converted to N-R′-substituted amides, where alkyl = R′, by treatment with N₂O₄ to give an N-nitroso compound, followed by treatment of this with a primary amine (R′NH₂).¹⁹⁶⁹ Lactams can be converted to ring-expanded lactams (e.g., 117) if a side chain containing an amino group is present on the nitrogen. A strong base is used to convert the NH₂ to NH^–, which then acts as a nucleophile, expanding the ring by means of a transamidation.¹⁹⁷⁰ The discoverers call it the Zip reaction, by analogy with the action of zippers.¹⁹⁷¹

Lactams can be opened to ω-amino amides by reaction with amines at 10 kbar.¹⁹⁷²

OS I, 302 (but see V, 589), 450, 453; II, 461; III, 151, 404; IV, 52, 361. See also, OS VIII, 573.

16-77 Acylation of Amines by Other Acid Derivatives

Acylamino-de-halogenation or dealkoxylation

Acid derivatives that can be converted to amides include thiol acids (RCOSH), thiol esters (RCOSR),¹⁹⁷³ acyloxyboranes [RCOB(OR′)₂],¹⁹⁷⁴ α-keto nitriles, acyl azides, and nonenolizable ketones (see the Haller–Bauer Reaction12-34). N-Acylsulfonamides react with primary amines to the amide (AcNHR).¹⁹⁷⁵ Carbonylation reactions can be used to prepare amides and related compounds. The reaction of a primary amine, an alkyl halide with CO₂, in the presence of Cs₂CO₃/Bu₄NI, gave the corresponding carbamate.¹⁹⁷⁶

OS III, 394; IV, 6, 569; V, 160, 166; VI, 1004.

Imides can be prepared by the attack of amides or their salts on acyl halides, anhydrides, and carboxylic acids or esters.¹⁹⁷⁷ A good synthetic method for the preparation of acyclic imides is the reaction between an amide and an anhydride at 100 °C catalyzed by H₂SO₄.¹⁹⁷⁸ When acyl chlorides are treated with amides in a 2: 1 molar ratio at low temperatures in the presence of pyridine, the products are N,N-diacylamides [(RCO)₃N].¹⁹⁷⁹

This reaction is often used to prepare urea derivatives, an important example being the preparation of barbituric acid (118).¹⁹⁸⁰

When the substrate is oxalyl chloride (ClCOCOCl) and the reagent an unsubstituted amide, an acyl isocyanate (RCONCO) is formed. The “normal” product (RCONHCOCOCl) does not form, or if it does, it rapidly loses CO and HCl.¹⁹⁸¹

OS II, 60, 79, 422; III, 763; IV, 245, 247, 496, 566, 638, 662, 744; V, 204, 944.

16-78 Acylation of Azides

The reaction of an aldehyde with sodium azide and Et₄⁺ I(OAc)₂ or polymer-bound PhI(OAc)₂ leads to an acyl azide.¹⁹⁸² Acyl azides are also prepared directly from aldehydes using tert-butyl hypochlorite.¹⁹⁸³

F. Attack by Halogen at an Acyl Carbon

16-79 Formation of Acyl Halides from Carboxylic Acids

Halo-de-hydroxylation

The same inorganic acid halides that convert alcohols to alkyl halides (Reaction 10-48) also convert carboxylic acids to acyl halides.¹⁹⁸⁴ The reaction is the best and the most common method for the preparation of acyl chlorides. Bromides and iodides¹⁹⁸⁵ are also made in this manner, but much less often. Acyl bromides can be prepared with BBr₃ on alumina,¹⁹⁸⁶ or with ethyl tribromoacetae/PPh₃.¹⁹⁸⁷ Thionyl chloride¹⁹⁸⁸ is a good reagent, since the byproducts are gases and the acyl halide is easily isolated, but PX₃ and PX₅ (X = Cl or Br) are also commonly used.¹⁹⁸⁹ Hydrogen halides do not give the reaction. A particularly mild procedure, similar to one mentioned in Reaction 10-48, involves reaction of the acid with Ph₃P in CCl₄, whereupon acyl chlorides are produced without obtaining any acidic compound as a byproduct.¹⁹⁹⁰

Oxalyl chloride (113) and oxalyl bromide are mild and often superior reagents, since the oxalic acid byproduct decomposes to CO and CO₂, and the equilibrium is thus driven to the side of the other acyl halide.¹⁹⁹¹ These reagents are commonly the reagent of choice, particularly when sensitive functionality is present elsewhere in the molecule.

Acyl fluorides can be prepared by treatment of carboxylic acids with cyanuric fluoride.¹⁹⁹²C,N-Chelated di-n-butyltin(IV) fluoride has been used to prepare acyl fluorides.¹⁹⁹³ Acid salts are also sometimes used as substrates and acyl halides are used as reagents in an exchange reaction:

which probably involves an anhydride intermediate. This is an equilibrium reaction that must be driven to the desired side.

OS I, 12, 147, 394; II, 74, 156, 169, 569; III, 169, 490, 547, 555, 613, 623, 712, 714; IV, 34, 88, 154, 263, 339, 348, 554, 608, 616, 620, 715, 739, 900; V, 171, 258, 887; VI, 95, 190, 549, 715; VII, 467; VIII, 441, 486, 498.

16-80 Formation of Acyl Halides from Acid Derivatives

Halo-de-acyloxy-substitution

Halo-de-halogenation

These reactions are most important for the preparation of acyl fluorides.¹⁹⁹⁴ Acyl chlorides and anhydrides can be converted to acyl fluorides by treatment with polyhydrogen fluoride–pyridine solution¹⁹⁹⁵ or with liquid HF at –10 °C.¹⁹⁹⁶ Formyl fluoride, which is a stable compound, was prepared by the latter procedure from the mixed anhydride of formic and acetic acids.¹⁹⁹⁷ Acyl fluorides can also be obtained by reaction of acyl chlorides with KF in acetic acid¹⁹⁹⁸ or with diethylaminosulfur trifluoride (DAST).¹⁹⁹⁹ Carboxylic esters and anhydrides can be converted to acyl halides other than fluorides by the inorganic acid halides mentioned in Reaction 16-79, as well as with Ph₃PX₂ (X = Cl or Br),²⁰⁰⁰ but this is seldom done. Halide exchange can be carried out in a similar manner. When halide exchange is done, it is always acyl bromides and iodides that are made from chlorides, since chlorides are by far the most readily available.²⁰⁰¹

OS II, 528; III, 422; V, 66, 1103; IX, 13. See also, OS IV, 307.

G. Attack by Carbon at an Acyl Carbon²⁰⁰²

16-81 The Conversion of Acyl Halides to Ketones with Organometallic Compounds²⁰⁰³

Alkyl-de-halogenation

Acyl halides react cleanly and under mild conditions with lithium dialkylcopper reagents (see Reaction 10-58)²⁰⁰⁴ to give high yields of ketones.²⁰⁰⁵ The R′ group may be primary, secondary, or tertiary alkyl or aryl and may contain iodo, keto, ester, nitro, or cyano groups. The R groups that have been used successfully are methyl, primary alkyl, and vinylic. Secondary and tertiary alkyl groups can be introduced by the use of PhS(R)CuLi (Reaction 10-58) instead of R₂CuLi,²⁰⁰⁶ or by the use of either the mixed homocuprate [(R′SO₂CH₂CuR)⁻ Li⁺],²⁰⁰⁷ or a magnesium dialkylcopper reagent ‘(RMeCuMgX).’²⁰⁰⁸ Secondary alkyl groups can also be introduced with the copper–zinc reagents [RCu(CN)ZnI].²⁰⁰⁹ The R group may be alkynyl if a cuprous acetylide (R²C≡CCu) is the reagent.²⁰¹⁰ Organocopper reagents generated in situ from highly reactive copper, and containing such functional groups as cyano, chloro, and ester, react with acyl halides to give ketones.²⁰¹¹

When the organometallic compound is a Grignard reagent,²⁰¹² ketones are generally not obtained because the initially formed ketone reacts with a second molecule of RMgX to give the salt of a tertiary alcohol (Reaction 16-82). Ketones have been prepared in this manner by the use of low temperatures, inverse addition (i.e., addition of the Grignard reagent to the acyl halide rather than the other way), excess acyl halide, and so on, but the yields are usually low, although high yields have been reported in THF at –78 °C.²⁰¹³ Pretreatment with a trialkylphosphine followed by the Grignard reagent gave the ketone.²⁰¹⁴ Using CuBr²⁰¹⁵ or a Ni catalyst²⁰¹⁶ with the Grignard reagent can lead to the ketone. Some ketones are unreactive toward Grignard reagents for steric or other reasons; these can be prepared in this way.²⁰¹⁷ Other methods involve running the reaction in the presence of Me₃SiCl²⁰¹⁸, which reacts with the initial adduct in the tetrahedral mechanism (Sec. 16.A.i), and the use of a combined Grignard–LiNEt₂ reagent.²⁰¹⁹ Certain metallic halides, notably ferric and cuprous halides, are catalysts that improve the yields of ketone at the expense of tertiary alcohol.²⁰²⁰ For this catalysis, both free radical and ionic mechanisms have been proposed.²⁰²¹