THE INFLUENCE OF SUBSTITUENTS ON THE OXIDATION OF HYDROCARBONS BY C . N . HINSHELWOOD Received 15th February, 1951 The influence of substituents on the ease of oxidation of hydrocarbons is described, and discussed in relation to theories of oxidation and electronic theories about reactivity in general. Hydrocarbon oxidation has been discussed on more than one occasion in recent years, so that a general survey of the problem will not be made in this paper which will deal rather with one special aspect of the problem, namely the influence of structure and substitution on reaction rate. In spite of various ambiguities and complexities in the situation there is general agreement that in the range of what are conventionally called low temperatures, that is from 1 5 0 O to 300° C, the oxidation of hydrocarbons takes place by a mechanism depending on the formation of a peroxide which slowly decomposes into radicals, thereby setting up branching chains.1 R-0-O-R’ (where R’ may be H) splits into R-O- and R’-0- which in general break down, e.g.by the reaction The smaller hydrocarbon radical attacks oxygen and continues the chain. The range characteristic of this so-called low temperature mechanism is some zooo to 300’ lower than that in which the pyrolytic reactions be- come conveniently measurable. The contrast is clearly due in part to the reactivity of the oxygen biradical, but since the initial rate of attack of oxygen on hydrocarbon is quite slow compared with that attained when the branching chains have developed to the maximum extent, a second major factor is the relative ease of splitting of the central bond in RO-OR’.In the Discussion on the Labile Molecule 2 it was pointed out that the oxidation reactions are enormously more sensitive than the decomposition reactions to the structure of the hydrocarbon. This fact has, in part at least, a kinetic explanation. The rate of the branching chain reaction is of the form RICHS-O- = R1 + HCHO. B(I - e--At) A according as the system is stationary or non-stationary. In either case A or A’ measures the degree of branching and is a difference of two terms, one of which includes the velocity constant for the splitting of the peroxide into radicals. Variations in this velocity constant have therefore a magnified influence on the rate. There is nothing comparable to this situation in the reactions of pyrolysis.This is one reason for the contrast, though another may well be that the -0-O- link itself is specially sensitive to the nature of the attached groups. The expression for the oxidation rate contains two factors, one pro- portional to the initial rate of attack of the oxygen molecule on the hydrocarbon (chain-initiating process) , the other dependent upon the rate of branching. In principle, both map be influenced by the structure Semenov, Chemical Kinetics and Chairt Reactions (Oxford, Clarendon Press, 1935) ; Cullis and Hinshelwood, E;araday SOC. Discussions, 1947, 2, 117. a Faraday SOC. Discussions, 1947, 2, 117; Hinshelwood, J . Chem. SOC., B(eA’l - I) A’ ’ or 194% 531. 266C. N. HINSHELWOOD 267 of the hydrocarbon, but the second is likely to be the more important, since its effect is enhanced in the manner described.We shall consider, therefore, the influence of changes in the structure of R (including the introduction of non-hydrocarbon groups as substituents) on the reactions (a) R-0-OX = RO- + XO- (b) RH + 0, = R- + H-0-0-, and with special attention to the first. The facts recently brought to light about the structural effect may be summarized as follows. I. The rate of oxidation is very much lowered by the introduction of extra methyl groups into a hydrocarbon.2 The mode of initial attack by oxygen which leads to the peroxide most actively capable of chain-branch- ing is that made on a CH, group as remote as possible from methyl. 2. If there is no place for the attack except a methyl group, as in CH,CH,, CH,COCH, or CH,OCH, there is a veyy great reduction in rate of oxidation compared with that of a corresponding compound containing a CH, group., 3.The substitution of chlorine increases oxidation rate, and acts in two ways, firstly by destroying the symmetry of any methyl group into which the substituent enters, and secondly by a direct inductive effect of the chlorine atom.4 If definite -numerical magnitudes are as- signed to the stabilizing influence of methyl groups and to the activating influence of chlorine atoms, a coherent relation appears between the oxidation rates and the calculated " stability factors '' for a whole series of hydrocarbons and chloro compounds. Acetone, it is true, oxidizes only about a quarter as fast as propane under similar conditions, but about 20 times as fast as ethane.The slowness compared with propane is due to the destruction of the CH, group on passage from CH,CH,CH, to CH,COCH,. 2-Pentanone reacts about as East as pentane and 3-pentanone several times faster. Butanone reacts about as fast as butane. In the passage from CH,CH,CH,CH, LO CH,COCH,CH, two equivalent CH, positions become reduced to a single one, and the fact that in spite of this the rate is maintained shows that the inductive effect of the carbonyl oxygen is essentially one favouring oxidation. 5. The introduction of an amino group leads to an increased oxidation rate.6 With propane and butane the effect is not very great but with ethane it is most marked, once again, presumably because the symmetry of a methyl group is destroyed and a CH, group appears in The molecule.6. Ethers are very much more easily oxidized than paraffins,' diethyl ether reacting about 2500 times as rapidly as pentane. 7. With unsubstituted hydrocarbons, chloro compounds, amino compounds and ketones the rate increases very rapidly with the lengthen- ing of the normal carbon chain, but with the ethers the increase from ethyl ether to higher ethers is relatively small, as has recently been found by Dr. Eastwood. The enormous increase caused by the 0 atom appears to have saturated the possible response of the molecule. It will be observed that the only substituent which stabilizes the mole- cule towards oxidation is the methyl group. This is an electron-repelling group, with an inductive action represented by - I in the conventional symbolism.If we regard ethers as hydrocarbons with the substituent 4. Carbonyl groups likewise tend to increase oxidizability.6 Mulcahy, Trans. Faraday SOC., 1949, 45, 537. Cullis, Hinshelwood and Mulcahy, R o c . Roy. SOC. A , 1949, 196, 160. Cullis and Smith, Trans. Faraday SOC., 1950, 46, 42. 6 Bardwell (in press). 'Malherbe and Walsh, Trans. Faraday SOC., 1950, 46, 835 ; Eastwood (unpublished work).2 68 INFLUENCE OF SUBSTITUENTS OR, then the groups which accelerate oxidation are OR, C1 and NH,, and all these are of the + I type. The moment C+-0- of the carbonyl group indicates that in ketones also a flow of electrons into the sub- stituent occurs. We may conclude that the two major influences of structure on oxidation are firstly, the great tendency of methyl groups to preserve themselves intact and, secondly, a stabilizing influence of electron access to the seat of reaction. The first influence, which wherever possible displaces the seat of reaction away from methyl groups, is probably conditioned by the symmetry of their structure and the consequent possibility of electron delocalization.The second depends upon the response to electron accession or recession of the reactions ( a ) and (b) referred to above. Walsh 8 has argued that bonds between strongly electronegative elements should be strengthened by electron-repelling groups, since these facilitate the expansion of atomic orbitals and allow increased overlap without the occurrence of nuclear repulsion.If this is so, then the stabilization of RO-OR’ by the -I groups and its weakening by +I groups would provide a satisfactory interpretation of the observed effects. Electron displacement phenomena must show saturation and this fact would explain why introduction of -0- (in diethyl ether) is not reinforced by a lengthening of the carbon chain, whereas lengthening of the chain in a simple paraffin causes progressive and marked increases in the rate of oxidation. The influence of electronic displacements on reaction (b) is, as already pointed out, less important, and is likely to be smaller in any case, since the C-H bond exists between atoms with a much less pronounced electro- chemical character than oxygen. It is surprisingly difficult to form an unambiguous opinion about the effect which electron-accession should have on the strength of this bond.There is, however, another possible line of argument about the ease of reaction which, with all due reserve, seems to be as follows. RH + 0, = R + HO,, the attacking oxygen must be changed into the biradical form under the influence of RH. In this change electrons are drawn from the O=O bond, and the very initial stages of the necessary re-distribution may well be helped by the action of a positive centre in R which would tend to attract electrons towards it. If this were so +I substituents would favour attack and -1 substituents would weaken it. Thus methyl groups, by causing a flow of electrons towards the seat of reaction, would weaken the attack and the other substituents would intensify it. It is, of course, not very difficult to formulate qualitative arguments which might seem to lead to an opposite conclusion. But in general the influence of substituents on the reactivity of organic molecules seems most profitably to be assessed in terms of the ease of approach of the attacking agent, and the initiation of the requisite reorganization. If this point of view is applicable here also, the effects of the substitution on reaction (a) and on reaction (b) are in the same direction and reinforce one another so that a general interpreta.tion of the observed changes in oxidizability becomes possible. Physical Chemical Laboratory, In the process South Parks Road, Oxford. a Walsh, J . Chem. SOC., 1948, 398.