Free radical chemistry of β-lactones. Arrhenius parameters for the decarboxylative cleavage and ring expansion of 2-oxetanon-4-ylcarbinyl radicals. Facilitation of chain propagation by catalytic benzeneselenol

Article abstract of DOI:10.1021/ja980447w

2-Oxetanon-4-ylcarbinyl radicals undergo facile ring opening with cleavage of the C-O bond to give 3-butenoxyl radicals which in turn suffer loss of carbon dioxide to provide allyl radicals. When the initial radical is generated from a bromolactone with Bu₃SnH and AIBN, chain propagation is poor owing to the relatively slow abstraction of hydrogen from the stannane by the allyl radical. The inclusion of catalytic Ph₂Se₂, reduced in situ to PhSeH, provides for much smoother cleaner reactions because of the better hydrogen donating capacity of the selenol. The oxetanon-4-ylcarbinyl radical derived from 6-benzyl-1-(bromomethyl)-8-oxa-7-oxobicyclo[4.2.0]octane is anomalous and undergoes a radical ring expansion in competition with the fragmentation process. Possible reasons for this anomaly are presented as are Arrhenius functions for the fragmentation and rearrangement. The Arrhenius function for the fragmentation of a simple 2-oxetanon-4-yl radical is also presented. Conditions are described under which the fragmentation of 2- oxetanon-4-yl radicals may be suppressed.

Full text of DOI:10.1021/ja980447w

8298
J. Am. Chem. Soc. 1998, 120, 8298-8304
Free Radical Chemistry of â-Lactones. Arrhenius Parameters for the
Decarboxylative Cleavage and Ring Expansion of
2-Oxetanon-4-ylcarbinyl Radicals. Facilitation of Chain Propagation
by Catalytic Benzeneselenol
David Crich* and Xue-Sheng Mo
Contribution from the Department of Chemistry, UniVersity of Illinois at Chicago, 845 West Taylor Street,
Chicago, Illinois 60607-7061
ReceiVed February 9, 1998. ReVised Manuscript ReceiVed June 4, 1998
Abstract: 2-Oxetanon-4-ylcarbinyl radicals undergo facile ring opening with cleavage of the C-O bond to
give 3-butenoxyl radicals which in turn suffer loss of carbon dioxide to provide allyl radicals. When the
initial radical is generated from a bromolactone with Bu3SnH and AIBN, chain propagation is poor owing to
the relatively slow abstraction of hydrogen from the stannane by the allyl radical. The inclusion of catalytic
Ph2Se2, reduced in situ to PhSeH, provides for much smoother cleaner reactions because of the better hydrogen
donating capacity of the selenol. The oxetanon-4-ylcarbinyl radical derived from 6-benzyl-1-(bromomethyl)-
8-oxa-7-oxobicyclo[4.2.0]octane is anomalous and undergoes a radical ring expansion in competition with the
fragmentation process. Possible reasons for this anomaly are presented as are Arrhenius functions for the
fragmentation and rearrangement. The Arrhenius function for the fragmentation of a simple 2-oxetanon-4-yl
radical is also presented. Conditions are described under which the fragmentation of 2-oxetanon-4-yl radicals
may be suppressed.
Introduction
to 4-pentenyl radicals. In its simplest, unsubstituted form this
rearrangement has a rate constant of 5 × 103 s-1 at 25 °C.11-13
Our interest in the free radical rearrangements of esters and
lactones14-16 led us to consider the possible reactions of
2-oxetanon-4-ylcarbinyl radicals. These might be subject to ring
expansion or fragmentation by cleavage of either the C-C or
C-O bonds (Scheme 1). Cleavage of the C-C bond would
lead to a resonance-stabilized R-carboxyl radical but in the
higher energy anti conformer. Should cleavage of the C-O
bond occur this would be followed by decarboxylation to give
a resonance-stabilized allyl radical. The radical ionic pathways
favored for cleavage of some â-(phosphatoxy)alkyl radicals14
also dictated consideration of a heterolytic pathway for rupture
of the C-O bond; however, we recognized that this would also
be followed by rapid decarboxylation and so would be func-
tionaly indistinguishable from the pure radical pathway. The
fragmentation of â-(acyloxy)alkyl radicals to alkenes and
carboxyl radicals is an extremely rare event and, with one
apparent exception,17 has only been reported when the newly
formed multiple bond is part of an aromatic ring.18
One of the most rapid radical rearrangements known is the
reversible1,2 cleavage of oxiranylcarbinyl radicals to allyloxy
radicals,3,4 for which the rate constant has been estimated, by
competition kinetics, to be ∼1010 s-1 5
. Thus, this rearrangement
is approximately some 2 orders of magnitude more rapid than
the prototypical simple cyclopropylcarbinyl to homoallyl6 ring
opening. A feature of the cleavage of oxiranylcarbinyl radicals,7
is the usual kinetic preference for rupture of the C-O rather
than the C-C bond. On the basis of ab initio MO calculations
Pasto advanced 3.57 kcal‚mol-1 as the activation energy for
cleavage of the C-O bond in the oxiranylmethyl radical8 and
concluded that the inclusion of the oxygen atom in the three-
membered ring not only facilitates cleavage of the ring by
rupture of the C-O bond but also retards cleavage by C-C
bond scission. Somewhere near the other end of the spectrum
of radical rearrangments9,10 is the cleavage of cyclobutylcarbinyl
(1) Marples, B. A.; Rudderham, J. A.; Slawin, A. Z.; Edwards, A. J.;
Hird, N. W. Tetrahedron Lett. 1997, 38, 3599-3602.
(2) Ziegler, F. E.; Petersen, A. K. J. Org. Chem. 1995, 60, 2666-2667.
(3) Grossi, L.; Strazzari, S. J. Chem. Soc., Chem. Commun. 1997, 917-
918.
(4) Krosley, K. W.; Gleicher, G. J. J. Phys. Org. Chem. 1993, 6, 228-
232.
(5) Krishnamurthy, V.; Rawal, V. H. J. Org. Chem. 1997, 62, 1572-
1573.
(6) Newcomb, M.; Glenn, A. G. J. Am. Chem. Soc. 1989, 111, 275-
277.
(7) Barton, D. H. R.; Motherwell, R. S. H.; Motherwell, W. B. J. Chem.
Soc., Perkin Trans. 1 1981, 2363-2367.
(8) Pasto, D. J. J. Org. Chem. 1996, 61, 252-256.
(9) Newcomb, M. Tetrahedron 1993, 49, 1151-1176.
(10) Ingold, K. U.; Griller, D. Acc. Chem. Res. 1980, 13, 317.
(11) Walton, J. C. J. Chem. Soc., Perkin Trans. 2 1989, 173-177.
(12) Beckwith, A. L. J.; Moad, G. J. Chem. Soc., Perkin Trans. 2 1980,
1083-1092.
(13) Ingold, K. U.; Maillard, B.; Walton, J. C. J. Chem. Soc., Perkin
Trans. 2 1981, 970-974.
(14) Beckwith, A. L. J.; Crich, D.; Duggan, P. J.; Yao, Q. Chem. ReV.
1997, 97, 3273-3312.
(15) Choi, S.-Y.; Crich, D.; Horner, J. H.; Huang, X.; Martinez, F. N.;
Newcomb, M.; Wink, D. J.; Yao, Q. J. Am. Chem. Soc. 1998, 120, 211-
212.
(16) Crich, D.; Beckwith, A. L. J.; Filzen, G. F.; Longmore, R. W. J.
Am. Chem. Soc. 1996, 118, 7422-7423.
S0002-7863(98)00447-8 CCC: $15.00 © 1998 American Chemical Society
Published on Web 07/29/1998

Free Radical Chemistry of â-Lactones
J. Am. Chem. Soc., Vol. 120, No. 33, 1998 8299
Scheme 1
Scheme 3
The pericyclic extrusion of CO2 from â-lactones is a well-
known reaction.26 However, the rate constants (10-1-10-4
s-1 27-30
for this type of cleavage are typically several orders
)
of magnitude slower than even the cyclobutylcarbinyl rear-
rangement and therefore such chemistry was not expected to
be a serious complication in this study. Here, we report in full
our study of the free radical chemistry of 4-oxetanonylcarbinyl
radicals.31
Scheme 2
Results and Discussion
Initially we prepared the known spirocyclic â-lactone 132 and
33
subjected it to reduction by 2 equiv of SmI2 in THF at -78
°C, followed by warming to room temperature when the
triarylmethane 5 was isolated in 76% yield. This result may
be rationalized in terms of rapid fragmentation of the initial
Sm(III) ketyl 2 to give carboxyl radical 3, which then decar-
boxylates to provide 4. A second electron transfer then gives
the corresponding triarylmethyl anion which, on workup,
provides 5 (Scheme 3). We cannot altogether rule out the
possibility that the initial ketyl undergoes a second one-electron
reduction to give a dianion, which is followed by a two-electron
fragmenation and decarboxylation. However, given that the
triphenylacetoxy anion does not undergo decarboxylation in
alkaline aqueous solution at room temperature unless in a
photoexcited state,34 we consider this to be unlikely.
We then turned to the reaction of bromolactones with
stannanes. Substrate 7 was prepared by hydrogenation of the
â-lactone 6,35 which, in turn, like 824 was obtained from the
seco acid by bromolactonization. A further bromo-â-lactone
12 was obtained by Wittig olefination of 9,36 followed by
saponification, and kinetic bromolactonization (Scheme 4).
Indirect evidence for the heterolytic cleavage of some
R-oxygen-substituted â-(acyloxy)alkyl radicals in polar media
has been advanced by the groups of Norman and Schulte-
Frohlinde.14,19-21 On the basis of the differing behavior of
oxiranylcarbinyl and cyclopropylcarbinyl radicals, we reasoned
that the most likely pathway would be b (Scheme 1). Moreover,
we reasoned that any such cleavage would occur several orders
of magnitude more rapidly than the cyclobutylcarbinyl rear-
rangement and so be observable under preparative conditions.
The literature provided a single example of the ring opening of
an oxetanylcarbinyl radical which, indeed, occurred with
preferential cleavage of the C-O bond.22 The literature also
revealed three examples of the reduction of 4-bromoalkyl-â-
lactones (Scheme 2) which were reported to proceed in good
yield, with no mention of the type of fragmentation anticipated
here.23-25
Reaction of 6 with Bu3SnH and AIBN at 30 °C, under
irradiation from a sunlamp provided, after 38 h, the hydrocarbon
(24) Shibata, I.; Toyota, M.; Baba, A.; Matsuda, H. J. Org. Chem. 1990,
55, 2487-2491.
(25) Askani, R.; Keller, U. Liebigs 1988, 61-68.
(26) Pommier, A.; Pons, J.-M. Synthesis 1993, 441-449.
(27) Ocampo, R.; Dolbier, W. R.; Bartberger, M. D.; Paredes, R. J. Org.
Chem. 1997, 62, 109-114.
(28) Imai, T.; Nishida, S. J. Org. Chem. 1980, 45, 2354-2359.
(29) Mulzer, J.; Zippel, M.; Bruntrup, G. Angew. Chem., Int. Ed. Engl.
1980, 19, 465-466.
(30) Krabbenhoft, H. O. J. Org. Chem. 1978, 43, 1305-1311.
(31) Preliminary communication: Crich, D.; Mo, X.-S. J. Org. Chem.
1997, 62, 8624-8625.
(32) Zaugg, H. E. Org. React. 1954, 8, 305-363.
(33) Molander, G. A.; Harris, C. R. Chem. ReV. 1996, 96, 307-338.
(34) Faria, J. L.; Steenken, S. J. Am. Chem. Soc. 1990, 112, 1277-1279.
(35) Ganem, B.; Holbert, G. W.; Weiss, L. B.; Ishizumi, K. J. Am. Chem.
Soc. 1978, 100, 6483-6491.
(17) Van Arnum, S. D.; Stepsus, N.; Carpenter, B. K. Tetrahedron Lett.
1997, 38, 305-308.
(18) Barton, D. H. R.; Dowlatshahi, H. A.; Motherwell, W. B.; Villemin,
D. J. Chem. Soc., Chem. Commun. 1980, 732.
(19) Gilbert, B. C.; Larkin, J. P.; Norman, R. O. C. J. Chem. Soc., Perkin
Trans. 2 1972, 794-802.
(20) Behrens, G.; Koltzenberg, G.; Schulte-Frohlinde, D. Z. Naturforsch.
1982, 37c, 1205-1227.
(21) Behrens, G.; Bothe, E.; Koltzenberg, G.; Schulte-Frohlinde, D. J.
Chem. Soc., Perkin Trans. 2 1980, 883-889.
(22) Kim, S.; Lim, K. M. J. Chem. Soc., Chem. Commun. 1993, 1152-
1153.
(23) Mead, K. T.; Park, M. Tetrahedron Lett. 1995, 36, 1205-1208.

8300 J. Am. Chem. Soc., Vol. 120, No. 33, 1998
Crich and Mo
Scheme 4
Scheme 5
Scheme 6
16 in 63% isolaed yield. We envisage 16 as having been formed
by the anticipated radical fragmentation followed by oxidation
of the resulting cyclohexadienyl radical 15 and loss of a proton
(Scheme 5). The formal oxidation of cyclohexadienyl-type
radicals to aromatic systems is a common occurrence in the
chemistry of radical addition to arenes and is additionally
characterized by poor chain propagation and the requirement
for considerable amounts of radical initiator.37 Indeed, to obtain
the recorded result, it was necessary to work with 0.4 mol equiv
of AIBN.
The saturated analogue (7) of 6 was irradiated with the
sunlamp in benzene at 60 °C in the presence of Bu3SnH and
1
AIBN for 8 h, after which examination by H NMR spectros-
copy revealed a complex mixture devoid of substrate. Prepara-
tive TLC enabled isolation of an 88:12 mixture of isomeric
olefins 18 and 19 in 22% yield and a complex mixture of dimeric
products in 31% yield. Again these results were readily
interpreted in terms of the anticipated radical fragmentation
giving the allylic radical 17 followed either by quenching by
the stannane to give the two olefins or by dimerization (Scheme
6). The isolation of the dimers in greater combined yield than
the two olefins suggests that chain propagation was not efficient,
and this was again reflected in the requirement for 0.4 mol equiv
of AIBN as an initiator for the reaction to proceed to completion.
We reasoned that the poor chain propagation might be overcome
by the inclusion of a catalytic quantity of Ph2Se2, reduced in
situ to PhSeH (and PhSeSnBu3),37-39 which, with its superior
hydrogen donating capacity,40 would overcome this problem.
The benzeneselenol is constantly regenerated by reaction of the
PhSe• radical with stoichiometric Bu3SnH. In the event,
working at 40 °C in benzene, using only 10 mol % of di-tert-
butyl peroxalate (DBPO) as an initiator and 10 mol % of Ph2Se2
(2.1 × 10-3 M) and 1.35 equiv of Bu3SnH, the bromolactone
was consumed in less than 3 h and the olefin 18 isolated in
85% yield. Similarly, in benzene reflux, the inclusion of 10
mol % of Ph2Se2 in the reaction mixture enabled the amount of
AIBN to be reduced to 10 mol % when smooth conversion of
the substrate to alkene 18, isolated in 84% yield after only 1.5
h. Thus, the catalytic quantity of PhSeH resulted in dramatic
reductions of the reaction times and in much cleaner, smoother
reactions. No dimers were formed in these PhSeH-catalyzed
reactions, indicating that the selenol operates by efficient
quenching of the allyl radical 17. Moreover, quenching of 17
was highly regioselective, providing only 18 (18:19 > 95:5).
A further example was provided by the monocyclic â-lactone
8, a simple model for the literature reactions presented in
Scheme 2. In line with the above results, when the reaction
was conducted with excess Bu3SnH in the absence of catalytic
PhSeH, it was slow and complex. However, operating in the
presence of 5 mol % Ph2Se2 (2.6 × 10-3 M) and 10 mol %
Scheme 7
AIBN in benzene at reflux provided a very clean reaction
mixture consisting only of the isomeric olefins 22 and 23. No
evidence was found for the formation of the simple reduction
product 24 under these conditions (Scheme 7). However, when
a similar reaction was conducted at room temperature in the
presence of DBPO (10 mol %) as initiator and using a full
equivalent of Ph2Se2 (5.3 × 10-2 M), <5% of 22 and 23 was
1
formed as judged by H NMR spectroscopy but the reduction
product 24 was isolated in 80% yield. A high concentration of
PhSeH is therefore effective in preventing the radical fragmen-
tation reaction.
Treatment of â-lactone 12 with Bu3SnH and AIBN alone was
again inefficient, owing to poor chain propagation by the
intermediate allyl radical. However, exposure of 12 to Bu3SnH
and 5 mol % of Ph2Se2 (1.5 × 10-3 M), with initiation by only
5% of DBPO at room temperature, resulted in complete
consumption of 12 after 3 h. 1H NMR spectroscopy revealed
a clean reaction mixture in which olefin 27, isolated in 78%
yield, was accompanied by two other minor products. The
experiment was repeated with 30 mol % of Ph2Se2 (9.0 × 10-3
M) when 27 and the two, previously minor, products were
formed in the ratio 1:1.8:0.4. Chromatographic separation
enabled isolation of 27 in 25% yield and a 4:1 mixture of the
two other products, which spectroscopic investigation revealed
to be the reduction product 28 and the rearranged lactone 30,
respectively (Scheme 8). Enhancement of the benzylic meth-
(36) Hodgson, A.; Marshall, J.; Hallett, P.; Gallagher, T. J. Chem. Soc.,
Perkin Trans. 1 1992, 2169-2174.
(37) Crich, D.; Hwang, J.-T. J. Org. Chem. 1998, 63, 2765-2770.
(38) Crich, D.; Yao, Q. J. Org. Chem. 1995, 60, 84-88.
(39) Crich, D.; Hwang, J.-T.; Liu, H. Tetrahedron Lett. 1996, 37, 3105-
3108.
1
ylene protons in the H NMR spectrum on double irradiation
of the bridgehead hydrogen led to the assignment of 30 as the
cis-fused structure depicted. In a further experiment using a
(40) Newcomb, M.; Varick, T. R.; Ha, C.; Manek, M. B.; Yue, X. J.
Am. Chem. Soc. 1992, 114, 8158-8163.

Free Radical Chemistry of â-Lactones
J. Am. Chem. Soc., Vol. 120, No. 33, 1998 8301
Scheme 8
log(kH) ) 10.35 - 1.76/(2.3RT)
(1)
41
,
kF is the rate constant for the fragmentation providing 27,
and kR is that for the rearrangement leading ultimately to 30.
In this manner the rate constants kF(25 f 26) and kR(25 f 29)
were determined to be 3.8 × 106 and 1.8 × 106 s-1, respectively,
at 40 °C.
Subsequently the reaction of 12 with Bu3SnH was carried
out in the presence of a fixed concentration of PhSeH over a
85° range of temperature enabling Arrhenius plots to be made
for the fragmentation and rearrangement of radical 25. Both
were straight lines which enabled the extraction of the Arrhenius
equations (eqs 2 and 3).
log(kF) ) (8.3 ( 0.3) - (2.5 ( 0.4)/(2.3RT)46
log(kR) ) (8.2 ( 0.2) - (2.7 ( 0.2)/(2.3RT)46
(2)
(3)
Next we turned to the fragmentation of â-lactone 8. In this
instance we directly studied the fragmentation at a range of
different temperatures enabling determination of the Arrhenius
equation (eq 4) in which kF′ is the rate constant for this
full molar equivalent of Ph2Se2 (3.0 × 10-2 M), the reduced
log(kF′) ) (13.3 ( 0.2) - (8.7 ( 1.0)/(2.3RT)46 (4)
1
product 28 was formed in >95% yield as determined by H
NMR spectroscopy of the crude reaction mixture. Thus, in one
or another of the various examples, all distinguishable processes
in Scheme 1 with the exception of C-C bond cleavage have
been observed.
fragmentation. The Arrhenius parameters for the fragmentation
of radical 20 (eq 4) appear perfectly straightforward for such a
radical fragmentation reaction and invite no special comment.
They lead to a rate constant of 8.2 × 106 s-1 for the opening of
this radical at 25 °C. This is some 3 orders of magnitude greater
than that of the cyclobutylmethyl radical,11-13 as predicted in
the initial analysis of the problem. With these measured values
of the rate constants for fragmentation, we can now reconsider
the previous literature reports of the reactions of oxetanoyl-
carbinyl radicals. The rate constant for the fragmentation of 8
at 80 °C is calculated to be 8.1 × 107 s-1. From this rate
constant and that calculated for trapping of the n-butyl radical
by Bu3SnH at 80 °C (6.3 × 106 s-1),47 we calculate that a
Bu3SnH concentration of 50 M would be necessary to obtain
an 80% yield of the reduction product 24 and a 20% yield of
the combined fragmentation products (22 and 23) in benzene
at reflux, without catalysis by the selenol. This concentration
of Bu3SnH is significantly higher than that reported by Shibata
(Scheme 2, 0.5 M)24 and must cast some doubt on the reported
yield. Similarly, we calculate that Mead operated with Bu3SnH
concentration of at least 9 × 10-2 M to obtain a 70% yield of
reduction product at -78 °C (Scheme 2).23 These yields,
temperatures, and concentrations (Scheme 2) are to be contrasted
with the 80% isolated yield of 24 isolated on reducing 8 in
benzene at room temperature with 0.13 M Bu3SnH, assisted by
0.05 M PhSeH (Scheme 7). The lack of fragmentation in the
third example of Scheme 225 presumably derives from the fact
that this would result in the formation of a cyclobutene whose
strain would offset that lost on â-lactone opening.
We next turned to the full kinetic characterizaton of these
new radical rearrangements. First, we decided to investigate
the cleavage/fragmentation of 12 as this was the only example
in which a rearrangement was observed, and second, we chose
to study 8 as being more representative of the type of â-lactone
most likely to be encountered in synthetic schemes. As usual
we have made use of our catalytic adaptation41 of Newcomb’s
PhSeH clock reaction.40 In this method trapping by a catalytic
quantity of PhSeH at a known concentration is used as the clock
reaction. The PhSeH is constantly regenerated by reaction of
the PhSe• radical with stoichiometric Bu3SnH. As with any
type of indirect radical clock method, the validity of the result
depends on the applicability of the clock reaction employed.
Here, the assumption is made that the clock reaction, quenching
of primary alkyl radicals by PhSeH (and Bu3SnH), is not
affected by the â-acyloxy group. â-Oxygen effects in pure
radical,42-44 as opposed to radical cation,45 reactions are typically
small, and the assumption should be correct to a first ap-
proximation.
Initially, 12 was reacted at 40 °C with Bu3SnH and DBPO
as an initiator in the presence of a range of concentrations of
PhSeH. This enabled plots to be made of the ratios of the
reduction product (28) with both the fragmentation product (27)
and the rearrangement product (30) against the concentration
of PhSeH. In each case linear plots were obtained whose slopes
are equal to kH/kF and kH/kR, respectively, wherein kH is the
rate constant for trapping of a primary alkyl radical by PhSeH
at 40 °C, calculated from the known Arrhenius equation (eq 1)
In contrast to eq 4, the Arrhenius parameters for the
fragmentation and rearrangement (eqs 2 and 3, respectively) of
radical 25 are somewhat unusual. Indeed the comportment of
radical 25 is unusual, at least among the examples studied here,
insofar as it provided the only example in which a detectable
amount of the ring expansion process was observed. The
activation energies in eqs 2 and 3, while low, are reasonably
consistent with those for other highly exothermic radical
(41) Crich, D.; Jiao, X.-Y.; Yao, Q.; Harwood: J. S. J. Org. Chem. 1996,
61, 2368-2373.
(42) Crich, D.; Beckwith, A. L. J.; Chen, C.; Yao, Q.; Davison, I. G. E.;
Longmore, R. W.; Anaya de Parrodi, C.; Quintero-Cortes, L.; Sandoval-
Ramirez, J. J. Am. Chem. Soc. 1995, 117, 8757-8768.
(43) Busfield, W. K.; Grice, I. D.; Jenkins, I. D. J. Chem. Soc., Perkin
Trans. 2 1994, 1079-1086.
(44) Busfield, W. K.; Grice, I. D.; Jenkins, I. D.; Monteiro, M. J. J.
Chem. Soc., Perkin Trans. 2 1994, 1071-1077.
(45) Crich, D.; Mo, X.-S. J. Am. Chem. Soc. 1997, 119, 249-250.
(46) Errors are at the 95% confidence interval (3.2 s).
(47) Chatgilialoglu, C.; Ingold, K. U.; Scaiano, J. C. J. Am. Chem. Soc.
1981, 103, 7739-7742.

8302 J. Am. Chem. Soc., Vol. 120, No. 33, 1998
Crich and Mo
Scheme 9
Returning to the theme of catalysis of chain transfer by
PhSeH, we note that this was very successful when the
intermediate radical was an allyl radical, as in Schemes 6-8.
However, PhSeH failed to catalyze the quenching of of the
cyclohexadienyl 15. Thus, even in the presence of a full 1 equiv
of PhSeH, the cleavage of 6 was unchanged and provided the
oxidative cleavage product 16 (Scheme 5). This causes us to
draw attention to recent work from this laboratory37 in which it
was found that cyclohexadienyl 32 was readily quenched by
PhSeH, in a catalytic system similar to the one used here, giving
good isolated yields of cyclohexadienyls. In contrast it was
found that the isomer 33 underwent “oxidation” even in the
presence of PhSeH. As was noted,37 the cyclohexadiene C-H
and PhSeH Se-H bond energies are very similar and any
additional stabilization of the cyclohexadienyl radical by the
substituents will prevent effective reduction by the selenol.
fragmentation reactions, e.g. the cubylcarbinyl radical (EA
)
3.7 kcal‚mol-1), the 2,2-diphenylcyclopropylcarbinyl radical (EA
) 2.0 kcal‚mol-1), and the bicyclo[2.1.0]pent-2-yl radical (EA
) 5.2 kcal‚mol-1).48-51 The log A values are however unusually
small, both for fragmentations9,52 and for acyloxy migrations,14,53
and suggest52 tight transition states for these reactions. The close
similarity between the two equations (eqs 2 and 3) and the log
A values suggest to us that both reactions proceed along the
same reaction coordinate to a bridged transition state (31) then
diverge to the fragmentation and rearrangement products
(Scheme 9). Transition state 31 is directly analogous to the
three-electron, three-center cyclic transition state for the 1,2-
migration of â-(acyloxy)alkyl radicals, for which there is ample
literature precedent.14 The alternative five-electron, five-center
cyclic transition state, corresponding to that of a 2,3-migration
of an acyloxyalkyl radical, would be highly strained and requires
the unlikely closure of the initial radical 25 onto the â-lactone
carbonyl oxygen. Moreover, the observed rate constant for the
rearrangement of 25 to 29 is consistent with those previously
determined for 1,2-acyloxy shifts but not with those for the
corresponding 2,3-shifts.14 An alternative pathway involving
radical ionic fragmentation (Scheme 1, path c), followed by
competing fragmentation and recombination, is considered to
be highly unlikely on the grounds that extensive experimentation
has so far failed to provide any evidence for such a mechanism
in â-(acyloxy)alkyl radical rearrangements.14 We also note that
radical 25 has the potential to undergo cleavage of the exocyclic
C-O, of an exocyclic C-C bond, or of the transannular C-C
bond. In this it is not unlike numerous bicyclo[3.2.0]heptan-
1-oxyl radicals which prefer cleavage of the transannular bond.54
It seems likely, as Dowd and Zhang have suggested,55 that in
such 1-bicyclo[4.2.0]octylalkyl and oxyl radicals and their
[3.2.1]heptyl congenors there is a better overlap of the singly
occupied orbital with the transannular bond than with the
exocyclic bonds, which leads to preferential cleavage of the
former. In the context of radical 25 this would mean that
cleavage of the C-O is retarded. The alternate formation of
the strained transition state (31) for the acyloxy migration
therefore becomes possible (Scheme 9).
Finally, we raise the possibility of an additional mechanism
for the radical fragmentations presented here. We have
considered the fragmentations in terms of stepwise processes
involving expulsion of â-acyloxy radicals followed by rapid
decarboxylation (Scheme 1). We have also noted the possibility
of a radical ionic fragmentation followed by decarboxyation
(Scheme 1). We ruled out the possibility of concerted loss of
CO2 either from the substrate followed by reduction of an allyl
bromide or from the reduction products owing to the known
kinetics of this type of concerted loss of CO2 from â-lactones.
Moreover, all experiments were conducted at temperatures at
which blank experiments indicated the â-lactones to be stable,
at least on the time scale of the reactions. However, and without
being able to provide a solution at the present time, we feel
justified in raising the possibility of a concerted loss of CO2
from the first formed radical (Scheme 10). The suggestion is
Scheme 10
that the 4π-pericyclic fragmentation may be accelerated by
several orders of magnitude by conjugation with the exocyclic
radical. We are prompted to offer this suggestion by the
enormous rate enhancements, of up to 10 orders of magnitude,
seen when an exocyclic anion is conjugated with a pericyclic
transition state, as in the oxyanion Cope rearrangment.56-58 At
the present time the prospects for distinguishing between the
stepwise radical pathway and the concerted pathway do not
appear good; the rate constant for the reactions measured here
and the near diffusion controlled loss of CO2 from carboxyl
radicals conspire against the possibility of trapping any such
intermediate species.
(48) Choi, S.-Y.; Eaton, P. E.; Newcomb, M.; Yip, Y. C. J. Am. Chem.
Soc. 1992, 114, 6326-6329.
(49) Newcomb, M.; Manek, M. B.; Glenn, A. G. J. Am. Chem. Soc.
1991, 113, 949-958.
(50) Bowry, V. W.; Lusztyk, J.; Ingold, K. U. J. Am. Chem. Soc. 1991,
113, 5687-5698.
(51) Newcomb, M.; Johnson, C. C.; Manek, M. B.; Varick, T. R. J. Am.
Chem. Soc. 1992, 114, 10915-10921.
(52) Benson, S. W. Thermochemical Kinetics, 2nd ed.; Wiley: New York,
1976.
(53) Another isolated example of a â-(acyloxy)alkyl rearrangement, that
of the 2,3,4,6-tetraacetylglucopyranos-1-yl radical, has a log A of 8.1: Korth,
H.-G.; Sustmann, R.; Groninger, K. S.; Leisung, M.; Giese, B. J. Org. Chem.
1988, 53, 4364-4369.
(54) Dowd, P.; Zhang, W. J. Org. Chem. 1992, 57, 7163-7171.
(55) Dowd, P.; Zhang, W. Chem. ReV. 1993, 93, 2091-2115.
(56) Evans, D. A.; Golob, A. M. J. Am. Chem. Soc. 1975, 97, 4765-
4766.
(57) Gajewski, J. J.; Gee, K. R. J. Am. Chem. Soc. 1991, 113, 967-
971.
(58) Yoo, H. Y.; Houk, K. N.; Lee, J. K.; Scialdone, M. A.; Meyers, A.
I. J. Am. Chem. Soc. 1998, 120, 205-206.

Free Radical Chemistry of â-Lactones
J. Am. Chem. Soc., Vol. 120, No. 33, 1998 8303
spectroscopy revealed only 2264 and (E)- and (Z)-2362 in the ratio 1:2:
4. No evidence was found, by H NMR spectroscopy, for formation
Experimental Section
1
General Procedures. 1H and 13C NMR spectra were recorded as
CDCl3 solutions at 300 and 75 MHz, respectively, with chemical shifts
(δ) in parts per million downfield from tetramethylsilane as an internal
standard. All solvents were dried and distilled by standard methods.
Extracts were dried over Na2SO4 and solvents removed in vacuo.
Microanalyses were performed by Midwest Microlabs (Indianapolis,
IN).
of reduction product 24. When a similar reaction was conducted
without Ph2Se2 a complex mixture was obtained. Pertinent spectral
data for 22, and (E)- and (Z)-23 were in good agreement with the
1
literature.64,62 22: H NMR δ 2.20 (td, J ) 7.7, 6.6 Hz, 2H); 2.60 (t,
J ) 7.7 Hz, 2H), 4.85 (m, 2H), 5.70 (m, 1H), 7.1 (m, 5H). (E)-23: 1H
NMR δ 1.70 (d, J ) 6.0 Hz, 3H), 3.30 (d, J ) 6.3 Hz, 2H), 5.55 (m,
2H), 7.25 (m, 5H). (Z)-23: 1H NMR δ 1.73 (d, J ) 4.9 Hz, 3H), 3.40
(d, J ) 5.1 Hz, 2H), 5.55 (m, 2H), 7.25 (m, 5H).
Reaction of 1 with SmI2. To a stirred solution of 132 (76 mg, 0.25
mmol) in THF/MeOH (3/1) (10 mL) was added dropwise SmI2 (5 mL,
0.1 M in THF) by a syringe at -78 °C under Ar. The reaction was
complete within 2 min. Saturated NH4Cl was then added and the
aqueous phase extracted with CH2Cl2. The organic layer was dried
and concentrated. The residue was dissolved in a minimum of CH2Cl2
and treated with hexane to give a yellow solid 5,59,60 which was collected
Reaction of 8 with Bu3SnH, DBPO, and 100% PhSeH in Benzene
at Room Temperature. To a solution of 8 (27 mg, 0.106 mmol),
Ph2Se2 (33 mg, 1 equiv), and DBPO (2.5 mg, 0.1 equiv) in benzene (2
mL) under Ar at room temperature was added Bu3SnH (66 µL, 2.3
equiv), followed by stirring at room temperature for 4 h. Removal of
solvent and inspection of the reaction mixture by 1H NMR spectroscopy
revealed 24 (>95%) and 22 and 23 (<5%). The reaction mixture was
dissolved in CH2Cl2 and MeOH and treated with NaBH4 at 0 °C and
quenched by the addition of water. The aqueous phase was extracted
with CH2Cl2, and the combined organic solvents were dried (Na2SO4)
and concentrated to give a residue, which after TLC on silica gel
(eluent: hexane/EtOAc ) 9/1) afforded 2424 (14.6 mg, 80%) as an
oil: 1H NMR δ 1.44 (d, J ) 6.1 Hz, 3H), 3.02 (dd, J ) 9.31, 14.5 Hz,
1H), 3.16 (dd, J ) 5.7, 14.5 Hz, 1H), 3.46 (m, 1H), 4.45 (m, 1H), 7.27
(m, 5H); IR (neat) ν (cm-1) 1810.
1
by filtration (49.5 mg, 76%): mp 107-109 °C (CH2Cl2/hexane); H
NMR δ 5.02 (br. s, 1H), 5.46 (s, 1H), 6.69 (d, J ) 10.5 Hz, 2H), 6.92
(d, J ) 10.5 Hz, 2H), 7.20 (m, 10H); IR (neat) ν (cm-1) 3388.
Reaction of 6 with Bu3SnH and AIBN. A solution of 635 (70 mg,
0.24 mmol), Bu3SnH (100 µL, 1.5 equiv), and AIBN (17 mg, 0.4 equiv)
in benzene (3 mL) was irradiated at 30 °C with a sunlamp with
monitoring by TLC. After 38 h at the same temperature, 6 was
consumed. Removal of the volatiles afforded a residue, which was
purified by preparative TLC on silica gel (eluent: hexane) to give 1661
as a colorless oil (25.4 mg, 63%): 1H NMR δ 3.99 (s, 2H), 7.2 (m,
6H), 7.32 (m, 4H).
Reaction of 12 with Bu3SnH, DBPO, and 5% of PhSeH. 12 (61.8
mg, 0.2 mmol), Ph2Se2 (3.1 mg, 0.05 equiv), and DBPO (2.4 mg, 0.05
equiv) were dissolved in benzene (6.5 mL) under Ar and treated with
Bu3SnH (66 µL, 1.2 equiv) at room temperature, followed by stirring
for 3 h. The solvent was then removed in vacuo whereupon inspection
Reaction of 7 with Bu3SnH and AIBN. A mixture of 7 (120.4
mg, 0.407 mmol), Bu3SnH (178 µL, 1.5 equiv), and AIBN (29 mg,
0.4 equiv) in benzene (4 mL) was irradiated at 60 °C with a sunlamp
for 8 h. After the solvent was removed in vacuo, examination of the
reaction mixture by 1H NMR spectroscopy showed a complex reaction
mixture containing several olefins and the absence of starting material.
Preparative TLC on silica gel (eluent: hexane/EtOAc ) 16/1) enabled
the separation of two closely migrating bands. One band contained
the dimers, as a complex mixture of diastereomers (22 mg, 32%);
HRMS: Calcd for C26H30 342.2347 (M+•), found 342.2346), while the
other was found to consist of 18 and its regioisomer 19 in the ratio
1
of the reaction mixture by H NMR spectroscopy indicated that only
trace amounts of 28 and 30 were formed in addition to 27. Column
chromatography on silica gel (eluent: hexane/CHCl3 ) 3:1) gave pure
2765 (29 mg, 78%) as a colorless oil: 1H NMR δ 1.59 (m, 4H), 1.74
(s, 3H), 1.88 (m, 2H), 2.03 (m, 2H), 3.37 (s, 2H), 7.17 (m, 3H), 7.27
(m, 2H).
Reaction of 12 with Bu3SnH, DBPO, and 30% of PhSeH. The
above experiment was repeated with 100 mg of 12 and 30% of PhSeH.
1
Removal of solvent gave a mixture, inspection of which by H NMR
1
88:12 as determined by H NMR spectral analysis (15.4 mg, 22%).
spectroscopy revealed that not only 27 but also 28 and 30 were
generated in the ratio 1:1.8:0.4. The reaction mixture was again
dissolved in CH2Cl2 (5 mL) and methanol (1 mL) and treated with
NaBH4 at 0 °C. After quenching with water and extraction of the
aqueous phase with CH2Cl2, the combined organics were dried and
concentrated to a residue. After preparative TLC on silica gel (eluent:
CHCl3), this afforded 27 (15 mg, 24.8%) and a mixture of 28 and 30
as a mixture in the ratio ∼4:1 (41 mg, 55%). Anal. Calcd for
C15H18O2: C, 78.23; H, 7.88. Found: C, 77.93; H, 7.74%. Pure 28
was obtained in the form of a white solid by triturating the mixture of
18:62 1H NMR δ 1.26-1.57 (m, 4H), 1.86 (m, 2H), 2.02 (m, 2H), 3.25
(s, 2H), 5.47 (s, 1H), 7.20-7.26 (m, 5H). Olefin 1963 was confirmed
by the following diagnostic signals: 1H NMR δ 2.38 (m, 1H), 2.55
(dd, J ) 8.0, 13.1 Hz, 1H), 2.62 (dd, J ) 7.0, 13.1 Hz, 1H), 5.59 (m,
1H), 5.68 (m, 1).
Reaction of 7 with Bu3SnH, AIBN or DBPO, and PhSeH. To a
solution of 7 (27.4 mg, 0.093 mmol), Ph2Se2 (2.9 mg, 0.1 equiv), and
DBPO (2.2 mg, 0.1 equiv) in benzene (3 mL) under Ar was added
dropwise Bu3SnH (33 µL, 1.2 equiv) by syringe at 40 °C. After the
mixture was stirred for 3 h at that temperature, the solvent was
1
28 and 30 with hexane: mp 100-102 °C (dec); H NMR δ 1.62 (m,
1
8H), 1.75 (m, 2H), 2.76 (d, J ) 14.3 Hz, 1H), 3.28 (d, J ) 14.3 Hz,
1H), 7.28 (m, 5H); 13C NMR δ 16.0, 17.8, 23.2, 26.0, 31.1, 37.0, 58.7,
83.2, 127.0, 128.5, 130.3, 136.1, 174.6; IR (KBr) ν (cm-1) 1810. 30,
which was not obtained pure, was identified by the following diagnostic
signals: 1H NMR δ 2.24 (m, 1H), 2.86 (d, J ) 13.7 Hz, 1H), 3.07 (d,
J ) 13.7 Hz, 1H), 3.97 (dd, J ) 7.8, 8.6 Hz, 1H), 4.06 (dd, J ) 7.4,
8.6 Hz, 1H); 13C NMR δ 21.7, 24.1, 30.3, 36.4, 40.0, 46.8, 69.0, 127.1,
128.7, 130.4, 137.0, 181.0; IR (KBr) ν (cm-1) 1761.
evaporated to give a residue. Inspection of this residue by H NMR
spectroscopy revealed 18 to be the exclusive product. Preparative TLC
on silica gel (eluent: hexane/EtOAc ) 16/1) gave pure 18 (13.6 mg,
85%), which was identical to that described above. A similar
experiment was carried out with 10% of AIBN in place of DBPO in
refluxing benzene for 1.5 h, when an 84% yield of 18 was isolated.
Reaction of 8 with Bu3SnH, AIBN, and 5% PhSeH in Benzene
at Reflux. To a solution of 824 (132 mg, 0.52 mmol), Ph2Se2 (8 mg,
0.05 equiv), and AIBN (10 mg, 0.1 equiv) in benzene (10 mL) under
Ar was added Bu3SnH (188 µL, 1.35 equiv) by a syringe. The reaction
mixture was brought to reflux with stirring for 3 h. The reaction
mixture was then cooled to room temperature, and the solvent was
removed to afford a residue of which inspection by 1H NMR
Reaction of 12 with Bu3SnH, DBPO, and 100% of PhSeH. A
experiment similar to that described above was conducted in the
presence of 100% of PhSeH by using 100 mg of 12. Inspection of the
1
reaction mixture by H NMR spectroscopy revealed 28 (>95%), 30
(<2%), and 27 (<2%). 28 was isolated in 85% yield.
Determination of the Kinetics of Decarboxylation and Rear-
rangement of â-Lactone 12. Stock solutions of 12 (100 mg, 0.3234
mmol) in benzene (5 mL), Ph2Se2 (156 mg, 0.5 mmol) in benzene (5
mL), Bu3SnH (291 mg, 1.0 mmol) in benzene (5 mL), and DBPO (23.4
(59) Burns, J. M.; Wharry, D. L.; Koch, T. H. J. Am. Chem. Soc. 1981,
103, 849-856.
(60) Andrews, A. F.; Mackie, R. K.; Walton, J. C. J. Chem. Soc., Perkin
Trans. 2 1980, 96-102.
(61) Saito, S.; Ohwada, T.; Shudo, K. J. Am. Chem. Soc. 1995, 117,
11081-11084.
(64) Hill, D. H.; Parvez, M. A.; Sen, A. J. Am. Chem. Soc. 1994, 116,
2889-2091.
(62) Arnold, D. R.; Mines, S. A. Can. J. Chem. 1989, 67, 689-698.
(63) Matsushita, H.; Negishi, E.-I. J. Org. Chem. 1982, 47, 4161-4165.
(65) Chaudhari, P. N.; Rao, A. S. Indian J. Chem., Sect. B 1976, 14B,
165-167.

8304 J. Am. Chem. Soc., Vol. 120, No. 33, 1998
Crich and Mo
mg, 0.1 mmol) in benzene (1 mL) under Ar were prepared. The stock
solution of 12 (800 µL, 0.0518 mmol) in benzene was transferred by
a syringe to each of five 25-mL Pyrex test tubes equipped with septa
and stir bars, which had been prepared by evacuation and flushing with
Ar several times beforehand. Stock solutions of 130 µL (0.0130 mmol,
0.25 equiv), 156 µL (0.0155 mmol, 0.3 equiv), 182 µL (0.0181 mmol,
0.35 equiv), 208 µL (0.0207 mmol, 0.40 equiv), or 234 µL (0.0233
mmol, 0.45 equiv) of Ph2Se2 were added to these tubes, respectively,
followed by an amount of Bu3SnH in benzene (stock solution)
corresponding to that of Ph2Se2. After decolorization, further Bu3SnH
stock solution (337 µL, 1.3 equiv) was then added to each tube followed
by sufficient benzene (86-242 µL) to make the volume 1574 µL. Each
tube was then immersed into a boiling CH2Cl2 bath maintained at 40
°C, followed by addition of the stock solution of DBPO in benzene
(26 µL, 0.05 equiv), making the final volume in each tube 1.6 mL.
After the reaction mixtures were stirred at 40 °C for 4 h, the solvent
in Tables 2 and 3. Irradiation with a 140-W medium-pressure mercury
lamp (Hanovia) was necessary when the reactions proceeded at -15
or -25 °C (entries 4 and 5 in Table 2). The solvent was removed in
vacuo, and the residues were examined by 1H NMR spectoscopy, which
gave the ratios of 27/28 and 30/28 recorded in Tables 2 and 3,
respectively (Supporting Information). The data for 40 °C were taken
from the results of Table 1 (Supporting Information).
Determination of the Arrhenius Parameters for Decarboxylation
of 8. Stock solutions of 8 (140 mg, 0.6486 mmol) in benzene (5 mL),
Bu3SnH (291 mg, 1.0 mmol) in benzene (5 mL), Ph2Se2 (156 mg, 0.5
mmol) in benzene (5 mL), and DBPO (23.4 mg, 0.1 mmol) in benzene
(1 mL) under Ar were prepared. A stock solution of lactone 8 (450
µL, 0.0494 mmol), 250 µL of the solution of Ph2Se2 (0.025 mmol,
0.506 equiv), and 450 µL of the Bu3SnH solution (0.089 mmol, 1.81
equiv) were transferred to each of five 25-mL Pyrex test tubes equipped
with septa and stir bars, which were evacuated and flushed with Ar
several times beforehand. Each tube was then equilibrated at the
specific temperature shown in Table 4. To each reaction mixture was
then added 50 µL of the benzene solution of DBPO (0.005 mmol, 0.1
equiv) followed by stirring under Ar for 4 h. Irriadiation with a 140-W
medium-pressure mercury lamp (Hanovia) was necessary when the
reaction proceeded at 11 °C (284 K). The solvents were removed in
vacuo (cold water bath temperature), and the residues were examined
by 1H NMR spectroscopy. In each case the substrate was completely
consumed. Integration of the benzyl signals of 24 at δ 3.16 (PhCHaHb,
Ha, or Hb), 22 at δ 2.60 (PhCH2CH2), (E)-23 at δ 3.30 (PhCH2CH),
and (Z)-23 at δ 3.4 (PhCH2CH) gave the ratios of (22 + 23)/24 recorded
in Table 4 (Supporting Information).
1
was removed in vacuo and the residues were examined by H NMR
spectoscopy. In each case the substrate was completely consumed.
Integration of the benzydryl singals (PhCHaHb) of 28 at δ 3.28 (Ha
or Hb), 30 at δ 3.07 (Ha or Hb), and 27 at 3.37 (Ha ) Hb) gave the
ratios of 28/30 and 28/27 recorded in Table 1 (Supporting Information).
Determination of the Arrhenius Parameters for Decarboxylation
and Rearrangement of 12. Stock solutions of 12 (200 mg, 0.6468
mmol) in toluene (5 mL), Ph2Se2 (156 mg, 0.5 mmol) in toluene (5
mL), Bu3SnH (291 mg, 1.0 mmol) in toluene (5 mL), and DBPO (23.4
mg, 0.1 mmol) in toluene (1 mL) under Ar were prepared. The stock
solution of 12 (400 µL, 0.0518 mmol) was transferred by a syringe to
each of four 25-mL Pyrex test tubes equipped with septa and stir bars,
which were evacuated and flushed with Ar several times beforehand.
A toluene solution of Ph2Se2 (208 µL, 0.0207 mmol, 0.4 equiv) was
added to each of these tubes followed by an amount of Bu3SnH in
toluene (104 µL, stock solution) corresponding to that of Ph2Se2.
Further stock solution of Bu3SnH in toluene (337 µL, 1.3 equiv) was
then added to each tube followed by 525 µL of toluene to make the
volume 1574 µL. Each tube was then equilibrated at the required
temperature and treated with the stock solution of DBPO in toluene
(26 µL, 0.05 equiv). The total volume in each tube was finally made
up to 1.6 mL. The reactions was stirred at the temperature indicated
Acknowledgment. We thank the NSF (CHE 9625256) for
support of this work.
Supporting Information Available: Details of the prepara-
tion of compounds 7 and 10-12 and Tables 1-4 of kinetic
data (5 pages, print/PDF). See any current masthead page for
ordering information and Web access instructions.
JA980447W