Organic Synthesis and Functional Group Transformations

Functional group interconversion

Functional group interconversion (FGI) means changing one functional group into another while the carbon skeleton stays essentially intact. In synthesis, it is the basic tool for moving from an available starting material to a target. For each step, the question is which reagent converts the wanted group and leaves the others untouched. To study FGI effectively, learn each conversion together with its reagent, its selectivity, and the conditions that decide the outcome. A useful way to organise the material is by oxidation level. Reductions lower the oxidation state of carbon (alkene to alkane, ketone to alcohol, ester to alcohol), oxidations raise it, and some FGIs are rearrangements or substitutions that keep the oxidation level the same.

Reduction and hydrogen-transfer reagents. Diimide, generated from hydrazine with a copper catalyst and oxygen, reduces a C=C bond by a concerted syn transfer of two hydrogen atoms. With a limited amount of the reagent, one of two alkenes in a diene can be reduced selectively. Metal hydrides differ in strength and selectivity. LiAlH_4 reduces esters and carboxylic acids, whereas DIBAL-H at low temperature can stop at the aldehyde stage, and an excess of DIBAL-H carries the reduction of an ester on to the primary alcohol. LiBH_4 reduces an imide carbonyl to a primary alcohol and so can be used to remove a chiral auxiliary. Sodium borohydride is milder and is used mainly for aldehydes and ketones. Catalytic hydrogenation with Pd/C saturates alkenes, and a desulfonylation with sodium amalgam removes a sulfonyl group next to a carbanion. Hydrogenolysis over Raney nickel cleaves a weak N–O bond, which converts an isoxazoline into a \beta-hydroxy imine. Dissolving-metal Birch reduction reduces an aromatic ring to a 1,4-diene, and it reduces a conjugated enone to the saturated ketone via the enolate. Reduction of a protonated C=N bond with Sn/HCl converts an isoquinoline into a tetrahydroisoquinoline. Low-valent titanium, generated from TiCl_3 and zinc–copper, is a single-electron reductant that couples two carbonyl groups in a pinacol reaction to give a vicinal diol.

Oxidation and oxygen transfer. Nucleophilic epoxidation of an electron-poor enone uses alkaline hydrogen peroxide (Weitz–Scheffer) or dimethyldioxirane, and both give an \alpha,\beta-epoxy ketone. Peracids such as m-CPBA epoxidise electron-rich alkenes, and a Lewis acid such as BF_3·OEt_2 can then open the epoxide by a Meinwald-type shift. Osmium tetroxide with NMO gives a syn-diol from an alkene, and the existing hydroxyl groups in the substrate decide the face of attack. Cold dilute permanganate also gives a cis-diol, and periodate cleaves a 1,2-diol into two carbonyl compounds. Ozonolysis followed by a reductive work-up cleaves a C=C bond to two carbonyl groups, which can then be used in a further ring-forming step such as an intramolecular aldol condensation. Allylic oxidation by selenium dioxide or manganese dioxide converts an allylic CH_2 or an allylic alcohol into an enone, and Dess–Martin periodinane oxidises a secondary alcohol to a ketone. The Parikh–Doering reagent (SO_3·pyridine/DMSO) oxidises alcohols, but it does not transfer oxygen to a C=C bond. Bromination at an allylic position with NBS and a radical initiator, followed by hydrolysis of the bromide, is a standard route from an alkene to an allylic alcohol.

Substitution, elimination and ring closure. An alkoxide can displace a tosylate intramolecularly by an SN2 reaction with inversion, but only when the nucleophile and the leaving group are trans-diaxial on the ring. A bromohydrin, formed by adding Br_2 and water across an alkene, closes to an epoxide under base because the alkoxide displaces the neighbouring bromide. Lithium–halogen exchange at low temperature is fast, and a neighbouring alkoxy group can direct the exchange to the adjacent halogen. An organolithium or Grignard reagent adds to an aldehyde to give a secondary alcohol. A Weinreb amide accepts one equivalent of a Grignard reagent and stops at the ketone, because the chelated tetrahedral intermediate does not collapse until work-up, so it is the standard route to methyl ketones. Enol ethers hydrolyse under acid to ketones, which is useful when an alkyl halide has been alkylated onto a masked ketone carbanion. Sodium hydride with dimethyl carbonate installs a methoxycarbonyl group on a ketone to give a \beta-keto ester. A chiral N-acyl oxazolidinone can be enolised with LDA and alkylated with high facial selectivity, and the auxiliary is later removed reductively. Enzymes such as Candida antarctica lipase B catalyse acetylation with isopropenyl acetate, which is irreversible because the enol by-product tautomerises to acetone. The enzyme acetylates one alcohol enantiomer faster, so the reaction is a kinetic resolution. Intramolecular ring closure with a phosphonium ylide gives a cyclopropyl phosphonium salt, which on deprotonation gives a cyclopropylidene ylide.

Carbon–carbon bond and ring-forming FGIs. A Wittig reaction converts a carbonyl into an alkene, and the Horner–Wadsworth–Emmons variant gives mainly (E)-\alpha,\beta-unsaturated esters. A vinyl triflate, made from a ketone, can be coupled with an alkene such as methyl acrylate in a Heck reaction. A Shapiro reaction converts a tosylhydrazone of a ketone into an alkene with the double bond on the less substituted side. Ring-closing metathesis joins two terminal alkenes and releases ethylene, and ring-opening cross metathesis of a strained cyclobutene opens the ring to a divinylcyclobutane. A divinylcyclobutane undergoes a boat-like Cope [3,3] shift on heating to give a cyclooctadiene. A Diels–Alder reaction with a vinyl sulfone builds a six-membered ring carrying a sulfonyl group, which can be removed later. A 1,3-dipolar cycloaddition of a nitrile oxide to an alkene gives an isoxazoline with the alkene geometry retained. Benzyne, formed from a 2-(trimethylsilyl)aryl triflate with fluoride, reacts with enolates and alkenes. An allylic trifluoroacetate ionises with a Lewis acid to an allyl cation, which can add to a diene in a formal [4+3] cycloaddition, and a \beta-silyl group then leaves to give an alkene. A palladium(0) catalyst ionises an allylic acetate to a \pi-allyl complex, which can insert into a pendant alkene to close a five-membered ring. A radical generated from an alkyl iodide with tributyltin hydride cyclises onto a nearby alkene, and 5-exo-trig addition is preferred to 6-endo-trig. Carbonyl compounds and carbanions can also be linked by an intramolecular aldol condensation that closes a new ring.

Skeletal rearrangements. Several named rearrangements change the ring size or the connectivity while converting one functional group into another. In the Beckmann reaction, an oxime rearranges to an amide, but an oxime whose carbon is attached to a quaternary carbon can instead fragment to a nitrile and a stabilised carbocation. The Favorskii rearrangement of an \alpha-halo ketone with alkoxide proceeds through a cyclopropanone and gives a ring-contracted ester, and when the enolate is formed on the side that has no halogen, the same carbonyl-to-ester conversion can occur with a different product. The Gabriel–Colman rearrangement of an N-phenacylphthalimide with base expands the imide ring to an isoquinoline. A di-\pi-methane rearrangement occurs under irradiation of a 1,4-diene, which forms a vinylcyclopropane. Norrish type I cleavage breaks the bond between a photoexcited ketone carbonyl and an adjacent carbon, and the resulting radicals recombine to give a new ring. Photolysis of a 1,2-benzisoxazole cleaves the N–O bond, and the nitrogen then closes to an azirine that rearranges to a benzoxazole. The Bartoli indole synthesis converts an ortho-substituted nitroarene into an indole when excess vinyl Grignard reagent is used, and the ortho substituent ends up at the 7-position of the indole. Reimer–Tiemann conditions generate dichlorocarbene, which adds to the ring of a naphthol and expands the ring to a seven-membered benzotropone. A benzylic ketone with two methyl groups can be dimethylated by sodium hydride and methyl iodide, and the gem-dimethyl ketone then fragments under Beckmann conditions to a nitrile bearing an isopropenyl group.

Organometallic and miscellaneous transformations. Rhodium hydride species add across alkenes by hydrometallation, and the hydride usually goes to the less hindered terminal carbon, which gives a linear alkyl metal. CO insertion into a metal–alkyl bond gives a metal acyl, and a strained cyclopropene can open at a ruthenium centre to a vinylcarbene. A methyl group on a metal carbonyl complex can be transferred to a metal-bound acyl carbon, and when a labelled CO is present, it is redistributed among the vacant sites in a statistical way. Ring-opening of a bicyclic alkene with ethylene generates a cis-1,2-divinylcyclobutane. The ring-contraction and Diels–Alder steps in complex sequences are usually planned backwards from the target, so that each intermediate has a clear FGI route.

Stereochemical bookkeeping. When assigning products, check three things: which face of a ring the reagent approaches (steric or directing-group control), whether an SN2 step has inverted the centre, and the CIP priority order when naming a configuration. For example, in a secondary alcohol bearing OAc, phenyl, methyl and H, the priority order is OAc > phenyl > methyl > H, and the configuration follows from the drawing with H pointing away. Syn-directed cyclopropanation by Simmons–Smith reagents is delivered to the face of the alkene bearing a free hydroxyl group, and the outcome of a second step must then be checked on the opposite face. For cyclic products, a cis-fused ring usually results from a concerted or intramolecular ring closure, while trans-fused products require special conditions.

Choosing a route. A good FGI answer names the reagent, explains why the chosen reagent acts on one group and not another, and checks the stereochemistry and the number of equivalents. Many incorrect options fail on one of these points, for example a reagent that reduces the wrong carbonyl, a step that adds a ring that the sequence never closes, or a product that requires an extra oxidation or reduction that the conditions do not supply.

Functional group interconversion

Functional group interconversion (FGI) means changing one functional group into another while keeping the carbon skeleton, or while adding a known carbon-extending step. In exam problems the task is to choose a reagent that acts on the intended group, at the intended position, with the intended stereochemistry. The following transformations are the ones to know.

• Hydroboration–oxidation — adds H and OH across an alkene with anti-Markovnikov regiochemistry. A bulky borane such as thexylborane stops after one addition to a terminal alkene, so it discriminates between a terminal C=C and a more hindered ring C=C.
• Kinetic enolate trapping — LDA at low temperature removes the proton from the less hindered α-carbon of an unsymmetrical ketone, and TMSCl traps the enolate as a silyl enol ether. Sulfenylation followed by oxidation to the sulfoxide and gentle heating removes PhSOH by syn elimination and gives an α,β-unsaturated ketone.
• Babler–Dauben oxidative transposition — a tertiary allylic alcohol is oxidised by PCC so that the double bond moves into conjugation with the new carbonyl group.
• Heck cyclisation — an intramolecular Pd(0)-catalysed carbopalladation onto an acrylate closes a five-membered ring, and β-hydride elimination followed by isomerisation can aromatise an indoline to an indole. Hydrogenolysis removes an N-benzyl group.
• Acetyl installation — a terminal alkyne adds to a ketone as an acetylide, and Hg(II)-catalysed hydration of the alkyne gives a methyl ketone by Markovnikov addition. Dehydration of the tertiary alcohol then gives the conjugated enone. Alternatively, an enol triflate couples with a vinyl ether under Pd catalysis, and acid hydrolysis of the enol ether unveils the acetyl group.
• Aldehyde from ketone (one-carbon homologation) — a methoxymethylene Wittig reagent converts C=O into an enol ether, and acid hydrolysis gives the aldehyde on the former carbonyl carbon. The Corey–Chaykovsky spiro-epoxide can be isomerised by a Lewis acid (Meinwald rearrangement) to an aldehyde. The Shapiro reaction of a tosylhydrazone gives a vinyllithium that is formylated by DMF.
• Dithiane as an acyl anion equivalent — a 2-lithio-1,3-dithiane adds to a carbonyl compound, and HgCl_2-mediated hydrolysis of the thioacetal releases a carbonyl group, but this gives an α-hydroxy carbonyl compound rather than a simple aldehyde.
• Ketone surrogates from Diels–Alder adducts — a nitroalkene dienophile gives a nitro-norbornene, which a titanium(III) reagent converts into a ketone. An α-acetoxy acrylonitrile gives a cyanohydrin acetate, and base hydrolysis releases the ketone.
• Beckmann-type fragmentation — when the carbon α to an oxime carbon can form a stable tertiary or quaternary carbocation, acid-promoted fragmentation gives a nitrile and a carbocation rather than a lactam.
• Zweifel-type halogenation — a hydroborated terminal alkyne gives an alkenylboronate, and halogenation with base-promoted 1,2-migration converts it into a vinyl halide with inversion of the alkene geometry.
• SmI_2 ketyl–olefin cyclisation — SmI_2 reduces a ketone to a ketyl radical, which adds 5-exo onto an enoate, forming a fused ring with a tertiary alcohol at the ring junction.
• Acetal protecting-group removal — a MEM ether is an acetal that can be cleaved by Lewis acids such as ZnBr_2 or by mild acid such as PPTS in tert-butanol. Benzyl-type hydrogenolysis and fluoride reagents do not remove an alkyl acetal.
• Wagner–Meerwein ring contraction — a secondary carbocation next to a quaternary carbon can trigger a 1,2-alkyl shift that contracts a six-membered ring to five and moves the cation onto the gem-dimethyl carbon. Loss of a proton then gives an alkene.
• Diene formation followed by Diels–Alder — a Heck reaction of a cyclic alkenyl triflate with a vinyl ether gives an alkoxy-substituted diene, which reacts with an acrylate in a regioselective [4+2] cycloaddition.

Two pairs are often confused. Acetylide addition followed by Hg-catalysed hydration gives a methyl ketone, whereas hydroboration of a terminal alkyne gives an aldehyde. The Corey–Chaykovsky spiro-epoxide route must be distinguished from direct Wittig homologation, since only the epoxide route requires a Lewis acid step to reach the aldehyde.

Grignard reagent additions

Grignard reagents and organolithium reagents add to the electrophilic carbon of a carbonyl group, giving alcohols after acidic work-up. Addition to an aldehyde gives a secondary alcohol, and addition to a ketone gives a tertiary alcohol. When the substrate already has a stereocentre next to the carbonyl group, the new carbinol centre forms as a mixture of diastereomers, and the ratio is set by steric control of the transition state.

For quantitative questions, the steps are: convert the mass of the starting material into moles using its molar mass, multiply by the fractional yield, and then split the product among the diastereomers according to their ratio. This gives the mass of each product.

The choice of organometallic also changes the outcome. Phenyllithium adds mainly in a 1,2-manner to a simple enone, whereas a diphenylcuprate adds conjugately to the β-carbon.

Organometallics are also prepared by directed metalation. A MOM ether can direct sec-butyllithium to the ortho position of an aromatic ring. Transmetalation with ZnCl_2 gives an arylzinc reagent, which undergoes Negishi cross-coupling with an aryl bromide under Pd(PPh_3)_4 catalysis, and the MOM group survives.

Carbonyl condensation reactions

Carbonyl condensation and related carbonyl additions form new carbon–carbon bonds by reacting an enolate or another carbon nucleophile with a carbonyl compound or an activated alkene. The following points are tested.

• Michael addition of a dithiane anion — 2-alkyl-1,3-dithianes are deprotonated by n-BuLi at C-2, and the anion adds conjugately to an α,β-unsaturated sulfoxide. The sulfinyl group is retained in the adduct until it is removed by heating, which gives a vinyl dithiane. Hydrolysis of the dithioacetal then releases a simple vinyl ketone.
• Sulfur ylides — a non-stabilised dimethylsulfonium methylide is a hard nucleophile that adds 1,2 to an enone carbonyl and gives an epoxide. A stabilised dimethylsulfoxonium methylide is softer and adds 1,4 to the enone, giving a cyclopropane fused to the ketone.
• Chiral auxiliary alkylation — an N-acyl oxazolidinone is deprotonated by LDA to a Z-enolate, in which lithium chelation fixes the geometry. A bulky substituent on the auxiliary blocks one face, so an alkyl halide attacks from the other face. LiAlH_4 then cleaves the imide to a primary alcohol and releases the auxiliary.

These reactions are often confused with simple alkylation. The stereochemical outcome of an auxiliary-controlled enolate alkylation depends on the enolate geometry, not on the alkyl halide alone, so the geometry must be specified before the face selectivity is assigned.

Electrophilic addition to alkenes

Electrophilic addition to a C=C bond starts with attack of an electrophile on the π electrons, giving a carbocation or a bridged onium ion, and the second step is attack of a nucleophile on that intermediate. Markovnikov regiochemistry follows from the more stable carbocation. Concerted additions behave differently, because they are stereospecific.

• Simmons–Smith cyclopropanation — the iodomethylzinc carbenoid generated from CH_2I_2 and a Zn–Cu couple adds a CH_2 unit across the double bond in a single concerted, syn step. The alkene geometry is therefore transferred directly to the ring, so a cis alkene gives a cis-disubstituted cyclopropane, and any stereocentre elsewhere in the molecule is untouched.
• Diels–Alder regiochemistry with electron-rich dienes — a diene carrying a donor group at C-1 has its most nucleophilic terminus at the far carbon (C-4). This terminus bonds to the β-carbon of an electron-poor dienophile, and the ring carbon next to the donor bonds to the α-carbon bearing the electron-withdrawing group. After hydrolysis of the silyl enol ether and loss of the donor group, a conjugated enone with the substituent at the γ-position results.