The mechanism of the β-(acyloxy)alkyl radical rearrangement: Substituent and solvent effects

Full text of DOI:10.1021/ja963153o

12838
J. Am. Chem. Soc. 1996, 118, 12838-12839
Scheme 1
The Mechanism of the â-(Acyloxy)alkyl Radical
Rearrangement: Substituent and Solvent Effects
Athelstan L. J. Beckwith* and Peter J. Duggan*,1
Research School of Chemistry
Australian National UniVersity, GPO Box 4
Canberra, ACT 2601, Australia
ReceiVed September 9, 1996
Ever since it was first reported,2 the â-(acyloxy)alkyl radical
rearrangement (Scheme 1), an intramolecular process that has
no intermolecular analog, has attracted almost continuous
mechanistic investigation. Early results indicated that it involves
a five-membered cyclic transition structure (1) and hence can
be regarded as an open shell pericyclic process.3 However, more
recent work on this and related rearrangements3,4 has supported
the view that the mechanism might sometimes involve a tight
radical-cation/anion pair (2) and/or a three-membered cyclic
transition structure 3.3,4c,5 These conclusions have been based
on the comparison of results from widely different systems. In
order to obtain a more cohesive set of data, we have now
examined the substituent effects, the solvent dependence, and
the product distribution of an 17O label for a single type of
â-(acyloxy)alkyl radical.
Scheme 2
The 2-(butanoyloxy)-2-phenylpropyl radical 5a was chosen
as the parent system (Scheme 2) because (a) the rearrangement
of the corresponding acetate is known to proceed smoothly;6
(b) it has a molecular weight amenable to labeling studies that
employ 17O NMR spectroscopy;3 (c) the rearrangement rate was
expected to be in a range suitable to be measured accurately by
the tributyltin hydride method;3,5 (d) substituent effects could
be readily examined by changing substitution on the phenyl ring;
and (e) it undergoes a neophyl rearrangement that is relatively
insensitive to polar effects7 and hence can be used as a kinetic
yardstick for calibration of the rates of rearrangement of 5a and
its substituted derivatives.
0.72) - (7.35 ( 1.2)/2.3RT where kr and kn are the rate constants
for the â-(acyloxy)alkyl and the neophyl rearrangements,
respectively. Substitution of the Arrhenius parameters for kH,
the rate constant for the reaction of Bu3SnH with neopentyl
radical11 gave the temperature dependence for kr and kn (eqs 1
and 2).
In the usual way,3,5,8 the bromide 4a9 was treated with Bu3-
SnH (0.0156-0.125 M) and AIBN (2,2′-azobisisobutyronitrile)
in benzene at various temperatures (11.5-100.0 °C), and the
yields of 6a, 7a, and 8a9 were accurately determined by GC.
The usual treatment of the data10 gave log[(kr/kH)/M] ) (3.24
( 0.36) - (8.34 ( 0.55)/2.3RT and log[(kn/kH)/M] ) (1.37 (
log[(kr)/s-1] ) (11.7 ( 0.5) - (11.0 ( 1.3)/2.3RT (1)
log[(kn)/s-1] ) (9.9 ( 0.8) - (10.1 ( 2.0)/2.3RT (2)
(1) Current address: Department of Chemistry, School of Molecular
Sciences, James Cook University of North Queensland, Townsville, Qld,
4811, Australia.
(2) Surzur, J. M.; Teissier, P. C. R. Acad. Sci., Ser. C 1967, 264, 1981.
Surzur, J. M.; Teissier, P. Bull. Soc. Chim. Fr. 1970, 3060. Tanner, D. D.;
Law, F. C. P. J. Am. Chem. Soc. 1969, 91, 7535.
The log A terms for kr and kn are both somewhat smaller
than those previously reported for these types of reactions3,5,12
but the rate constant kr at 75 °C (6.2 × 104 s-1) is mid-range
for a â-(acyloxy)alkyl rearrangement.3 However, the neophyl
migration rate constant kn at 75 °C (3.6 × 103 s-1) for 5a is
somewhat smaller than usual.12
(3) Beckwith, A. L. J.; Duggan, P. J. J. Chem. Soc., Perkin Trans. 2
1993, 1673 and references cited therein.
(4) (a) Wilt, J. W.; Keller, S. M. J. Am. Chem. Soc. 1983, 105, 1395
(ester migration onto a silyl radical). (b) Crich, D.; Filzen, G. F. J. Org.
Chem. 1995, 60, 4834 (rearrangement of â-(nitroxy)alkyl and â-(sul-
fonatoxy)alkyl radicals). (c) Crich, D.; Yao, Q.; Filzen G. F. J. Am. Chem.
Soc. 1995, 117, 11455. Crich, D.; Jiao, X.-Y. J. Am. Chem. Soc. 1996,
118, 6666 (phosphate ester migration). (d) Crich, D.; Yao, Q. Tetrahedron
1994, 50, 12305 (rearrangement of â-(vinyloxy)alkyl radicals).
(5) Beckwith, A. L. J.; Duggan, P. J. J. Chem. Soc., Perkin Trans. 2
1992, 1777.
Solvent effects were determined by measuring the relative
rate constants kr/kH and kn/kH at 75 °C in six solvents.13
Correlation of the results (Table 1) with the solvent polarity
parameter ET14 gave log[(kr/kH)/M] ) 0.020ET - 2.793 and log-
[(kn/kH)/M] ) -0.004ET - 3.135. Since kn is virtually
independent of solvent polarity,7 kH must show a weak solvent
(6) Beckwith, A. L. J.; Thomas, C. B. J. Chem. Soc., Perkin Trans. 2
1973, 861.
(7) Beckwith, A. L. J.; Ingold, K. U. In Rearrangements in Ground and
Excited States; de Mayo, P., Ed.; Academic Press: New York, 1980; Vol.
1, pp 170 -174.
11
dependence. Normalization against the value of log kH in
(8) Second-order conditions were employed (0.95 equiv of Bu3SnH). The
following equations were also used: [7]final/[8]final ) kr/kn; kR ) kr + kn;
[7]final + [8]final ) (kR/kH)ln{([Bu3SnH]init + kR/kH )/([Bu3SnH]final + kR/
kH)}.
(11) Johnston, L. J.; Lusztyk, J.; Wayner, D. D. M.; Abeywickreyma,
A, N.; Beckwith, A. L. J.; Scaiano, J. C.; Ingold, K. U. J. Am. Chem. Soc.
1985, 107, 4594.
(12) Lindsay, D. A.; Lusztyk, J.; Ingold, K. U. J. Am. Chem. Soc. 1984,
106, 7087.
(13) The rate constants for reactions in benzene at 75 °C were calculated
from the Arrhenius parameters. The others were averages of three to five
determinations, each at different stannane concentrations.
(14) Reichardt, C. Angew. Chem., Int. Ed. Engl. 1965, 4, 29.
(9) All new compounds were characterized by 1H and 13C NMR, IR,
MS, and HRMS or microanalysis.
(10) Product ratios were determined at four different stannane concentra-
tions at each of four temperatures for the ester migration and at each of
three temperatures for the neophyl rearrangement.
S0002-7863(96)03153-8 CCC: $12.00 © 1996 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 118, No. 50, 1996 12839
Table 1. Kinetic Data for the Rearrangement of 5a in Various
Solvents at 75 °C
Table 2. Rearrangement of 17O-Labeled Substrates
17O distributions in % (17O NMR chemical shift)a
solvent
ET
10-6 kH (M-1 s-1
)
10-4 kr (s-1
)
substrateb
rearrangedc
C6H12
C6H6
DME
DMF
EtOH
MeOH
30.9
34.5
38.2
43.8
51.9
55.5
6.17
6.37
6.60
6.95
7.48
7.74
3.7
6.4
5.8
8.4
13.0
16.4
solvent radical
-CdO
-O-
-CdO
-O-
C6H6
5a
84 (370)
100 (368)
100 (375)
84 (370)
16 (160)
14 (346)
39 (346)
86 (162)
61 (160)
100 (160)
67 (162)
C6H6
5bd
5ce
5a
C6H6
CH3OH
16 (160)
33 (346)
a The ppm values are relative to external 17O H2O. b Compound 4.
c Compound 7. d Heating of 4b in benzene for an extended period did
not result in any detectable scrambling of the ester oxygens. e The
labeling pattern in 7c was determined from a mixture of 6c, 7c, and
8c, the ratios of which were determined accurately by GC.
benzene gave log[kH/M-1 s-1] ) 0.004ET + 6.667. Values of
kH determined from this equation were used to determine the
values of kr given in Table 1. A plot of the data for kr gave
log[kr/s-1] ) 0.024ET + 3.882 (r ) 0.978). This indicates a
weak dependence of kr on solvent polarity by comparison with
purely ionic reactions but a substantial one for a radical reaction.
For a limited survey of substituent effects the rate constants
kr and kn for the p-methoxy- and p-cyano-substituted radicals
5b9 and 5c9 were determined at 75 °C15 by the stannane method
and compared with those for 5a. As expected for a reaction
with little polar character, kn for 5b, 5a, and 5c (7.6, 3.8, and
100 × 103 s-1, respectively) correlated well with Creary’s
radical substituent parameter16 (σc•) but not with σp+. However,
in accord with the view that the ester migration involves a polar
transition structure, kr for 5b, 5a, and 5c (16.9, 6.0, and 1.6 ×
104 s-1, respectively) showed a good correlation (r ) 0.985)
with σp+, the F value (-0.71) being substantial for a radical
reaction. As expected for a substrate (5d) containing the highly
electron-attracting trifluoroacetoxy group, kr at 75 °C (2.5 ×
106 s-1) was relatively large (cf., butyrate 5c kr ) 1.6 × 104
s-1).17
Although the solvent and substituent effects clearly support
the hypothesis that the â-(acyloxy)alkyl radical rearrangement
of 5a involves a polarized transition state and/or intermediate,
the key question remains of whether the mechanism changes
with change of solvent or substitution.
In an attempt to resolve this question, the three radical
precursors 4a-c labeled with 17O mainly in the carbonyl group18
were treated with Bu3SnH and AIBN.3,5 The labeling patterns
in the bromides and the rearranged products 6a-c were
determined by the 17O NMR method.3 The results (Table 2)
reveal that under conditions that do not greatly favor a highly
polarized transition state or an intermediate similar to 2 (e.g.,
the rearrangements of 5a and 5c in benzene), complete
transposition of the ester oxygens occurs. Under conditions that
do favor such a polarized transition state or intermediate (e.g.,
rearrangement of 5b in benzene and of 5a in methanol),
considerable scrambling of the ester oxygens occurs. However,
attempts to trap intermediate radical-cation/anion pairs such as
2 with methanol or with benzoate or azide anions were
unsuccessful.
The significance of the present work is that it clearly
establishes a relationship between electronic environment,
rearrangement rate, and the degree of oxygen scrambling in a
single system. Conditions expected to favor the formation of
dipolar transition states or intermediates accelerate the rate of
the rearrangement and increase the degree of scrambling of the
ether and carbonyl oxygen atoms in the migrating acyloxy
group.
The most straightforward interpretation is that radicals such
as 5 can undergo an acyloxy shift by more than one mechanism.
One of these must be relatively nonpolar in nature and involves
a five-membered transition state similar to 1.19 Another must
be relatively dipolar and may involve a radical-cation/anion pair
similar to 2 or a highly polarized version of 3. Indeed, 2 and
3 can reasonably be regarded as variations of the same contact
radical-cation/anion pair arising from differing degrees of
tightness. It has recently been suggested20 that all of these
reactions proceed by essentially one mechanism involving the
collapse of an intermediate contact radical-cation/anion pair.
However, until this hypothesis can be further explored experi-
mentally, we believe that the concept of a dichotomy of
mechanism agrees best with all the available evidence, both
experimental and computational.21
(15) Rate constants were calculated from an average of four determina-
tions, each at a different stannane concentration.
(16) Creary, X.; Mehrsheikh-Mohammadi, M. E.; McDonald, S. J. Org.
Chem. 1987, 52, 3254.
(17) A similar difference in rates for the rearrangement of radicals
containing the acetate or trifluoroacetate group has been previously
recorded: Barclay, L. R. C.; Lusztyk, J.; Ingold, K. U. J. Am. Chem. Soc.
1984, 106, 1793.
(18) 17O-Butyryl chloride was prepared as described previously5 and
heated with the appropriate bromohydrin in dimethylaniline at 50-70 °C.
The labeled product was isolated by means of flash chromatography.
(19) The five-membered transition state 1 is thought to have some polar
character: Beckwith, A. L. J.; Radom, L.; Saebo, S. J. Am. Chem. Soc.
1984, 106, 5119.
Supporting Information Available: Details of the synthesis of
substrates and products and descriptions of kinetics and analytical
methods (14 pages). See any current masthead page for ordering and
Internet access instructions.
(20) Sprecher, M. Chemtracts: Org. Chem. 1994, 7, 120.
(21) Zipse, H. J. Chem. Soc., Perkin Trans. 2 1996, 1797.
JA963153O