Rearrangements and neighboring group effects

Chapter 28 

Rearrangements and o Neighboring Group Effects Nonclassical Ions

Rearrangements and neighboring group effects: intramolecular audiophilic attack

• Carbonium ions, we know, can rearrange through migration of an organic group or a hydrogen atom, with its pair of electrons, to the electron-deficient carbon. Indeed, when carbonium ions were first postulated as reactive intermediates (p. 160), it was to account for rearrangements of a particular kind. Such rearrange. the driving force behind all carbonium ion reactions is the need to provide electrons to the electron-deficient carbon. When an electron-deficient carbon is generated, a near-by group may help to relieve this deficiency. It may, of course, remain in place and release electrons through the molecular framework, inductively or by resonance. Or and this is what we are concerned with here it may actually carry the electrons to where they are needed. Other atoms besides carbon can be electron-deficient in particular, nitrogen and oxygen and they, too, can get electrons through rearrangement. The most important class of molecular rearrangements is that involving 1,2-shifts to electron-deficient atoms. It is the kind of rearrangement that we shall deal with in this chastens still provide the best single clue that we are dealing with a carbonium ion reaction.

• An electron-deficient carbon is most commonly generated by the departure of a leaving group which takes the bonding electrons with it. The migrating group is, of course, a nucleophile, and so a rearrangement of this sort of v amounts to intramolecular audiophilic substitution. Now, as we have seen, nucleophilic substitution can be of two kinds, S^2 and SN! - Exactly the same possibilities exist for a re-arrangement: it can be SN2-like, with the migrating group helping to push out the leaving group in a single-step reaction; or it can be SN l-like, with the migrating.

 

• Neighboring group effects involve the same basic process as rearrangement. Indeed, in many cases there is rearrangement, but it is hidden. What we see on the surface may be this:

• If a neighboring group is to form a bridged cation, it must have electrons to form the extra bond. These may be unshared pairs on atoms like sulfur, nitrogen, oxygen, or bromine; * electrons of a double bond or aromatic ring; or even, in some cases, an electron. 

• In making its nucleophilic attack, a neighboring group competes with outside molecules that are often intrinsically much stronger nucleophiles. Yet the evidence clearly shows that the neighboring group enjoys for its nucleophilic power a tremendous advantage over these outside nucleophiles. Why is this? The answer is quite simple: because it is there.

• The neighboring group is there, in the same molecule, poised in the proper position for attack. It does not have to wait until its path happens to cross that of the substrate; its "effective concentration" is extremely high. It does not have to give up precious freedom of motion (translational entropy) when it becomes locked into a transition state. Between it and the reaction center there are no tightly clinging solvent molecules that must be stripped away as reaction takes place. Finally, the electronic reorganization changes in overlap that accompanies reaction undoubtedly happens more easily in this cyclic system.

• Enzymes function by accelerating, very specifically, rates of the organic reactions involved in life processes. They evidently do this by bringing reactants together into exactly the right positions for reaction to occur. Underlying much enzyme activity, it appears, are what amount to neighboring group effects.

Hofmann rearrangement. Migration to electron-deficient nitrogen

• Let us begin with a reaction that we encountered earlier as a method of synthesis of amines: the Hofmann degradation of amides. Whatever the mechanism of the reaction, it is clear that rearrangement occurs, since the group joined to carbonyl carbon in the amide is found joined to nitrogen in the product.

• The reaction is believed to proceed by the following steps:

• Step (1) is the halogenation of an amide. This is a known reaction, an Haloimide being isolated if no base is present. Furthermore, if the N-haloimide isolated in this way is then treated with base, it is converted into the amine. Step (2) is the abstraction of a hydrogen ion by hydroxide ion. This is reasonable behavior for hydroxide ion, especially since the presence of the electron withdrawing bromine increases the acidity other amide. Unstable salts have actually been isolated in certain of these reactions. Step (3) involves the separation of a halide ion, which leaves behind an electron-deficient nitrogen atom. In Step (4) the actual rearrangement occurs. Steps (3) and (4) are generally.

HOFMANN REARRANGEMENT. INTRA- OR INTERMOLECULAR?

• believed to occur simultaneously, the attachment of R to nitrogen helping to push out halide ion. That is, migration is SN2-like, and provides an chimeric assistance. Step (5) is the hydrolysis of an isocyanate (R N - C=O) to form an amine and carbonate ion. This is a known reaction of isocyanates. If the Hofmann degradation is carried out in the absence of water, an isocyanate can actually be isolated.

• Like the rearrangement of carbonium ions that we have already encountered (Sec. 5.22), the Hofmann rearrangement involves a 1,2-shift. In the rearrangement of carbonium ions a group migrates with its electrons to an electron-deficient carbon; in the present reaction the group migrates with its electrons to an electron-deficient nitrogen. We consider nitrogen to be electron-deficient even though it probably loses electrons to bromide ion while migration takes place, rather than.

• The strongest support for the mechanism just outlined is the fact that many of the proposed intermediates have been isolated, and that these intermediates have been shown to yield the products of the Hofmann degradation. The mechanizing is also supported by the fact that analogous mechanisms account satisfactorily for observations made on a large number of related rearrangements. Furthermore, the actual rearrangement step fits the broad pattern of 1,2-shifts to electrondeficient atoms.

• In addition to evidence indicating what the various steps in the Hofmann degradation are, there is also evidence that gives us a rather intimate view of just how the rearrangement step takes place. In following sections, we shall see what some of that evidence is. We shall be interested in this not just for what it tells us about the Hofmann degradation, but because it will give us an idea of the kind of thing that can be done in studying rearrangements of many.

Hofmann rearrangement. Stereochemistry at the migrating group

• When optically active a-phenyl propionamide undergoes the Hofmann degradation, a-phenylethylamine of the same configuration and of essentially the same optical purity is obtained:

• Rearrangement proceeds with complete retention of configuration about the chiral center of the migrating up.

• These results tell us two things. First, nitrogen takes the same relative position on the chiral carbon that was originally occupied by the carbonyl carbon. Second, the chiral carbon does not break away from the carbonyl carbon until it has started to attach itself to nitrogen. If the group were actually to become free during its migration, we would expect considerable loss of configuration and hence a partially racemic product. (If the group were to become free really free we would expect reaction to be, in part, intermolecular, also contrary to fact.) We may picture the migrating group as moving from carbon to nitrogen via a transition state, I, in which carbon is pentavalent: 

• The migrating group steps from atom to atom; it does not jump. There is much evidence to suggest that the stereochemistry of all 1,2-shifts has this common feature: complete retention of configuration in the migrating group.

Hofmann rearrangement. Timing of the steps

• We said that steps (3) and (4) of the mechanism are believed to be simultaneous, that is, that loss of bromide ion and migration occur in the same: 

• One reason for believing this is simply the anticipated difficulty of forming a highly unstable intermediate in which an electronegative element like nitrogen has only a sextet of electrons. Such a particle should be even less stable than primary carbocations, and those, we know, are seldom formed; reaction takes the easier, $^2- like path. Another reason is the effect of structure on rate of reaction. Let us examine this second.

• When the migrating group is aryl, the rate of the Hofmann degradation is increased by the presence of electron-releasing substituents in the aromatic ring; thus substituted benzarones show the following order of reactivity:

•  Now, how could electron release speed up Hofmann degradation? One way could be through its effect on the rate of migration. Migration of an alkyl group must involve a transition state containing pentavalent carbon, like I in the preceding section. Migration of an aryl group, on the other hand, takes place via structure like V. This structure is a familiar one; from the standpoint of the migrating aryl group, rearrangement is simply electrophilic aromatic substitution, with the electron-deficient atom nitrogen, in this case acting as the attacking reagent. In at least some rearrangements, as we shall see, there is evidence that structures.

• like V are actual intermediate compounds, as in the ordinary kind of electrophilic aromatic substitution (Sec. 11.16). Electron-releasing groups disperse the developing charge on the aromatic ring and thus speed up formation of V. Viewed in this way, substituents affect the rate of rearrangement the migratory aptitude of an aryl group in exactly the same way as they affect the rate of aromatic nitration, halogenation, or sulfonation. (As we shall see, however, conformational effects can sometimes completely outweigh these electronic effects.)

• There is another way in which electron release might be speeding up reaction: by speeding up formation of the electron-deficient species in equation (3). But the observed effect is a strong one, and more consistent with the development of the positive charge in the ring itself, as during rearrangement.

Rearrangement of hydroperoxides. Migration to electron-deficient oxygen

• The phenyl group is joined to carbon in the hydroperoxide and to oxygen in phenol: clearly, rearrangement takes place. This time, it involves a 1,2-shift to electron deficient oxygen.

• Acid converts (step 1) the peroxide I into the protonated peroxide, which loses (step 2) a molecule of water to form an intermediate in which oxygen bears only six electrons. A 1,2-shift of the phenyl group from carbon to electron-deficient oxygen yields (step 3) the "carbonium" ion II, which reacts with water to yield (step 4) the hydroxy compound III. Compound HI is a hemi-acetal (Sec. 19.15) which breaks down (step 5) to give phenol and.

• Every step of the reaction involves chemistry with which we are already quite familiar: protonation of a hydroxy compound with subsequent ionization to leave an electron-deficient particle; a 1,2-shift to an electron-deficient atom; reaction of a carbonium ion with water to yield a hydroxy compound; decomposition of a hemi-acetal. In studying organic chemistry, we encounter many new things; but much of what seems new is found to fit into old familiar patterns of behavior.

Rearrangement of hydroperoxides. Migratory aptitude

• The rearrangement of hydroperoxides lets us see something that the Hofmann rearrangement could riot: the preferential migration of one group rather than another. That .is, we can observe the relative speeds of migration the relative migratory aptitudes of two groups, not as a difference in rate of reaction, but as a difference in the product obtained. In cumene hydroperoxide, for example, any one of three groups could migrate: phenyl and two methyl's. If, instead of phenyl 

• methyl was to migrate, reaction would be expected to yield methanol and acetophenone. Actually, phenol and acetone are formed quantitatively, showing that a phenyl group migrates much faster than a methyl. 

• It is generally true in 1,2-shifts that aryl groups have greater migratory aptitudes than alkyl groups. We can see why this should be so. Migration of an alkyl group must involve a transition state containing pentavalent carbon (IV). Migration of an aryl group, on the other hand, takes place via a structure of the biennium ion type (V); transition state or actual intermediate, V clearly offers an easier path for migration than does IV.

Pinacol rearrangement Migration to electron-deficient carbon

• Upon treatment with mineral acids, 2,3-dimethyl-2,3-butanediol (often called pinacol) is converted into methyl tert-butyl ketone (often called pinacolone). The glycol undergoes dehydration, and in such a way that rearrangement of the carbon skeleton occurs. Other glycols undergo analogous reactions, which are known collectively .as pinacol rearrangements.

• The pinacol rearrangement is believed to involve two important steps: (1) loss of water from the protonated glycol to form a carbonium Iori; and (2) rearrangement of the carbonium ion by a 1,2-shift to yield the protonated ketone.

• When the groups attached to the carbon atoms bearing OH differ from one another, the pinacol rearrangement can conceivably give rise to more than one compound. The product actually obtained is determined (a) by which OH group is lost in step (1), and then (b) by which group migrates in step (2) to the electron deficient carbon thus formed. For example, let us consider the rearrangement of l-phenyl-l,2-propanediol. The structure of the product actually obtained, methyl benzyl ketone, indicates that the benzyl carbonium ion (I) is formed in preference to the secondary carbonium ion (II), and that H migrates in preference to -CH3.

Pinacoid deamination. Conformational effects

• Unlike their aromatic counterparts, however, these diazonium ions are extremely unstable, and lose nitrogen rapidly to give products that strongly suggest intermediate formation of carbonium ions.

• If such an amino group is located alpha to a hydroxyl group, then treatment with nitrous acid causes a reaction closely related to the pinacol rearrangement, pina colic deamination':

• This system permits many studies not possible with pinacols, since here theft electron-deficiency is generated at a pre-determined position: at the carbon that held the amino group.

• Let us examine the stereochemistry of pinacolyl deamination in some detail. , In this we shall see the operation of a factor we have not yet encountered in rearrangements: conformational effects. More import.

• When optically active 2-amino-l,l-diphenyl-l -propanol is treated with nitrous acid, there is obtained 1,2-diphenyl-l-propanone of inverted configuration but lower optical purity than the starting material. Reaction has taken place with.

• racemization plus inversion: stereochemistry typical of SN! reactions, and consistent with the idea of an open carbonium ion as interned.

• In a series of elegant experiments, Clair Collins (of Oak Ridge National Laboratory) has given us intimate details about the reaction: the intermediacy of open carbonium ions, their approximate life-time, and the conformational factors that affect their chemistry. Collins, too, carried out deamination of optically active 2-amino-l,l-diphenyl-l -propanol, but his starting material was labeled stereo specifically

• (I) with carbon- 14 in one of the phenyl groups. He resolved

 

• the products and, by degradation studies, determined the location of the radioactive label in each. The inverted product had been formed exclusively by migration of the labeled group, Ph* ; the product of retained configuration was formed exclusively by migration of the unlabeled group, Ph. (The 12% of retention observed by Collins agrees, of course, quite well with the results of the earlier simple stereochemical study.)

• On the basis of these results, Collins pictured the reaction as taking place as shown in Fig. 28.1. Three conclusions were drawn, (a) An open carboning ion stormed. If, instead, migration of phenyl was concerned with loss of N2 , attack

• The course of rearrangement is thus determined largely by the conformation of the first-formed ion and, to a lesser extent, of the ion most easily formed from it by limited rotation. These conformations reflect, in turn, the most stable conformation of the parent diazonium ion.  

• Conformational factors can determine more than the stereochemistry of rearrangement. In light of Collins' findings, let us examine work done earlier by D. Y. Curtin (of the University of Illinois) with 2-amino-l-anisyl-l-phenyl1-propanol. This resembles Collins* labeled compound (p. 899), except that an Ainsly group (p-methoxyphenyl group) takes the place of one of the phenyls. Here, the competition in migration is between a phenyl and an Ainsly, instead of between labeled and unlabeled phenyl.

• In the deamination of VI, migration of Ainsly was found to exceed that of phenyl, 94:6. This, we might say, is to be expected: with its electron-releasing methoxy group, Ainsly migrates much faster than phenyl. But in the deamination of the diastereomer VII, phenyl migration was found to exceed that of Ainsly, 88:12. Clearly, migratory aptitude is not the controlling factor in the reaction of VII nor then, most probably, in the reaction of VI, either.

• In these particular reactions, then, just which group migrates is controlled, not electronically by intrinsic migratory aptitude, but sterically by conformational factors. This does not negate the idea of migratory aptitude. Groups do differ in their tendencies to migrate, and in some cases the effects of such differences can be very great. What we see here is simply that conformational factors can, sometimes, outweigh migratory aptitudes. 

• The strength of these electronic factors depends on how badly they are needed. In SN2-like rearrangements, where the migrating group is needed to help push out the leaving group, differences in migratory tendencies are very great. (That is, the migration terminus is an unreactive reagent and hence is highly selective.) Indeed, as we shall see in Sec. 28.12, the strength of the effects of substituents in migrating aryl groups can be used to measure the relative importance of SN iliac and SN2-3tke rearrangements.

Neighboring grope effects: stereochemistry

• When treated with concentrated hydrobromic acid, the bromohydrin 3-bromo-2-butanol is converted into 2,3-dibromobutane. This, we say, involves nothing out of the ordinary; it is simply nucleophilic attack (SN 1 or SN2) by bromide ion on the protonated alcohol. But in 1939 Saul Windstein (p. 474) and Howard J. Lucas (of the California Institute of Technology) described the stereochemistry of this reaction and, in doing this, opened the door to a whole new concept in organic chemistry: the neighboring group.

• In one of the products (I) from the erythron bromohydrin, there is retention of configuration. But in the other product (II), there is inversion, not only at the carbon that held the hydroxyl group, but also at the carbon that held bromine a carbon that, on the surface, is not even involved in the reaction. How is one to account for the fact that exactly half the molecules react with complete retention, and the other half with this strange double inversion.

• Reaction consists of two successive nucleophilic substitutions. In the first one the nucleophile is the neighboring bromine; in the second, it is bromide ion from outside the molecule. Both substitutions are pictured as being SN2-like

Neighboring group effects: rate of reaction. An chimeric assistance

• Like other alkyl halides, mustard gas (ftj0'-dichlorodiethyl sulfide) undergoes hydrolysis. But this hydrolysis is unusual in several ways: (a) the kinetics is order, with the rate independent of added base; and (b) it is enormously faster than hydrolysis of ordinary primary alkyl chlorides.

• C1CH2CH2-S-CH2CH2C1 -^^ CICH2CH2-S-~CH2CH2OH

• We have encountered this kind of kinetics before in SN! reactions and know, in a general way, what it must mean: in the rate-determining step, the substrate is reacting unimolecular to form an intermediate, which then reacts rapidly with solvent or another nucleophile. But what is this intermediate? It can hardly be the carbonium ion. A primary cation is highly unstable and hard to form, so that primary alkyl chlorides ordinarily react by SN2 reactions instead; and here we have electron-withdrawing sulfur further to destabilize a carbonium ion.

• This is another example of a neighboring group effect, one that shows itself not in stereochemistry but in rate of reaction. Sulfur helps to push out chloride ion, forming a cyclic sulfonium ion in the process. As fast as it is formed, this inter mediate reacts with water to yield the product.

• Reaction thus involves formation of a cation, but not a highly unstable carbonium ion with its electron-deficient carbon; instead, it is a cation in which every atom has an octet of electrons. Open chain sulfonium ions, R3S + , are well known, stable molecules; here, because of angle strain, the sulfonium ion is less stable and highly reactive but still enormously more stable and easier to form than a carbonium ion.

• The first, rate-determining step is unimolecular, but it is SN2-like. As with other primary halides, a nucleophile is needed to help push out the leaving group. Here the nucleophile happens to be part of the same molecule. Sulfur has unshared electrons it is willing to share, and hence is highly nucleophilic. Most important, it is there: poised in just the right position for attack. The result is an enormous increase in rate.

• The trans sulfide group evidently gives strong an chimeric assistance. Why cannot the cis sulfide? The answer is found in the examination of molecular models. Like other nucleophiles, a neighboring group attacks carbon at the side away from the leaving group. In an open-chain compound like mustard gas or like diastereomer of 3-bromo-2-butanol rotation abuts a carbon-carbon bond can bring the neighboring group into the proper position for back-side attack: anti to the leaving group (Fig. 28. 6a). But in cyclohexane derivatives, 1,2-substituents are anti to each other only when they both occupy axial positions possible only for trans substituents (Fig. 28.66). Hence, only the trans chloride shows the.

• When 2-acetoxycyclohexyl toluate is heated in acetic acid there is obtained, as expected, the diacetate of 1,2-cyclohexanediol. The reactant exists as diastereomers, and just what happens and how fast it happens depends upon which diastereomer we start with. The cis rosulate yields chiefly the trans diacetate. Reaction take* the usual course for nucleophilic substitution, predominant inversion. But the trans rosulate also yields trans diacetate. Here, apparently, reaction takes place with retention, unusual for nucleophilic substitution, and in contrast to what is observed for the cis isomer. Two pieces of evidence show us clearly.

• The apparent retention of configuration in the reaction of the trans toluate is a neighboring group effect. The neighboring group is acetoxy, containing oxygen with unshared electrons. Through back-side nucleophilic attack, acetoxy helps to push out the rosulate anion (1) and, in doing this, inverts the configuration at the

Neighboring group effects. Neighboring aryl

• In 1949, at the University of California at Los Angeles, Donald J. Cram published the first of a series of papers on the effects of neighboring aryl groups, and set off a controversy that only recently, after twenty years, shows signs of being resolved. Let us look at just one example of the kind of thing he discovered. Solvolysis of 3-phenyl-2-butyl toluate in acetic acid yields the acetate. The toluate contains two chiral centers and exists as two racemic modifications.

• too, does the acetate. Solvolysis is completely stereospecific and proceeds, it at first appears, with retention of configuration: racemic erythron toluate gives only racemic erythron acetate, and racemic throe toluate gives only racemic three acetate (Fig. 28.7). When, however, optically active throe toluate is used, it is found to

• an intermediate bridged ion a benzenediol ion. This undergoes nucleophilic attack (2) by acetic acid at either of the two equivalent carbons to yield the product.

• In all this, H. C. Brown played a role familiar to him: that of gadfly the organic chemist's conscience forcing careful examination of ideas that had been accepted perhaps too readily because of their neatness. The turning point in this part of the great debate was marked by the joint publication of a paper by Brown and Schleyer setting forth essentially the interpretation we have just.

Neighboring group effects: nonclassical ions

• The rearrangement of carbonium ions was first postulated, by Merwin (p. 160) in 1922, to account for the conversion of camphene hydrochloride into isononyl chloride. Oddly enough, this chemical landmark is the most poorly 

• understood of all such rearrangements. With various modifications in structure, this bicyclic system has been for over 20 years the object of closer scrutiny than any other in organic chemistry. We can see, in a general way, how this particular rearrangement could take place. Camphene hydrochloride loses chloride ion to form cation I, which rearranges by a 1,2-alkyl shift to form cation II. Using models, and keeping careful

• Reaction of the exon borylated is SN2-like, as shown in Fig. 28.9: back-side attack by C-6 on C-l helps to push out borylated and yields the bridged ion in a single step. The geometry of the endo borylated does not permit such back-side attack, and consequently it undergoes an SN l-hike reaction: slow formation of the open cation followed by rapid conversion into the bridged ion.

• The two diastereomers yield the same product, racemic oxo acetate, because they react via the same intermediate. But only the oxo borylates reacts with an chimeric assistance, and hence it reacts at the faster rate. 

• What Windstein was proposing was that saturated carbon using a electron could act as a neighboring group, to give an chimeric assistance to the expulsion of a leaving group, and to form an intermediate bridged cation containing pentavalent carbon. Bridged ions of this kind, with delocalized bonding a electron, have become known as nonclassical ions.

• In 1970, Olah reported that he had prepared a stable nonbony cation in SbF5-SO2. From its p.m., Cmar, and Raman spectra, he concluded that it has, indeed, the nonclassical structure with delocalization of a electrons. The 2-phenyU nonbony cation, on the other hand, has the classical structure; this benzylic cation, stabilized by electrons from the benzene ring, has no need of bridging. The tertiary 2-methylnorbornyl cation is intermediate in character: there is partial 9 delocalization and hence bridging, but weaker than in the unsubstituted cation. (Interestingly enough, delocalization in the 2-methyl cation seems to come, not from the C6~C7 bond, but from the C6 H bond; Olah pictures the back lobe of the carbon-hydrogen bond overlapping the p orbital of C2.)