Radical Clock Reactions
Johnson, C.C.; Lippard, S.J.; Liu, K.E.; Newcomb, M. J. Am. Chem. Soc. 1993, 115, 939-947.
Introduction:
Understanding an organic reaction mechanism is an integral part of synthetic chemistry. If one knows the path through which the reactants are transformed into products, then the product composition, including stereochemistry, can be predicted before the reaction is actually carried out. There exist several types of reactive intermediates found in the main types of organic mechanisms. These include carbocations, carbanions, carbenes, and radicals.

Radicals occupy an important niche in organic reactions and determining if radical intermediates are involved in the mechanism is invaluable. In 1980, Griller and Ingold introduced an indirect way of determining if the reaction proceeds via radical intermediates.1 The method involved free radical clock reactions. The idea behind these reactions is that an essentially irreversible radical rearrangement occurs with a known rate of rearrangement (kr) and can be used to probe a reaction mechanism.2 Organic chemists then use these calibrated clocks in two ways. First, they incorporate the radical clock into the molecule at a position where the radical is proposed to be formed. The reaction is then carried out and if a radical is formed in the mechanism, a portion of the reactants will undergo the irreversible radical rearrangement. Thus, the products will be a mixture of the rearranged and unrearranged reactants, indicating the presence of a radical. Second, if rearranged products are observed, the product ratios and the known rate of radical rearrangement allows one to indirectly calculate the rate constant of the reaction in question. Thus, a radical mechanism can be supported merely by the presence of rearranged products. The basis of the free-radical clock method is to create a competition between the free-radical rearrangement reaction (with a known kr) and the proposed mechanism in the presence of an appropriate quenching/trapping agent (Figure 1).3

Figure 1: Competition kinetics experiment. Radical R· rearranges to give R¢· in competition with trapping agent X-Y which gives the unrearranged and rearranged products R-X and R¢ -X.

This competition results in a ratio of products consisting of the unrearranged and rearranged radical precursors. Thus, if the rearranged product is observed, then a radical intermediate is most likely present. In addition, the rate constant for the unrearranged reaction can be determined from the ratio of the products (Figure 2).1,4


Figure 2: Radical Clock Study of [2+2] Photocycloaddition.

It should be noted, however, that failure to observe any rearranged products does not eliminate a radical pathway. Instead, the reaction in question maybe faster than the radical's rearrangement rate, so no rearrangement is observed. In that case, a faster radical clock is necessary.

Thus, the whole idea of radical clock studies hinges on the ability of the radical clock to rearrange. Two of the most popular radical clock rearrangements are the cyclopropylmethyl and the 5-hexenyl types (Figure 3).3,5 The cycloproylmethyl undergoes a very rapid ring opening rearrangement, which is enthalpically favorable, in order to relieve angle strain. Likewise, the 5-hexenyl radical rearranges extremely fast to give the five-membered ring; formation of the five-membered ring is both entropically and enthalpically more favorable than formation of a cyclohexyl ring.


Figure 3: Cyclopropylmethyl and 5-hexenyl radical rearrangements.
Whole "horlogeries" (in French, denotes a store where clocks are sold)1 have been developed ranging from primary alkyl radicals to the benzylic radical clocks (Table 1).1,3,6 (Note: This table is not meant to be all inclusive, instead it gives a general flavor of the types of radical clocks used). As can be seen, the whole span of rate constants is represented from the slowest to the fastest. The fastest known radical clocks are at the edge of the rate of transition state collapse which has been calculated to be around 1x1013 s-1.9 Also, it is worth noting that the majority of radical clocks fall into the primary alkyl radical category. Why this is so is not very clear since heteroatom radical chemistry is fairly extensive.

Table 1: Types of Radical Clocks and Their Corresponding Rate Constants.

It is also interesting to note that like real clocks, the radical clocks can be tuned and made to "run" faster or slower by adding or removing substituents that stabilize the radical. As shown in Figure 4, varying X and Y can increase or decrease the rate constant of the radical clock.3


Figure 4: Tuning Radical Clocks

However, tuning is usually only able to adjust the rate by a factor of ten in either direction. Thus, if the reaction to be studied has a rate very different from the radical clock that is being tuned, one would be better off choosing a radical clock with a comparable rate constant.

As stated previously, these radical clocks are then incorporated into the molecule to be studied at the position where a radical is believed to be formed. The reaction is then run and product analysis is done to see if any rearranged product due to radical formation is observed. Examples of the use of radical clocks to probe mechanisms include the [2+2] photocycloaddition.

Radical intermediates were proposed for the [2+2] photocycloadditions of enones by Bauslaugh in 1970. This was supported by a study done by Rudolph and Weedon in which the photocycloaddition of cyclopentenone and vinylcyclopropane gave the expected cyclobutyl containing compounds (II and III) as well as the ring opened compounds (I and IV) (Figure 5).4 In the experiment, the vinylcyclopropane acted as the masked cyclopropylmethyl radical clock. In the reaction, as shown in Figure 5, the vinylcyclopropane can add intitally to produce four distinct radical intermediates. Of these, 2 and 3 lead to the expected cyclobutyl containing compounds (II and III). However, intermediates 1 and 4 have the radical formed so that a cyclopropylmethyl radical is present. This radical can rapidly rearrange as shown in Figure 3 leading to the ring opened products I and IV. The presence of these rearranged products indicates the formation of a diradical intermediate in the mechanism.

Figure 5:  [2+2] photocycloaddition of cyclopentenones and vinylcyclopropane.

In 1975, Brace and Van Elswyk used the 5-hexenyl radical clock to show that the reductive dehalogenation of 1-perfluoroalkyl-2-iodoalkenes with zinc and acid was radical in nature (Figure 6).10 During the reaction, a radical intermediate is formed beta to the perfluoroalkyl group. This radical is now part of the 5-hexenyl radical clock and will undergo rearrangement in competition with reduction. As can be seen, more than 75% of the product isolated was the rearranged cyclopentyl perfluoroalkyl compound, indicating the presence of a radical in the mechanism.



Figure 6: Zinc/Acid Dehalogenation of perfluoroalkyl-iodoalkenes.

As a last example, Peralez and co-workers have used radical clocks to demonstrate the free-radical nature in Grignard reagent formation.11 It has been proposed that Grignard formation proceeds through the chain pathway shown in Scheme 1.14


Scheme 1: Proposed pathway for Grignard formation.

It was believed that a single electron transfer (SET) from magnesium to the alkyl halide occurred followed by fragmentation to give the alkyl radical and the Mg(I)X· species. To test this, researchers used an endo-5-(2¢ -haloethyl)-2-norbornene in which an endo-5-ethyl-2-norbornenyl radical was expected to be formed (Figure 7).11 This radical could then possibly cyclize with a rate constant of 1x107s-1. Table 2 gives these results, as can be seen, significant amounts of rearranged product are observed confirming the presence of radicals.11

Figure 7: Endo-5-ethyl-2-norbornenyl radical study of Grignard formation

Table 2: Reaction of endo-5-(2¢ -haloethyl)-2-norbornene with Mg° .

Further use of radical clocks in mechanistic studies are seen in the reduction of alkyl halides with sodium napthalenide12, the reaction of enones13, the Wittig rearrangement14, the reductive elimination reactions of dialkylmercury compounds15,16, dioxirane dihydroxylations17, and electrophilic fluorinations.18

It worth repeating here that failure to form rearranged products does not absolutely exclude a radical mechanism. The reaction in question may be faster than the radical clock used, or perhaps the substituents cause the clock to "run slower" than expected, leading to slower rearrangement. However, if even the fastest radical clock does not produce rearranged products, then a radical intermediate is deemed highly unlikely.

Despite this drawback, radical clocks have become increasingly important in probing the mechanisms of enzymes. This is because of two reasons. One, if any rearranged products are observed, then a radical must have been formed at some time and can confirm or support a radical mechanism. Secondly, the environment created in the active site of an enzyme makes the use of radical clocks ideal. This is because the active site has the capability of stabilizing any intermediate, radicals included, formed in the reaction process. Thus, the longer the radical is present, the higher the probability of seeing rearranged products.

By knowing the rates of these rearrangements, it is then possible to probe a mechanism for radical character and also measure the rate constant for the reaction of interest. One such enzyme that has been studied extensively is methane monooxygenase (MMO), which converts an inactivated C-H bond in methane into a C-OH to produce methanol and liberate energy.


Results:
 

This enzyme is an optimal system to utilize radical clocks because MMO can oxidize a wide variety of substrates, therefore, radical clock substrate probes can potentially derive mechanistic information about the hydroxylation step in the enzyme mechanism. The debate revolves around if during the hydroxylation reaction there is formation of a discrete substrate radical intermediate or if reaction proceeds through a concerted mechanism, where an oxygen atom inserts into a C-H bond and would not involve a radical. To establish the presence of a radical intermediate Liu et al. utilized four radical clock substrates, all having the cyclopropylcarbinyl skeleton (Figure 8).9

Figure 8: Four substrate radical clocks.

This is important because radical formation in these systems result in rapid ring opening, on the order of 108 s-1 and higher. What they were hoping to find was that if a radical intermediate was formed, there would be a good chance that the radical would isomerize to form, ultimately, a rearranged alcohol product.

Hydroxylation of these four substrates with MMO from M. capsulatus (Bath) were carried out at 45oC.7 Products were isolated by thorough extraction of the reaction mixture with diethyl ether or ethyl acetate. Isolated products were then analyzed with GC-MS. Identification of product components were carried out by comparing mass spectra of the products to those of authentic standards. Quantitation of products were obtained by internal standard method. The results are shown in Table 3.9

Table 3: Results from study.

The results of using the four hydrocarbon substrates as mechanistic probes in oxidations catalyzed by the MMO of M. capsulatus (Bath), all showed no observable rearrangement (meaning, less than 1% formed during the hydroxylation reaction). In every case, the major product(s) observed corresponded to the unrearranged alcohol. This suggest the hydroxylation reaction in M. capsulatus (Bath) does not involve radical formation.


Discussion:
 

The essence of radical clocks, as stated earlier, is to have an indirect method for determining the relative rates of radical reactions using essentially an irreversible intramolecular rearrangement as a mechanistic probe. If hydrogen atom abstraction from trans-2-phenylmethycyclopropane (1, Table 3) were to occur, the resulting cyclopropylcarbinyl radical could either rearrange and then be hydroxylated (1b, Table 3) or be hydroxylated before rearrangement occurred (1a, Table 3). Essentially, the two reactions (isomerization to form the rearranged product and trapping to form the unrearranged product) are competing and the relative compositions of products 1a and 1b will be dependent on the relative rates of trapping versus that of rearrangement. Generally stated, if, ktrapping > krearrangement then, the formation of product 1a will dominate the formation of product 1b, even though the formation of product 1b goes through the formation of the more stable secondary radical. These competitive kinetic reactions strongly suggest the presence of a radical intermediate and can give information of the relative lifetime of this intermediate along with the rate constants for radical trapping.

Using the rate constants for rearrangement of the corresponding cyclopropylcarbinyl radical (radical clock) the trapping rate constants for a radical intermediate would have to be at least 4x1013s-1 , 1x1013s-1 , 4x1010s-1 , 3x1011s-1 at 45oC for the formation of 1b, 2b, 3b, and 4b from Table 3 (assuming that a radical intermediate was formed). This estimation was calculated by the known rate constant for ring opening, k rearrangement , according to the equation below.


ktrapping= krearrangement([unrearranged alcohol]/[rearranged alcohol]).

The general trend seems to be that the trapping rate must be 100 times in magnitude larger than the rate of reorganization to totally negate the formation of rearranged product. However, it is known that the rate constants of trapping can not exceed the rate constants of decomposition of a transition state (1x1013 s-1). Since the radical rearrangement is already on the order of 1x1011 s-1 , it is very unlikely that the trapping reaction can be faster. Therefore, the absence of ring-opened (rearranged) products indicates that no radical intermediate was formed in hydroxylation.

More recently, with the increasing number of radical clock substrates capable of detecting even shorter-lived radicals, the question of a radical based intermediate in MMO, isolated from M. trichosporium OB3b, was studied.7 Jin and Lipscomb utilized methylcubane as the radical clock substrate since this relatively rigid hydrocarbon system undergoes a rapid rearrangement at a rate of 3x1010s-1. Methylcubane can potentially be hydroxylated at a number of positions, all of which lead to different products. If, a radical-based mechanism were occurring, the radical could be generated by homolytic cleavage at the C-H of the side chain methyl group or the C-H of the cubyl group, which would ultimately lead to cubylmethanol and different methylcubanols as products (Figure 9). In kinetic competition with this reaction is the fast rearrangement, which could lead to a rearranged alcohol product assuming that the radicals live long enough to rearrange (Figure 9).7


Figure 9: Potential products from (A) direct hydroxylation of methylcubane, (B) the formation of a cubylcarbinyl radical intermediate that rearranges rapidly.
The results of Jin and Lipscomb's study reveal that upon reaction of this methylcubane substrate with MMO from M. trichosporium OB3b (in the presence of NADH and O2) yielded several different products in varying compositions in the reaction GC trace and the product profiles. The lowest intensity peak, which is made up of 3 overlapping peaks, was attributed to the three possible methylcubanols resulting from hydroxylation at the cubyl C-H positions (2-, 3-, and 4-methylcubanol). The medium intensity peak corresponded to hydroxylation at the C-H of the side chain methyl group to form cubylmethanol. The most interesting result came from the highest intensity peak. They found that this peak was a rearranged product from the radical probe, specifically RP2 (Figure 9B). This assignment implies the formation of the cubylcarbinyl radical during the reaction, which would be consistent with the proposed radical-based reaction pathway. When this same radical clock (methylcubane) was used as a substrate in M. capsulatus (Bath), no rearranged alcohol product was observed, as indicated in the product distribution chart (Table 6). These results, as stated earlier, support the notion that a significant component of the hydroxylation reaction pathway does not proceed through a radical intermediate from M. capsulatus.


Table 6: Product distributions.

The problem with studying these radical clock probes to rationalize the presence or absence of a radical intermediate is that a major underlining assumption has to be made as far as the reactivity or rate of rearrangement of these substrate in the enzyme's active site. The assumption made is that the rate of rearrangement (krearrangement) in an enzyme active site is essentially the same as that found in solution, where they were calibrated. For some cases this may not be valid assumption because proteins, through steric constraints or other interactions, can alter the reactivity of the probe from that observed free in solution. If the ring opening were impeded, through the intimate association of enzyme and substrate, the clocks would "run slow" in the enzyme active site. This concern was addressed by using probes that do not dramatically change shape on rearrangement, such as the case previously mentioned using methylcubane. The radical formed by H atom abstraction in methylcubane ended up on the same carbon atom so there is no need for substrate translocation in the active site before trapping. The oxidation of methylcubane by MMO from M. trichoporium OB3b yielded several rearranged products. The point here is that methylcubane does not undergo a major change in shape, therefore, if the ring opening reaction is partly masked in the enzyme active site some rearranged product would still be observable. Whereas, using the cyclopropyl carbinyl radical, which undergoes a major change in shape, masking of this ring opening reaction could have drastic effect of the product composition.8 The rate of trapping could potentially completely mask the ring opening reaction and lead the notion that this reaction is not going through a radical intermediate.


Conclusion:
 

Though much debate still revolves around the mechanism of catalysis found in MMO, the use of cyclopropylcarbinyl compounds (Liu et al.) was quite ingenious. The rapid ring opening equilibrium between the cylclopropylcarbinyl radical and its open chain form are the most widely studied of all the radical clock reactions (Refer to Figure 3). This reaction is considered essentially irreversible in the direction that leads to ring opening. This reaction, along with a number of other examples of radical ring opening clocks, is driven by the strain energy of a small ring that is ruptured in the rearrangement. Cyclopropylcarbinyl and its alkyl substituted derivatives have fast radical rearrangements with rate constants for ring opening ranging from ~1x108s-1 to about 4x109s-1, a 40 fold increase in ring opening. The incorporation of radical stabilizing groups on the cyclopropyl ring results in extremely fast ring opening. These radical rearrangements can be used to probe mechanisms to determine if radical intermediates are formed.


References:
 

1.  Griller, D.; Ingold, K.U. Acc. Chem. Res. 1980, 13, 317-323.
2.  Bowry, V.W.; Hollis, R.; Hughes, L.; Ingold, K.U. J. Org. Chem. 1992, 57, 4284-4287.
3.  Newcomb, M. Tetrahedron. 1993,Vol.49(6), 1151-1176.
4.  Kaprinidis, N.A.; Lem, G.; Schuster, D.I. Chem. Rev. 1993, 93, 3-22.
5.  Crich, D.; Motherwell, W.B. Free Radical Chain Reactions in Organic Synthesis. Academic Press: London, 1992.
6.  Ballestri, M.; Chatgilialoglu, C.; Timokhin, V.I. J. Org. Chem. 1998, 63, 1327-1329.
7.  Lipscomb, J.D.; Jin, Y. Biochemistry. 1999,38, 6178-6186.
8.  Matthew, L.; Warkentin, J. J. Am. Chem. Soc. 1986, 108, 7981-7984.
9.  Johnson, C.C.; Lippard, S.J.; Liu, K.E.; Newcomb, M. J. Am. Chem. Soc. 1993, 115, 939-947.
10.  Brace, N.O.; Van Elswyk, J.E. J. Org. Chem. 1976, 41(5), 766-770.
11.  Chanon, M.; Negrel, J-C.; Peralez, E. Tetrahedron. 1995, Vol.51(46), 12601-12610.
12.  Barbas, J.T.; Garst, J.F. J. Am. Chem. Soc. 1974, 96(10), 3239-3246.
13.  House, H.O.; Weeks, P.D. J. Am. Chem. Soc. 1975, 97(10), 2778-2784.
14.  Garst, J.F.; Smith, C.D. J. Am. Chem. Soc. 1976, 98(6), 1526-1537.
15.  Kochi, J.K.; Nugent, W.A. J. Organometallic Chem. 1977, 124, 327-347.
16.  Kochi, J.K.; Nugent, W.A. J. Am. Chem. Soc. 1976, 98(17), 5405-5406.
17.  Banks, J. T.; Garden, S.J.; Ingold, K.U.; Vanni, R. Tetrehedron Letters. 1995, Vol.36(44), 7999-8002.
18.  Differding, E.; Rüegg, G. Tetrahedron Letters. 1991, Vol.32(31), 3815-3818.
Question 1:
 
It has been reported that the rate constant for the isomerization of 1, 2-di-tert-butylethylene (1) is on the order of k1= >3.2x105, making this a reasonable radical clock for measurement of radical reaction in the middle region of the kinetic scale. In contrast, the rate of isomerization of di-tert-butylacetylene (2) is 5000 slower than that compound 1, making this reaction kinetically insignificant and essentially a useless radical clock. Explain these results.


Answer 1:
 

During the rearrangement, a 1,2 alkyl shift occurs and the resulting transition state (or intermediate) has a three center cyclic structure and there is considerably more strain energy in the alkylidenecyclopropane (B).
Reference to Question 1:

Griller, D.; Ingold, K.U. Acc. Chem. Res. 1980, 13, 317-323.

Question 2:
 

Propose a mechanism for the round trip radical rearrangement of the methylcubyl radical.


Answer 2:
 
 

Reference to Question 2:
 

Newcomb, M. Tetrahedron. 1993, Vol. 49(6), 1151-1176.


Question 3:
 

It has been proposed by Ruzicka and coworkers that the radical clock studies of MMO from M. trichosporium OB3b with 1,1-dimethylcyclopropane yield results which are consistent with a radical intermediate. However, Ruzicka and others have proposed an alternative mechanism? What could this be?


Answer 3:
 

This alternative mechanism is a cationic mechanism in which MMO abstracts an H- from the substrate. This can cause the same type of rearrangement in the cyclopropyl group in the cationic mechanism as in the radical mechanism.


Reference to Question 3:
 

Ruzicka, F.; Huang, D.-S.; Donnelly, M.I.; Frey, P.A. Biochem. 1990, 29, 1696-1700.