Allylic Oxidation According To Kharasch Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.


The conversion of alkenes, both cyclic and acyclic, into allylic alcohols and allylic esters is an extremely useful tool in the world of organic synthesis. Allylic alcohols and esters are often intermediates in many preparative routes for pharmaceuticals, pesticides and natural products. With this in mind, the need to prepare these intermediates both efficiently and quite often stereospecifically has driven a field of organic chemistry for many years known as 'allylic oxidation'. There are many known ways of oxidising an allylic C-H bond, such as the use of selenium dioxide, transition metal catalysis, mercuric salts, metal organic frameworks, zeolites and enzyme catalysis, each of which will receive their own discussions later in the review. Perhaps the most well-known and well-studied of such transformations is the Kharasch-Sosnovsky reaction, first discovered in 1958. [1] Here, Kharasch reports the oxidations of the allylic C-H bond of cyclohexene and of 1-octene to the corresponding allylic ester by reaction of the olefin with tert-butyl peroxybenzoate in the presence of catalytic amounts of CuBr (Scheme 1).

Scheme 1: Allylic oxidation according to Kharasch

Since this initial discovery, many groups have worked on improving the yield, rate and stereoselectivity of the products, most notably the groups of Andrus [2] and Pfaltz [3] and also the Malkov group [4] with the development of chiral bisoxazoline-copper complexes2,3 and chiral bipyridyl ligands4 for asymmetric induction. These discoveries of asymmetric induction into the products further widened the field. Again, the Kharasch-Sosnovsky reaction will be discussed in greater detail in due course, along with applications in organic synthesis, but first, a general summary of allylic oxidation methods.

Allylic Oxidations Using Selenium Dioxide

A selection of oxidation products are available when an olefin is treated with selenium dioxide. The nature of the product is very much dependent on the solvent employed during the reaction for example, if the reaction is carried out in ethanoic acid, the allylic acetate is formed, in ethanol the ether is produced and the α,β-unsaturated ketone is formed when the reaction is carried out in water. [5] In 1972, Sharpless reported sufficient evidence for the elucidation of the mechanism for the oxidation which prior to this had been controversial. Scheme 2 shows the proposed mechanism.

Sharpless reported evidence that the [2,3] sigmatropic rearrangement from the allylseleninic acid (2) to the selenium ester (3) is a labile process and that the selenium ester is a meta-stable species. [6] 

Scheme 2: SeO2 Allylic Oxidation According to Sharpless

A big drawback for this type of reaction is the amount of reduced inorganic selenium(II) and organo-selenium by-products which serve almost as a deterrent for utilising this very reliable reaction. As a result of this, Sharpless developed a method by which catalytic amount of selenium dioxide could be used in the presence of a peroxide oxidant to produce allylic oxidation products in yields comparable to reactions involving sole stoichiometric selenium dioxide. [7] The use of the peroxide was to oxidise the selenium(II) hydroxide by-products back to selenium oxide which could then carry out further oxidation. The use of hydrogen peroxide as a co-oxidant however is known to produce the diol and epoxide oxidation products so tert-butyl hydroperoxide was used to ensure allylic oxidation proceeded without the formation of the unwanted products. It was found that more reactive (more substituted) olefins produced allylic alcohols in the presence of tert-butyl hydroperoxide and 1.5-2.0% selenium dioxide under mild conditions whereas the less reactive (less substituted) olefins required 0.5 equivalents of selenium dioxide with 2 equivalents of tert-butyl hydroperoxide.

Since this work by Sharpless and co-workers, oxidation using SeO2 in the presence of tert-butyl hydroperoxide has been applied in the total synthesis of miroestrol. [8] Scheme 3 shows the transformation employed as the penultimate step in a fourteen step synthesis.

Another interesting application of selenium dioxide allylic oxidation is the production of 2-acetylfurans from dienones as reported by Chung et al. [9] The dienone is simply refluxed in an appropriate solvent with two to four equivalents of selenium dioxide. Scheme 4 illustrates the transformation.

Scheme 4: Oxidation of a dienone to a 2-acetylfuran

It was proposed by Chung that the reaction proceeded via a [4+2] cycloaddition with the selenium dioxide acting as a dienophile and forming an initial six-membered seleno-lactone. Thermal decomposition of this intermediate leads to the 2-acetylfuran product and elemental selenium and water as by-products.

Transition Metal Catalysed Allylic Oxidation

A variety of transition metals can be used to catalyse allylic oxidation reactions. These include palladium, cobalt, manganese, cerium, titanium, vanadium and copper as previously mentioned. A method for the palladium-catalysed allylic oxidation of olefins was developed by McMurry and Kocovsky in 1984. [10] The group needed to produce a keto alcohol 2 to be used as a starting material in synthesis. In order to prepare 2, it was thought that a simple allylic oxidation of geranylacetone 1 would suffice. (Scheme 5) However, classical methods using SeO2 proved unsuccessful.

Scheme 5: The Desired Allylic Oxidation

In geranylacetone, there are six allylic positions so any successful allylic oxidation would need to be regioselective to ensure that only the terminal product formed. The method developed involved the use of catalytic amounts of Pd(OCOCF3) as it was known that palladium inserted into the terminal allylic C-H of geranylacetone to produce a π-allyl palladium complex. The use of an oxygen nucleophile such as o-methoxyacetophenone would then displace Pd(0) from the complex and an oxidising agent such as benzoquinone would complete the catalytic cycle by oxidising Pd(0) to Pd(II). This method proved successful for a variety of olefin substrates producing the allylic oxidation product in moderate to good yield.9

In 1988, Morimoto et al. [11] developed a method of allylic oxidation using cobalt(III), manganese(III) and cerium(IV) acetates in the presence of sodium bromide. Scheme 6 highlights the reaction carried out.

The presence of sodium bromide ensured that the desired product formed.

The use of cobalt(II) complexes has been shown to oxidise alkenes either by direct oxidation of the double bond to form an epoxide or by oxidation at the allylic position to form either an α,β-unsaturated ketone or an allylic alcohol. The nature of the product is dependent on the nature of the cobalt complex employed in the reaction. Reddy et al. [12] demonstrated this by the use of the cobalt complexes shown in figure 1.

Figure 1: Cobalt Complexes used by Reddy

Cobalt complex 3 was used to convert cyclohexene into the corresponding epoxide in the presence of 2-methylpropanal and dioxygen in acetonitrile and cobalt complex 4 was used to convert cyclohexene into a mixture of the α,β-unsaturated ketone and the allylic alcohol in identical conditions. The catalyst was present in 5 mol% with a ratio of 1:2 between the alkene and 2-methylpropanal. It was suggested that this difference in chemoselectivity may be a result of the charge of the complex. Complex 3 is charged and complex 4 is neutral. This difference in charge me cause different dioxygen species to form in the reaction such as O2., O2- and O22-, leading to different mechanistic pathways and producing different oxidation products.11

Porphyrin complexes of both titanium and vanadium have been used to oxidise alkenes at the allylic position. [13] A mixture of cyclohexene and 0.0015 mol% of a metallated porphyrin ring in benzene is simply irradiated with two 150 watt halogen bulbs under a controlled pressure of air. Products are again however a mixture of the epoxide, allylic alcohol and α,β-unsaturated ketone, with the relative ratios of products dependant on substitution about the porphyrin ring. Reaction times are also incredibly long at 180 hours so this method is not ideal for practice. A lot of the methods using transition metal catalysis appear to give a mixture of oxidation products and are not highly selective. These methods are therefore probably best used as academic exercises and are not so useful for practical purposes for example use for an allylic oxidation step in total synthesis.

A very recent paper by Jiang et al. [14] appears to have overcome the issue of selectivity in the oxidation products by using a copper-based MOF (metal organic framework). [Cu(bpy)(H2O)2(BF4)2(bpy)] (bpy = 4,4'-bipyridine) was used to catalyse the oxidation of cyclohexene using molecular oxygen as the oxidant. Again, four oxidation products are observed as Scheme 6 shows.

Use of the copper-MOF shows selectivity of the allylic hydroperoxide C over the allylic alcohol A, the α,β-unsaturated ketone B and the epoxide D. An interesting observation the group found is that when the copper-MOF is dehydrated, it is inactive as an oxidation catalyst. When rehydrated, the copper-MOF has a slightly different crystal structure and is once again active as an oxidation catalyst, showing even higher selectivity for the allylic hydroperoxide. Successive dehydration and rehydration cycles enhance the selectivity each time.

The Kharasch-Sosnovsky Reaction

The Kharasch-Sosnovsky utilises the special nature of an allylic C-H bond to form an allylic ester by reaction of an olefin with tert-butyl peroxybenzoate in the presence of a copper salt (Scheme 1). The original paper by Kharasch in 19581 reported the first allylic oxidation of this kind. Previous reactions using benzoyl peroxide as the oxidant gave rise to a mixture of isomeric allylic ester products but it was found that reaction with tert-butyl peroxybenzoate gave only one allylic ester. Kharasch reported that the reaction of 1-octene gave exclusively oct-1-en-3-yl benzoate with no isomeric oct-2-en-1-yl benzoate being detected. It was later found however that this was inaccurate due to technical limitations with detection of minor components at the time. Later work by Walling [15] found the initial claim by Kharasch on the exclusive formation of the 3-substituted allyl ester to be inaccurate and that the thermodynamically more stable 1-substituted allyl ester was also present in approximately a 10:1 ratio of the internal ester over the terminal ester product. (Scheme 6)

Scheme 6: The Regioselectivity of the Kharasch-Sosnovsky Reaction

Since Kharasch reported this method of allylic oxidation, much effort has been made to try to perform such reactions asymmetrically. Initial attempts made by Denney et al. [16] used copper salts of chiral acids to induce asymmetry into the product. (+)-α-ethyl camphorate was one of the chiral counter-ions employed for the reaction. Although the enantiomeric excess of the products was not measured directly, it was clear that this method of asymmetric induction was successful to a small extent as the product was able to rotate plane-polarised light.

The area remained virtually untouched until 1991 when Muzart [17] , again trying to perform the reaction asymmetrically, sought to investigate the effect of amino acids present in the reaction mixture on the nature product. Copper acetate was initially reacted with an amino acid in acetic acid to produce the copper salt in situ. Cyclohexene and the oxidant were then added to carry out the allylic oxidation. It was found that at least four equivalents of the amino acid were needed per copper acetate. This was to reduce the competition from the acetic acid in binding to the copper ion and gave rise to the highest optical yields. The cheap and commercial L-amino acids used in the reaction produced the S-enantiomer preferentially with enantiomer excesses not exceeding 30%.

Further enantioselective methods of the Kharasch-Sosnovsky reaction were developed almost simultaneously a few years later by the groups of Andrus2 and Pfaltz3. Prior work had shown that asymmetric copper catalysis of cyclopropanation and aziridination reactions could be achieved with the use of bisoxazoline ligands and with this in mind, the groups were inspired to apply this ideology to the asymmetric allylic oxidation quandary. Their curiosity prevailed with high enantiomeric excesses being observed for reactions involving the oxidation cyclopentene and cyclohexene. Scheme 7 illustrates the reaction carried out.

Scheme 7: An Asymmetric Kharasch-Sosnovsky Reaction

Using cyclopentene as the olefin and the copper-bisoxazoline complex ( R'=Me, R"=Ph), an enantiomeric excess (ee) of 80% was observed. A similar ee of 81% was observed with cyclohexene and the copper-bisoxazoline complex (R'=H, R"=tBu). Both of these reactions were carried out in acetonitrile at -20°C and although yields were moderate at 49% and 43% respectively, the high enantiomeric excesses observed were a promising success. With respect to linear alkenes allyl benzene and 1-octene under similar conditions, racemic mixtures of allylic ester were observed. Only when the reaction was carried out in benzene at 55°C were enantiomeric excesses seen at 36% and 30% respectively.

Similar work by the group of Pfaltz3 again using chiral bisoxazoline-type ligands to induce asymmetry into the allylic ester product gave reasonably good ee's at generally over 70%. These methods also improved on the mediocre yields reported by Andrus ranging from 60-80%. Figure 1 shows the selection of chiral ligands employed during the investigation.

Figure 1: Chiral Ligands Used By Pfaltz