Radical addition that takes place by chain reaction naturally divides into three categories. These three are reactions in which (a) a carbohydrate radical adds to an unsaturated noncarbohydrate, (b) a noncarbohydrate radical adds to an unsaturated carbohydrate, and (c) both reaction participants are carbohydrates. Each group can be further divided into mechanistic type, a distinction based upon whether a reaction has a bimolecular or a unimolecular chain-transfer step.
A. Addition of Carbohydrate Radicals to Unsaturated Noncarbohydrates
1. Bimolecular Chain Transfer
The mechanism shown in Scheme 1 describes a reaction that takes place by bimolecular chain transfer. The factors that control regio- and stereoselectivity in these reactions often depend on the type of carbohydrate radical involved; therefore, reactions of radicals centered on different carbon atoms in pyranoid and furanoid rings and in open-chain structures will be discussed separately.
a. Pyranos-1-yl Radicals
(1). Radical Formation
A classic example of radical addition occurring by bimolecular chain transfer is shown in Scheme 5.1,2 In this reaction the tri-n-butyltin radical abstracts a bromine atom from the glycosyl bromide 1 to produce the corresponding pyranos-1-yl radical 2, an intermediate that then adds to acrylonitrile to give the adduct radicals 3 and 4. Lastly, each of these radicals (3 and 4) abstracts a hydrogen atom from Bu3SnH in the chain-transfer step to form the C‑glycosides 5 and 6, respectively, and produce the chain-carrying tri-n-butyltin radical. The radical 2 also can abstract a hydrogen atom from Bu3SnH (or another, available hydrogen-atom transfer) to give the simple reduction product 7. (Compounds such as 5 and 6 commonly are called C‑glycosides, or, more formally, 1-C-linked glycosyl derivatives.17)
The data given in Tables 1-5 show that the bromine-atom abstraction described in Scheme 5 represents the most common method for producing pyranos-1-yl radicals. These radicals also are formed from phenyl selenides18,19 (Scheme 818) and glycosyl tellurides (Scheme 920) and in sequential reactions beginning either with glycals21 (Scheme 1021a) or carbohydrate nitrates.22
(2). Protecting Group Electronegativity
A significant change in reactivity occurs when the O-acetyl groups in a pyranosyl bromide are replaced by O-benzoyl groups.5,6 Reactions of the tetrabenzoate 16 (Scheme 11) and the corresponding tetraacetate 1 (Scheme 5) illustrate this situation.5,6 Addition of the tetra-O-benzoyl radical 17 to unsaturated compounds (Table 5) has regio- and stereoselectivity reminiscent of additions of the tetra-O-acetyl radical 2 (Table 1), but simple reduction is a much more important process in the reactions of 17. Under conditions normally employed for radical addition (entry 2, Table 5), simple reduction is, in fact, the major reaction pathway for the tetrabenzoate 16 but only a minor one for the tetraacetate 1 (eq 4).
It is interesting that different acyloxy groups have such different effects on radical reactivity even when these groups are well-separated from the radical center. Replacing O‑acetyl with O-benzoyl groups (i.e., converting 2 into 17) appears to either reduce the nucleophilicity of the resulting pyranos-1-yl radical (17) or increase the ability of this radical to abstract a hydrogen atom from Bu3SnH or both. Since phenyl groups are more electron-withdrawing than methyl groups, it is reasonable to expect that the tetra-O-acetyl radical 2 should be more nucleophilic than the tetra-O-benzoyl radical 17. Because reduced nucleophilicity of a radical decreases its rate of addition to an electron-deficient double bond, this benzoyl for acetyl replacement should slow the addition process. Reduced nucleophilicity also should make hydrogen-atom abstraction from Bu3SnH more rapid because the polarity of the tin-hydrogen bond should favor a transition state where the reacting radical is less nucleophilic (Scheme 12). The decrease in radical nucleophilicity in going from 2 to 17 correlates with an increase in the amount of hydrogen-atom abstraction at the expense of radical addition (eq 4).
When one considers the effect of O-acetyl and O-benzoyl groups on pyranos-1‑yl radical reactivity (eq 4), it is reasonable to expect that introducing a powerfully electron-withdrawing group near a radical center should substantially decrease the ability of the radical to add to an electron-deficient double bond and correspondingly increase its ability to abstract a hydrogen atom from Bu3SnH. This type of change in reactivity can be seen in eq 4, where the effect of 2-O-acetyl and 2-O-benzoyl groups is compared to that of a 2-keto group. The electron-withdrawing nature of the 2-keto group is great enough that pyranos-1-yl radical addition to acrylonitrile completely stops; only hydrogen-atom abstraction is observed.23
(3). Less Electron-Deficient Multiple Bonds
The addition reactions discussed thus far have been of pyranos-1-yl radicals adding to electron-deficient double bonds. Successful addition becomes problematic when reaction involves a compound with a multiple bond that is less electron-deficient than those in the unsaturated reactants shown in Tables 1-5. For example, although compounds 20 and 21 both contain electron-withdrawing groups, for neither of these are the groups close enough to the double bond to render it sufficiently electron deficient to undergo addition of a nucleophilic radical in the presence of Bu3SnH (eq 5).6 For reaction involving either 20 or 21, simple reduction is much faster than radical addition. If, however, only ineffective hydrogen-atom transfers are present in the reaction mixture, as is the case in the reaction shown in eq 6, where the tri-n-butyltin radical is generated by photolysis of hexabutylditin, the rate of simple reduction decreases and radical addition is observed.24
Styrene is a compound that is on the borderline between reactive and unreactive in pyranos-1-yl radical addition in the presence of Bu3SnH. In some reactions the double bond in styrene is not sufficiently electron-deficient for addition to be competitive with simple reduction,25,26 but in other cases addition does occur.19a,27 One example is shown in eq 7,27 where even though there is radical addition, the major product arises from simple reduction. The addition that does occur only takes place when Et3B is present in the reaction mixture, a finding that indicates coordination with Et3B in the transition-state structure may be critical in lowering the energy needed for addition to occur. As in the reaction shown in eq 7, simple reduction is the major pathway in other reactions where addition to styrene takes place.19a
b. Pyranosyl Radicals Centered at C-2, C-3, or C-4
A radical centered at C-2, C-3, or C-4 in a pyranoid ring undergoes regiospecific, stereoselective addition to a compound with an electron-deficient double bond.15,28,29 Because the kinetic anomeric effect is not operative in these reactions, stereoselectivity is determined primarily by steric effects.15 Since a radical centered at C-2, C-3, or C-4 in a pyranoid ring normally exists in a chair conformation, reaction that produces an equatorial substituent is likely to have a less hindered reaction pathway than one leading to an axial substituent. As the steric requirements of unsaturated reactants increase, reaction stereoselectivity also increases; thus, the radical 13 produces progressively greater percentages of products with equatorial substituents as the size and number of the groups attached to the double bond in the unsaturated reactants increase (Scheme 6).15
c. Pyranos-5-yl Radicals
The reactions of pyranos-5-yl radicals are subject to the kinetic anomeric effect in the same way as are those of pyranos-1-yl radicals. It is surprising, therefore, to find that the radical 23, derived from the bromide 22, adds to acrylonitrile with very little stereoselectivity (Scheme 13).30 The reason for this near lack of selectivity is not known, but the absence of a C-5 ring substituent makes the pyranos-5-yl radical 23 more conformationally mobile than radicals that have such a substituent (i.e., most pyranos-1-yl radicals) and raises the possibility that epimeric products 24 and 25 arise by reaction from different conformations.30 [A similar lack of stereoselectivity occurs during radical addition by pyranos-1-yl radicals that lack a conformation-stabilizing, C-5 substituent (see eq 16).]
(see eq 16 on p 385)
d. Radicals Centered at C-6 in Hexopyranoses and Hexopyranosides
Radicals centered at C-6 in hexopyranoses and hexopyranosides range in stability from those that are primary to ones that are stabilized by an attached oxygen atom. Both of these radical types add to unsaturated compounds.22,31–33 In the reaction shown in eq 8 the primary radical generated from the iodide 26 adds in good yield to the vinyl sulfone 27.31 The reaction in Scheme 14 pictures the addition of the oxygen-stabilized radical 28 to ethyl acrylate as one step in a sequential reaction.22 In this reaction the less hindered approach of ethyl acrylate to the convex face of the radical 28 causes highly stereoselective reaction to occur.
e. Furanosyl Radicals
Addition reactions involving furanosyl radicals are as broadly based as those of pyranosyl radicals; thus, reaction occurs with radicals centered at C‑1,34–40 C-3,28,41–44 C-4,45–50 C-5,51–55 and C-656 in structures containing furanoid rings. To appreciate the reactivity, modes of formation, and reaction stereoselectivity of these radicals, it is useful to consider a few typical examples.
The reaction shown in Scheme 15 involves the stereoselective addition of the furanos-1-yl radical 29 to the vinyl sulfone 30.34,36 Although the kinetic anomeric effect usually is the primary factor in determining reaction stereoselectivity for pyranos-1-yl radicals, its role in reaction of funanos-1-yl radicals is less important because a five-membered ring is closer to planarity. The kinetic anomeric effect would not favor reaction at either face of a ring with a planar geometry. Also, because five-membered rings tend to be more conformationally mobile that six-membered ones, reaction from different conformations of a furanos-1-yl radical are more likely than reactions from different conformations of a pyranos-1-yl radical. If reactions from multiple conformations are taking place, the possibility for formation of different stereoisomers increases. Stereoselectivity in the reactions of furanos-1-yl radicals (such as 29), therefore, should depend to a large extent on ring-substituent shielding of the radical center.
Shielding by ring substituents also explains reaction stereoselectivity of furanosyl radicals centered on atoms other than C-1. Thus, positioning of ring substituents accounts for product stereochemistry in the reactions shown in eq 941 (with an intermediate furanos-3-yl radical) and eq 1048 (with an intermediate furanos-4-yl radical). Methods for generating furanosyl radicals vary widely. Furanos-4-yl radicals often are generated from esters of N-hydroxypyridine-2-thione.45–50 The most common method for forming radicals centered at C-5 in addition reactions is by treatment of a deoxyiodo sugar with a zinc–copper couple (eq 11).51
f. Open-Chain and Related Radicals
When an open-chain, carbohydrate radical adds to an unsaturated compound, the normal result is to lengthen the carbon atom chain;57–63 for example, in the reaction shown in Scheme 16 the chain length increases by two carbon atoms.58 (In this reaction the chain decreases by one atom during radical formation and then increases by three as a result of radical addition.) In most reactions of this type the carbohydrate radical is generated from an ester of N-hydroxypyridine-2-thione. This type of ester also can be used to generate radicals in the aglycon portion of a glycoside. Subsequent addition of such a radical to an unsaturated compound not only lengthens and modifies the aglycon chain, but the carbohydrate moiety also acts as a chiral auxiliary in affecting reaction stereoselectivity (eq 12).64–66
2. Unimolecular Chain Transfer
There are several noteworthy advantages to unimolecular, chain-transfer reactions (Scheme 2). First, since a hydrogen-atom transfer, such as Bu3SnH, is not added to the reaction mixture, hydrogen-atom abstraction by R· to give the simple reduction product RH is less likely to occur. A reduced rate of hydrogen-atom abstraction means that addition to double bonds that are not electron deficient sometimes is fast enough to be observed. Also, when fragmentation of the adduct radical is rapid, as is often the case, fewer reactions of any type are able to compete with this product-forming step.
Unimolecular chain-transfer reactions form unsaturated products that have attractive possibilities for further transformation.67 Easily the most common reaction of this type involves addition of a carbohydrate radical to an allylstannane (eq 13).68 The reactions shown in Scheme 17 illustrate the way in which reduced hydrogen-atom abstraction and rapid radical fragmentation work together to allow an addition reaction to occur by unimolecular chain transfer when a similar reaction by bimolecular chain transfer is thwarted by simple reduction.23
a. Additions to Allylic Stannanes
(1). Pyranos-1-yl Radicals
Pyranos-1-yl radicals that add to allylic stannanes can be generated from glycosyl chlorides17,69–71 and bromides,23,69,72–77 phenyl thioglycosides,78,79 and glycosyl xanthates.17 The stereoselectivity of the addition process usually is controlled by the kinetic anomeric effect; thus, reaction between the glycosyl chloride 31 and allyltributylstannane preferentially produces the α‑C-glycoside 32 (eq 14).17,69 In the unusual event that the C-2 substituent is large enough to block the approach of an unsaturated compound to the α face of the pyranoid ring, the stereoselectivity of the reaction changes (eq 1569).17,69
In addition to the shielding effect large substituents have on reaction stereoselectivity, conformational mobility also has a role in determining this selectivity. The kinetic anomeric effect predicts a highly stereoselective reaction when a pyranos-1-yl radical exists primarily in a single conformation, which is the typical situation for hexopyranosyl radicals. If a radical is conformationally more mobile, as is the case for many pentopyranosyl radicals, including the one participating in the reaction shown in eq 16,72 the kinetic anomeric effect still is operative, but stereoselectivity is likely to be reduced (or in this case become nonexistent) because readily accessible conformers may produce different stereoisomers.
(2). Pyranosyl Radicals Centered on C-2, C-3, C-4, or C-6
Radical allylation causes replacement of substituents at C‑2,80,81 C-3,82,83 C‑4,68,84–88 and C‑668,82,84,89–91 in pyranoid rings with allyl or substituted allyl groups. Although the intermediate radicals in these reactions usually are generated by halogen-atom abstraction, compounds with O-thiocarbonyl groups also function as radical precursors.84,92 In addition, carbon-centered radicals capable of reacting with an allylstannane can be produced by cyclization reaction (Scheme 1893).93,94
Another method for forming a pyranosyl radical capable of participating in radical allylation is by electron transfer. Electron transfer usually involves a transition-metal complex as either the electron donor or acceptor (such reactions are discussed in Section IV), but this type of reaction also is believed to take place between Bu3Sn· and the keto group of an α-benzoyloxy ketone.95 Electron transfer from Bu3Sn· to the ketobenzoate 33 provides a pathway for forming a radical that then reacts with allyltributylstannane to complete the allylation process (Scheme 19).96
(3). Furanosyl Radicals
Unimolecular chain transfer also takes place in reactions that replace halogen and O-thiocarbonyl substituents with allyl (or allylic) groups in compounds with furanoid rings. Such reactions occur with both simple monosaccharides89,97–101 and with nucleosides.102–111 Among monosaccharides, replacement is known to take place at C-2,100 C-3,112 C-4,97 and C-5.89,113 For nucleosides, allyl groups replace halogen atoms,104,110 selenophenyl groups,108 and O-thiocarbonyl groups103–107,109,111,112 at the C‑2',104,110,111 C-3',105,107,109 and C-5' positions.108 A typical example is shown in eq 17.105
A substituted furanoid ring can become a chiral auxiliary when reaction takes place at a carbon atom in a substituent group. The reaction shown in Scheme 20, for example, describes a highly stereoselective allylation in which a soluble, polymer-supported furanose derivative serves as a chiral auxiliary.101 In addition to the striking stereoselectivity of this reaction, use of a polymer-supported reagent allows complete separation of the often-troublesome tin residues from the reaction product.
(4). Limitations on Radical Allylation
Significant restrictions exist on the stannanes that are synthetically useful in radical allylation84,90 because attaching a methyl (or related) group to C-1 or C-3 in an allyl stannane can complicate or even prevent reaction. One example of the difficulties encountered is illustrated by the behavior of the stannane 34, for which attempted addition reaction leads only to hydrogen-atom abstraction (eq 18).84 Even when addition does occur to an allylic system, hydrogen-atom abstraction can be a significant competing process (eq 19).90 Another type of difficulty is seen in the reaction shown in eq 20, where migration of the tri-n-butyltin group leads to formation of a mixture of stannanes; thus, 1,3-tin migration reduces the number of stannanes that can be effective reactants in an allylation process. As long as hydrogen-atom abstraction is not significant, compounds with substituents at the 2-position remain viable reactants if 1,3-tin migration is degenerate (eq 21). In the reaction shown in eq 22, for example, neither hydrogen-atom abstraction nor 1,3-tin migration prevent addition from occurring in good yield.81a
b. Additions to Vinylic Stannanes
One characteristic of the addition-elimination sequence that is central to radical allylation is that the double bond in the starting material shifts to a new position in the product (Scheme 2). Such a change in position does not occur in all addition-elimination reactions; in particular, reactions of vinylic stannanes reform the double bond in its original location (Scheme 21). Replacement of a halogen atom, as pictured in the reaction shown in eq 23,32 and an O-thiocarbonyl group,14,114 as described in eq 3,14 with a substituent containing a carbon–carbon double bond represent addition-elimination reactions that return a double bond to its original position. Such reactions can extend a carbon-atom chain32,115 or introduce a substituent at a ring carbon atom.14,32,114–119 These reactions are regiospecific and, when shielding groups force approach to a particular ring face, tend to be highly stereoselective.
c. Additions to Unsaturated Sulfones and Sulfides
There are several addition reactions in which an unsaturated sulfide24,120 or sulfone24,121–124 occupies the reactant role usually held by a stannane. The principal advantage in using sulfides or the more reactive sulfones is that the problems associated with the presence of tin-containing reagents can be avoided. Under some conditions sulfur-containing compounds are not sufficiently reactive,6 but when reaction does take place, it follows familiar patterns. If the kinetic anomeric effect is operative and a single conformation is dominant, reaction is highly stereoselective (eq 24).124 In situations where the kinetic anomeric effect cannot operate, sometimes stereoselectivity is low or completely lacking, an indication that group shielding does not strongly favor reaction on either face of the ring system involved (eq 25).122
B. Addition to Unsaturated Carbohydrates by Noncarbohydrate Radicals
1. Bimolecular Chain Transfer
a. α,β-Unsaturated Carbonyl Compounds
Addition of a radical to a carbohydrate containing an α,β-unsaturated carbonyl group is a regiospecific, stereoselective process that mechanistically follows the bimolecular, chain-transfer pathway (Scheme 1).125–131 The regioselectivity of this type of reaction can be explained in terms of frontier-orbital interactions and in terms of radical philicity. Both of these approaches are useful in understanding reactions with early transition states. In the reaction shown in eq 26,125 for example, the frontier-orbital interactions and the atomic orbital coefficients pictured in Figure 1 (and discussed earlier in this Chapter in Section II.B.2) predict the observed addition to the β-carbon atom of the α,β-unsaturated ketone. Radical philicity makes a similar predication because the nucleophilic hydroxymethyl radical would be expected to add to the less electron-rich carbon atom in the unsaturated ketone (Figure 2).
The stereoselectivity of radical addition to a carbohydrate containing an α,β-unsaturated carbonyl group is determined primarily by steric effects. In the reaction shown in eq 27 the hydroxymethyl radical adds preferentially to the α face of the pyranoid ring due to the shielding of the β face by the CH2OSiMe2t-Bu substituent at C-5.127 Steric effects associated with the ring substituent created by radical addition also contribute to the reaction stereoselectivity; thus, in the reactions described in Scheme 22, the relative amounts of products with the CH2R substituent projecting onto the less-hindered, β face of the ring become greater as the steric size of the R group increases.130
b. Enol Ethers
(1). Reactions with Heteroatom-Centered Radicals
The carbon-carbon double bond in an enol ether is electron-rich due to delocalization of nonbonded electrons from the attached oxygen atom (Figure 2). Addition reactions to enol ethers often begin with a radical centered on an electronegative sulfur134–138 atom or a less electronegative phosphorous132,133 atom. When a heteroatom-centered radical comes into contact with an unsaturated compound, the familiar possibilities of radical addition and hydrogen-atom abstraction are in competition. In general, the more electronegative, less polarizable radicals (e.g., those centered on oxygen and nitrogen atoms) abstract hydrogen atoms, and less electronegative, more polarizable radicals (e.g., those centered on phosphorous and sulfur atoms) add to multiple bonds. A typical example of radical addition is given in the reaction shown in eq 28,134 where AcS· adds to an enol ether to form a β-C-glycoside in a reaction that echoes stereoselectively the one shown in Scheme 23; that is, the product formed is the one expected from the kinetic anomeric effect controlling the hydrogen-atom abstraction that completes the reaction. In another example, a phosphorous-centered radical adds to the difluro enol ether 41 to form a mixture of products greatly favoring the β-C-glycoside 42 (eq 29).133
The structure of an enol ether is critical to the mechanism of its reaction with a sulfur-centered radical. This fact becomes evident when comparing the reactions shown in eq 28 and Scheme 24. The reaction pictured in eq 28 is a typical radical addition with a bimolecular, chain-transfer step. Although the reaction shown in Scheme 24 begins in a similar way by addition of a sulfur-centered radical to the unsaturated nucleoside 43 to give the adduct radical 44, a quite different mechanistic pathway then emerges,135–137 The defining feature of this new pathway is that the adduct radical 44 expels (EtO)2PO2- to generate the radical cation 45, which then accepts a electron from C6H5S- (or C6H5SH) to form the observed product.135–137 In this reaction chain transfer is a bimolecular process but one that takes place by an electron transfer reaction rather by group or atom abstraction.
(2). Reactions with Carbon-Centered Radicals
Carbon-centered radicals also add to enol ethers but only when these normally nucleophilic radicals are rendered electrophilic by one or more powerfully electron-withdrawing substituents attached to the radical center.139–144 A carbon-centered radical becomes sufficiently electrophilic to add to a carbohydrate containing an enol-ether linkage when alkoxycarbonyl substituents are attached to the radical center. Radicals of this type can be formed from ClCH(CO2CH3)2 by halogen-atom abstraction, as occurs in the reaction shown in Scheme 23.140 The β-C-glycoside 40 is the product expected from the kinetic anomeric effect controlling the reactions of the intermediate radical 38. Additional examples of carbon-centered radicals adding to glycals are found both later in this Chapter (Section IV.B) during discussion of nonchain reactions and in Chapter 21.
c. Compounds with Double Bonds Lacking Activating Substituents
Compounds in which the π system of a carbon–carbon double bond does not have an electron-donating or electron-withdrawing substituent attached still are able to react with radicals centered on sulfur145,146 or phosphorous16,147 atoms. An example involving a phosphorous-centered radical is shown in Scheme 7.16 The stereoselectivity in this reaction is determined in the same way as many reactions where hydrogen-atom abstraction is the final step, that is, by approach of the hydrogen-atom transfer (in this case H2P(=O)O-) to the less hindered face of the ring system. Electrophilic, carbon-centered radicals are similar to sulfur- and phosphorous-centered ones in that they too can add to a double bond lacking an activating substituent.148
2. Unimolecular Chain Transfer
Carbohydrates that have a double bond with a sulfur atom either directly attached149–153 or in an allylic position154,155 undergo reactions that take place by unimolecular chain transfer. The reaction shown in Scheme 25 is one where the sulfur atom is attached to the double bond.151 In this reaction the tri-n-butyltin radical adds to the double bond to give an adduct radical that then reforms the π system by fragmenting the carbon–sulfur bond. The reaction shown in eq 30 is one where Bu3Sn· adds to an unsaturated compound in which the double bond has an allylic, sulfur-containing substituent.154 This addition is followed by carbon-sulfur bond fragmentation that repositions the double bond.
C. Addition of Carbohydrate Radicals to Unsaturated Carbohydrates
1. Bimolecular Chain Transfer
There are reactions in which a carbohydrate radical adds to an unsaturated carbohydrate.130,156–165 Most of these reactions, including that shown in eq 31,156 involve the addition of a pyranos-1-yl radical to an α,β-unsaturated carbonyl compound. Other possibilities include addition of a primary radical to either an α,β-unsaturated ketone160 or a flourine-substituted enol ether.161 Radical addition also links two saccharide units through a phosphorous-containing bridge (eq 32).16
2. Unimolecular Chain Transfer
A modification of carbohydrate radical addition to an unsaturated compound is found in the reaction shown in eq 33, where a carbohydrate sulfone is converted into two, allyl-substituted carbohydrates (48α and 48β).166,167 The two propagation sequences operating in this reaction are shown in Scheme 26. Addition of the carbohydrate radical 49 to the carbohydrate sulfone 46 begins the first sequence. The second step is a unimolecular, β-fragmentation reaction that forms the products 48α and 48β and a sulfonyl radical. Decomposition of this radical produces the carbohydrate radical 49 needed to begin a new sequence. [Although this reaction sequence involves a carbohydrate radical (49) adding to an unsaturated carbohydrate (46), this joining together is temporary. The two carbohydrate units do not remain linked in the final products (48α and 48β).
The advantage that the aryl sulfone 47 brings to this reaction is that when used in excess, it is an unsaturated compound that protects the products 48α and 48β from further reaction without itself causing new products to be formed. The primary reason 47 does not lead to new product formation is that ArSO2·, produced in the second propagation sequence in Scheme 26, does not fragment to give Ar·, a radical that would form new products. (An added advantage to this procedure is that it avoids the use of tin-containing compounds.)
D. Addition to Compounds with Carbon–Carbon Triple Bonds
Although the majority of radical addition reactions of carbohydrates involve compounds with carbon–carbon double bonds, some take place with compounds containing carbon–carbon triple bonds.110,168-178 One reason for fewer reactions of compounds with triple bonds is that the pool of potential substrates is reduced by the inability of a pyranoid or furanoid ring to accommodate a triple bond internally. Also affecting addition to a compound with a triple bond is the fact that such a reaction produces a high energy, vinylic, adduct radical. To the extent that the energy of an adduct radical is reflected in the transition state for the reaction leading to its formation, the rate of radical addition to a compound with a triple bond will be smaller than that for comparable addition to a compound with a double bond.179 A reduced rate of addition means that simple reduction is more competitive.
An effective way to minimize the impact of simple reduction on radical addition (including addition in which a compound with a triple bond is one of the reactants) is to eliminate the hydrogen-atom transfer (e.g., Bu3SnH) from the reaction mixture by selecting reactions that take place by unimolecular chain transfer.110,168,170 In practice this often means choosing a reaction between a carbohydrate radical and an unsaturated stannane (e.g., 50) (eq 34).168 Although removing Bu3SnH from a reaction mixture reduces the possibility for simple reduction, it does not eliminate this reaction completely because a stannane such as 50 also can function as a hydrogen-atom transfer, although a less effective one than Bu3SnH.
When Bu3Sn· adds to a carbohydrate containing a triple bond, the product is a vinylic stannane.172–178 Simple addition,180 such as that shown in the reaction in eq 35,178 does take place, but sequential reaction is more common. In sequential reactions the vinylic radical produced by the initial addition undergoes radical cyclization (eq 36).172
E. Maximizing Radical Addition
As mentioned several times in this chapter (Sections III.A.1.a, III.A.2.a, and III.D), addition reactions that take place by bimolecular chain transfer (Scheme 1) are forced to compete with hydrogen-atom abstraction that leads to simple reduction. The data in Tables 1 and 2 and particularly that in Table 5 reinforce this observation by showing simple reduction to be a significant obstacle to high-yield radical addition. The procedures described in the next paragraph often are adopted to improve radical addition by discouraging simple reduction.
The opportunity for radical addition increases under the following conditions: (a) the unsaturated reactant is present in large excess, (b) the concentration of the hydrogen-atom transfer is maintained at a low level, and (c) the unsaturated reactant is altered in structure to be better able to react rapidly with a carbohydrate radical. Although each of these changes favors radical addition, they are not all always used because each has a drawback. Having an unsaturated reactant present in large excess (e.g., 10-20 equiv) effectively limits the possible reactants to small, nonprecious compounds (e.g., acrylonitrile, methyl acrylate, and the similar compounds listed in Table 1). Since lowering the concentration of Bu3SnH (typically done by its slow addition to the reaction mixture) reduces the rate of hydrogen-atom abstraction by the carbohydrate radical (eq 37) without changing the rate of radical addition, maintaining a low Bu3SnH concentration makes simple reduction less competitive; however, reduced Bu3SnH concentration means that other reactions of the adduct radical, such as telomer formation, become more competitive. Since nearly all carbohydrate radicals are nucleophilic, selecting an unsaturated reactant with an electron-deficient multiple bond increases the reaction rate but reduces the pool of possible reactants.