Lewis Acid Activation and Catalysis of Dialkylaminyl Radical Reactions

Article abstract of DOI:10.1021/jo962396w

Lewis acid activation and catalysis of dialkylaminyl radical reactions is demonstrated both qualitatively and quantitatively. Cyclization of the N-butyl-4-pentenaminyl radical (1) in the presence of a wide range of Lewis acids was shown to be efficient with good to excellent yields of cyclic products obtained in reactions conducted even at -78°C. Rate constants for fragmentation of the N-ethyl-2,2-diphenylethylaminyl radical (6), 6-exo cyclization of the N-methyl-6,6-diphenyl-5-hexenaminyl radical (7), and 5-exo cyclization of the N-methyl-5,5-diphenyl-4-pentenaminyl radical (8) in the presence of the Lewis acids LiBF₄, MgBr₂, and BF₃ were measured by laser flash photolysis (LFP) methods. The LFP studies demonstrated saturation kinetic behavior with respect to the Lewis acids. Equilibrium binding constants for the Lewis acids with the dialkylaminyl radicals and rate constants for reactions of the Lewis acid complexed dialkylaminyl radicals were obtained from nonlinear regression analysis of the observed rate constants.

Full text of DOI:10.1021/jo962396w

2704
J . Org. Chem. 1997, 62, 2704-2710
Lew is Acid Activa tion a n d Ca ta lysis of Dia lk yla m in yl Ra d ica l
Rea ction s
Chau Ha, Osama M. Musa, Felix N. Martinez, and Martin Newcomb*
Department of Chemistry, Wayne State University, Detroit, Michigan, 48202
Received December 30, 1996X
Lewis acid activation and catalysis of dialkylaminyl radical reactions is demonstrated both
qualitatively and quantitatively. Cyclization of the N-butyl-4-pentenaminyl radical (1) in the
presence of a wide range of Lewis acids was shown to be efficient with good to excellent yields of
cyclic products obtained in reactions conducted even at -78 °C. Rate constants for fragmentation
of the N-ethyl-2,2-diphenylethylaminyl radical (6), 6-exo cyclization of the N-methyl-6,6-diphenyl-
5-hexenaminyl radical (7), and 5-exo cyclization of the N-methyl-5,5-diphenyl-4-pentenaminyl radical
(8) in the presence of the Lewis acids LiBF4, MgBr2, and BF3 were measured by laser flash photolysis
(LFP) methods. The LFP studies demonstrated saturation kinetic behavior with respect to the
Lewis acids. Equilibrium binding constants for the Lewis acids with the dialkylaminyl radicals
and rate constants for reactions of the Lewis acid complexed dialkylaminyl radicals were obtained
from nonlinear regression analysis of the observed rate constants.
Sch em e 1
Cyclizations of carbon-centered radicals have become
one of the more important synthetic entries to five-
membered carbocycles. Although related cyclizations of
heteroatom-centered radicals to produce heterocycles are
less well developed, nitrogen-centered radicals are now
available from a variety of precursors, and pyrrolidine
formations via nitrogen radical cyclizations are well
documented.1,2 Neutral aminyl radicals, the first formed
species from most nitrogen radical precursors, are rela-
tively unreactive, but it has long been known that
protonation of aminyl radicals gives aminium cation
radicals that are much more reactive. Recent direct
kinetic studies of analogous dialkylaminyl radicals and
dialkylaminium cation radicals show that intramolecular
reactions of the protonated species are several orders of
magnitude faster than those of the neutral counter-
parts.3,4
protonated dialkylaminium cation radicals are available
from the results of this work and previous reports.
Aminyl radicals are also activated by complexation
Resu lts
with Lewis acids2,5 including such mild agents as MgBr2
6
and LiBF4,5 and the use of Lewis acids as activating
agents or catalysts instead of protic acids could offer
considerable advantages in synthetic applications. In
this work, we report evaluations of Lewis acid activation
of a simple 5-exo aminyl radical cyclization and laser
flash photolysis (LFP) kinetic studies of reactions of three
dialkylaminyl radicals in the presence of Lewis acids. The
qualitative studies show that a wide range of Lewis acids
activate the aminyl radical, and the LFP results provide
equilibrium binding constants and rate constants for
reactions of the Lewis acid complexed aminyl radicals
which can serve as basis reactions for kinetic scales of
these species. Direct comparisons of the kinetics of the
Lewis acid-complexed dialkylaminyl radical reactions
with those of uncomplexed dialkylaminyl radicals and
P r od u ct Stu d ies. The 5-exo cyclization of the N-bu-
tyl-4-pentenaminyl radical (1) was studied under a
variety of conditions. Radical 1 was produced in chain
reactions of the PTOC7 carbamate precursor 3. Both 1
and its protonated counterpart, dialkylaminium cation
radical 2, can cyclize in a 5-exo manner to give pyrro-
lidinylmethyl radicals 4. In the absence of other radi-
cophiles, intermediates 4 react in a chain propagation
step with 3 to give, ultimately, N-butyl-2-[(2-pyridylthio)-
methyl]pyrrolidine (5) (Scheme 1).8 Because the cycliza-
tion of neutral radical 1 is relatively slow, the yields of
cyclic products often are low in the absence of an acid.
The neutral acyclic radical 1 apparently does not react
with the PTOC carbamate precursor, but low yields can
result from radical-radical reactions (couplings and
disproportionations), and the relative amounts of cycliza-
tion and radical-radical derived products will be a
function of the rate of radical formation.
X Abstract published in Advance ACS Abstracts, April 15, 1997.
(1) Stella, L. Angew. Chem. Int. Ed. Engl. 1983, 22, 337-350.
(2) Esker, J . L.; Newcomb, M. In Advances in Heterocyclic Chemistry;
Katritzky, A. R., Ed.; Academic: San Diego, 1993; Vol. 58, pp 1-45.
(3) Horner, J . H.; Martinez, F. N.; Musa, O. M.; Newcomb, M.;
Shahin, H. E. J . Am. Chem. Soc. 1995, 117, 11124-11133.
(4) Musa, O. M.; Horner, J . H.; Shahin, H.; Newcomb, M. J . Am.
Chem. Soc. 1996, 118, 3862-3868.
Precursor 3 was allowed to react in the presence of
several Lewis acids at various temperatures. Reactions
(7) The acronym PTOC derives from pyridine-2-thioneoxycarbonyl.
PTOC carbamates as actually anhydrides of the thiohydroxamic acid
and a carbamic acid.
(8) Newcomb, M.; Deeb, T. M. J . Am. Chem. Soc. 1987, 109, 3163-
3165.
(5) Newcomb, M.; Ha, C. Tetrahedron Lett. 1991, 32, 6493-6496.
(6) Dickinson, J . M.; Murphy, J . A. J . Chem. Soc., Chem. Commun.
1990, 434-436.
S0022-3263(96)02396-1 CCC: $14.00 © 1997 American Chemical Society

Catalysis of Dialkylaminyl Radical Reactions
J . Org. Chem., Vol. 62, No. 9, 1997 2705
Ta ble 1. Yield s of P r od u ct 5 fr om Rea ction s of 3 in th e
P r esen ce of BF 3OEt2
Sch em e 2
solvent
THF
temp (°C)
[BF3]
% yield
-78
0
<4
74
98
59
85
80
82
22
65
81
32
14
39
46
68
98
44
69
82
97
0.05
0.10
0.15
0.05
0.10
0.15
0.012
0.025
0.05
0.075
0.01
0.02
0.03
0.04
0.05
0.012
0.025
0.038
0.05
-42
0
25
CH2Cl2
-78
salts MgBr2 and LiBF4 are noteworthy in that they
represent quite mild conditions. In general, the yields
of cyclic product 5 increased as the concentration of Lewis
acid was increased, but high concentrations of Lewis acid
were shown to reduce the yields of 5 in several cases. It
is possible that the Lewis acids can interfere with the
radical chain propagation step involving attack of inter-
mediate cyclic radical 4 on the PTOC precursor by
complexing with 3. Alternatively, Lewis acid complex-
ation by precursor 3 might further activate the already
reactive acyl derivative toward nucleophilic attack that
destroys 3.
Ta ble 2. Yield s of P r od u ct 5 fr om Rea ction s of 3 in th e
P r esen ce of Tita n iu m sa lts in CH2Cl2 a t -78 °C
Lewis acid
concn (M)
% yield
Ti(O-iPr)4
Ti(O-iPr)3Cl
0.025
0.012
0.025
0.038
0.050
0.012
0.025
0.038
0.050
0.012
0.025
0.05
>10
34
31
28
18
60
75
10
11
30
99
81
50
Ti(O-iPr)2Cl2
Kin etic Stu d ies. In order to evaluate the Lewis acid
activation in a quantitative sense, three dialkylaminyl
radical reactions were studied by LFP. Radicals 6-8
were produced by photolysis of the corresponding PTOC
carbamate precursors with 355 nm light from a Nd-YAG
laser. Each of the radicals reacts to give a diphenylalkyl
radical (Scheme 2) that is readily monitored by UV
spectroscopy. The kinetics of reactions were followed by
the same method as that previously used in kinetic
studies of the neutral radicals4 and their protonated
counterparts.3 The kinetics of the reactions of the neutral
dialkylaminyl radicals at 20 °C are as follows: radical 6
fragments too slowly for accurate kinetic measurements
(k < 5 × 103 s-1), radical 7 cyclizes with a rate constant
of ca. 7 × 103 s-1, and radical 8 cyclizes with a rate
Ti(O-iPr)Cl3
TiCl4
0.025
Ta ble 3. Yield s of P r od u ct 5 fr om Rea ction s of 3 in th e
P r esen ce of Lew is Acid s
Lewis Acid
solvent
temp (°C)
concn (M)
% yield
none
MgBr2
THF
THF
-78
-78
<4
69
57
38
57
18
35
12
20
39
27
30
41
0.012
0.025
0.050
0.075
0.050
0.075
0.012
0.03
LiBF4
THF
22
22
22
CH3CN
CH2Cl2
THF
THF
benzene
CH2Cl2
constant of 3.2 × 105 s-1 4
.
AgBF4
Et2AlCl
22
Reactions of radicals 6-8 were studied in the presence
of the Lewis acids BF3, LiBF4, and MgBr2 in THF at 20
°C, and the cyclization of radical 8 in the presence of BF3
was also studied at 4 °C. Accelerations of the fragmenta-
tion of 6 and cyclizations of 7 and 8 were obvious. The
reactions demonstrated saturation kinetics expected for
a rapid prior equilibrium between the uncomplexed
dialkylaminyl radical and the Lewis acid followed by rate-
limiting reaction of the Lewis acid complexed aminyl
radical. Figures 1-3 show the results for fragmentation
of radical 6 and cyclizations of radicals 7 and 8 at 20 °C.
For formation of a dialkylaminyl radical-Lewis acid
complex (eq 4), the equilibrium constant for complex
formation, KC, is given in eq 5. The observed rate
constant for formation of the diphenylalkyl radical prod-
uct is given in eq 6 where kx and Fx are, respectively,
the rate constants for reaction and the mole fractions of
the neutral aminyl radical (N) and the Lewis acid
complex (C). Even at low concentrations of Lewis acids,
the reactions occur essentially only from the complexed
species, and the approximation in eq 7 applies. Substitu-
tion for FC in eq 7 gives eq 8 where kC is the rate constant
-78
5
0.04
9-BBN-Bra
-78
0.025
0.050
0.10
a
9-Bromo-9-borabicyclo[3.3.1]nonane.
were initiated by irradiation with a 150 W tungsten
filament bulb, and the reaction progress was monitored
by TLC. After precursor 3 was depleted, the yields of
product 5 were determined by GC. Tables 1-3 contain
the results, some of which were previously communi-
cated.5
Essentially all of the Lewis acids tested appeared to
activate aminyl radical 1 toward cyclization, and catalytic
behavior was apparent in several cases. Because the
precursor, intermediate radicals, and products (and the
solvent THF, when used) can complex with the Lewis
acids, these results should be viewed qualitatively, but
some general conclusions are possible. The most effica-
cious Lewis acids were BF3 and titanium species of
intermediate Lewis acidity. The activating effects of the

2706 J . Org. Chem., Vol. 62, No. 9, 1997
Ha et al.
and any complexation between the Lewis acid and the
precursors or products cannot affect the total amount of
Lewis acid available. The rate constants for reaction of
the complexed dialkylaminyl radicals and the equilibrium
constants for complexation were solved by nonlinear
regression analysis according to eq 8, and the results are
given in Table 4. The results of the regression analyses
for the reactions of radicals 6-8 are shown in Figures
1-3 as solid lines.
LFP kinetic studies also were conducted at varying
temperatures with radical 7 and with radical 8 in the
presence of BF3. The observed rate constants are listed
in Table 5. Due to the large equilibrium constant for
complex formation with BF3, the populations of com-
plexed aminyl radicals predominated in all cases. For
example, the smallest mole-fraction of complexed aminyl
radical was in the low temperature studies with radical
8; using an equilibrium binding constant of 40 M-1 and
with the BF3 concentration at 0.1 M, 80% of radical 8
would be in the form of the complex at 4 °C.
F igu r e 1. Observed rate constants for fragmentation of
radical 6 in THF at 20 °C in the presence of Lewis acids. The
symbols are the experimental rate constants, and the lines are
the best fits from nonlinear regression analyses according to
eq 8. For LiBF4, the actual concentrations are twice the values
listed on the axis.
Using the weighted average equilibrium constant for
BF3 complexation at 20 °C of 56 M-1 (see Discussion) and
the experimental equilibrium constant for BF3 complex-
ation at 4 °C of 40 M-1, we estimated the equilibrium
constant for complexation at all temperatures studied
and calculated the mole-fraction of complexed species in
each case. The experimental values of kobs were then
divided by the appropriate mole fraction to give the
calculated rate constant for reaction of the complexed
species, kC, at each temperature; these values are also
listed in Table 5. From the calculated values of kC for
reactions of the BF3 complexes of radicals 7 and 8, one
computes the Arrhenius functions for these cyclizations
given in eq 9 (for the BF3 complex of radical 7) and eq 10
(for the BF3 complex of radical 8) where errors are 2σ
and R is in kcal/mol. One might note that the Arrhenius
functions in eqs 9 and 10 are only slightly different than
those one would obtain if the kobs values were used or if
the mole-fraction of complexed species was assumed to
be constant and in the range of 0.9 at all temperatures.
F igu r e 2. Observed rate constants for cyclization of radical
7 in THF at 20 °C in the presence of Lewis acids. See caption
for Figure 1.
log (kC s) ) (8.1 ( 0.3) - (1.7 ( 0.4)/2.3RT (9)
log (kC s) ) (8.7 ( 0.3) - (2.2 ( 0.4)/2.3RT (10)
As a check on the validity of the data treatment, one
can compare the computed rate constants at 20 °C from
eqs 9 and 10 with the rate constants determined at this
temperature from the complete saturation kinetic study
(Table 6 in Discussion). The two pairs of rate constants
differ by 6% for radical 7 and 4% for radical 8.
F igu r e 3. Observed rate constants for cyclization of radical
8 in THF at 20 °C in the presence of Lewis acids. The symbols
are experimental rate constants, and the lines are fits from
eq 8.
for reaction of the dialkylaminyl radical-Lewis acid
complex and KC is the equilibrium constant for complex-
ation.
Discu ssion
Dialkylaminyl radical 1 provides an excellent test
reaction for evaluation of Lewis acid activation. Radical
1 and related N-alkyl-4-pentenaminyl radicals have been
produced from a variety of sources: thermolysis and
photolysis of symmetrical tetrazenes;9,10 radical chain
reactions of N-chloroamines,11 PTOC carbamates8,12 and
R2N• + LA a (R2N•-LA)
(4)
KC ) [R2N•-LA]/([R2N•][LA])
kobs ) kNFN + kCFC
(5)
(6)
(7)
(8)
kobs ) kCFC
(9) Michejda, C. J .; Campbell, D. H.; Sieh, D. H.; Koepke, S. R. In
Organic Free Radicals; Pryor, W. A., Ed.; American Chemical Society:
Washington, DC, 1978; chapter 18.
(10) Newcomb, M.; Burchill, M. T. J . Am. Chem. Soc. 1983, 105,
7759-7760.
(11) Stein, J .-L.; Stella, L.; Surzur, J .-M. Tetrahedron Lett. 1980,
21, 287-288 and references therein.
kobs ) kCKC[LA]/(1 + KC[LA])
In the LFP experiments, the concentrations of the
PTOC carbamate precursors were small (<3 × 10-5 M),

Catalysis of Dialkylaminyl Radical Reactions
J . Org. Chem., Vol. 62, No. 9, 1997 2707
Ta ble 4. Exp er im en ta l Equ ilibr iu m Bin d in g Con sta n ts a n d Ra te Con sta n ts for Rea ction s of Dia lk yla m in yl Ra d ica ls in
th e P r esen ce of Lew is Acid s in THF a t 20 °Ca
6
7
8
acid
LiBF4
KC
10-4 × kC
8.4 ( 0.8
9.3 ( 0.1
56 ( 5
KC
10-4 × kC
15.2 ( 0.4
15.7 ( 0.2
680 ( 16
KC
10-4 × kC
670 ( 50
199 ( 3
1080 ( 70
980 ( 30
4.9 ( 1.4
43 ( 1
53 ( 21
6.8 ( 0.5
44.6 ( 1.5
55 ( 5
4.0 ( 0.5
44 ( 2
61 ( 14
40 ( 3.5
MgBr2
BF3
BF3 (4 °C)b
a
b
All uncertainties are 1σ. Equilibrium constants (KC) are in units of M-1, and rate constants (kC) are in units of s-1
.
Data obtained
at 4 °C.
Ta ble 5. Ra te Con sta n ts for Cycliza tion s of Ra d ica ls 7
a n d 8 in THF in th e P r esen ce of BF 3
will occur at the diffusion-controlled limit with spin
statistical selection of 0.25 (a rate constant of about 5 ×
109 M-1 s-1 in a low viscosity organic solvent at ambient
temperature), the pseudo first-order rate constant for
radical-radical reactions will be about 5 × 104 s-1 if the
radical concentration reaches 1 × 10-5 M. The cyclization
of radical 1 is only marginally fast enough to compete
with radical-radical reactions under typical synthetic
conditions.
Another feature that results in reduced yields of cyclic
product from simple 4-pentenaminyl radicals is the
reversibility of the cyclization reaction. Previously, our
group reported that the cyclization of radical 1 was
reversible.12 Recently, Maxwell and Tsanaktsidis have
claimed that our ring opening kinetic study could not be
reproduced and that the cyclization reaction of 1 was not
reversible.16,20 Subsequently, we found that the results
of Maxwell and Tsanaktsidis were artifacts apparently
due to contamination of their radical precursor samples.19
The equilibrium constant for cyclization of 1 is near
unity,19 and the product distributions from reactions of
1 reflect not only the rate constants for the unimolecular
radical reactions but also those of the second order radical
trapping reactions.
radical
temp (°C)a
[BF3] (M)
0.2
10-6 × kobs
10-6 × kC
b
c
7
0.0
10.0
20.0
30.0
40.0
50.0
0.0
1.5
1.5
8.0
8.0
10.0
13.2
13.2
20.0
30.0
40.0
50.0
5.4
5.8
6.0
7.4
8.3
9.0
8.4
8.4
8.3
9.2
9.2
9.6
9.9
6.1
6.4
6.6
8.0
8.9
9.3
10.7
9.5
9.0
10.3
9.9
11.8
11.0
10.5
11.9
14.3
15.5
17.8
8
0.1
0.2
0.3
0.2
0.3
0.1
0.2
0.3
0.1
0.1
0.1
0.1
9.8
10.0
12.3
13.6
15.9
a
b
( 0.2 °C. Observed rate constant in units of s-1 for reaction
in the presence of BF3. c Calculated rate constant in units of s-1
see text.
;
other thiohydroxamate precursors,13 and benzenesulfen-
amides14,15 and related arenesulfenamides;16 chemical10
or electrochemical17 oxidation of lithium amides; and ring
openings of aziridines.6,18 Surzur showed that protic
acids and Lewis acids resulted in high yields of pyrroli-
dine products from 1,1 and a comparison of the reactivity
of radical 1 and its protonated counterpart, aminium
cation radical 2, in the presence of various hydrogen atom
transfer trapping agents has been reported.12 Rate
constants for cyclization of both aminyl radical 1 and
aminium cation radical 2 are available from previously
reported competitive trapping studies and recent direct
kinetic studies.
In contrast to the low reactivity of neutral dialkylami-
nyl radicals, protonated dialkylaminium cation radicals
such as 2 (Scheme 1) react very efficiently in 5-exo
cyclizations. Again, one can calculate an approximate
rate constant for the unimolecular reaction. A competi-
tion study of cyclization of radical 2 in CH3CN at 25 °C
versus trapping by Bu3SnH gave a ratio of rate constants
of kSnH/kc ) 3 M-1 12
On the basis of the reported rate
.
constant for reaction of Ph3SnH with dialkylaminium
cation radicals at 25 °C in CH3CN (k ) 2.4 × 108 M-1
s-1)3 and with the assumption that Ph3SnH should be
about twice as reactive as Bu3SnH (as it is in reactions
with alkyl radicals21 ), the rate constant for cyclization
of 2 at 25 °C is about 4 × 107 s-1. This value is in good
agreement with a recently reported directly measured
rate constant, that for cyclization of radical 9 in CH3CN
From competition studies of the 5-exo cyclization of
radical 1 versus trapping by Bu3SnH, the rate constant
for cyclization of 1 at 50 °C is 4.3 × 104 s-1 12,19
.
Assuming
a log A value for cyclization of 1 of about 9.5, the rate
constant for cyclization of 1 at 25 °C would be only 1.5 ×
104 s-1. Thus, one can readily rationalize modest yields
of cyclic products formed from simple 4-pentenaminyl
radicals such as 1. Because radical-radical reactions
at 10 °C, which is 1 × 108 s-1 22
.
(12) Newcomb, M.; Deeb, T. M.; Marquardt, D. J . Tetrahedron 1990,
46, 2317-2328.
(13) Newcomb, M.; Weber, K. A. J . Org. Chem. 1991, 56, 1309-
1313.
(14) Bowman, W. R.; Clark, D. N.; Marmon, R. J . Tetrahedron 1994,
50, 1275-1294 and references therein.
The results in Tables 1-3 indicate that all of the Lewis
acids tested resulted in acceleration of the 5-exo cycliza-
(15) Beckwith, A. L. J .; Maxwell, B. J .; Tsanaktsidis, J . Aust. J .
Chem. 1991, 44, 1809-1812.
(16) Maxwell, B. J .; Tsanaktsidis, J . J . Am. Chem. Soc. 1996, 118,
4276-4283.
(17) Tokuda, M.; Miyamoto, T.; Fujita, H.; Suginome, H. Tetrahedron
1991, 47, 747-756, and references cited therein.
(18) Dickinson, J . M.; Murphy, J . A. Tetrahedron 1992, 48, 1317-
1326.
(20) Maxwell, B. J .; Tsanaktsidis, J . J . Chem. Soc., Chem. Commun.
1994, 533-534.
(21) Newcomb, M. Tetrahedron 1993, 49, 1151-1176.
(22) Newcomb, M.; Tanaka, N.; Bouvier, A.; Tronche, C.; Horner, J .
H.; Musa, O. M.; Martinez, F. N. J . Am. Chem. Soc. 1996, 118, 8505-
8506.
(19) Newcomb, M.; Musa, O. M.; Martinez, F. N.; Horner, J . H. J .
Am. Chem. Soc., in press.

2708 J . Org. Chem., Vol. 62, No. 9, 1997
Ha et al.
Ta ble 6. Ra te Con sta n ts for Rea ction s of Dia lk yla m in yl
tion reaction of 1. The excellent yields of cyclic product
5 obtained with BF3 and various titanium species and
the good yields of cyclic product found in the limited
studies with other Lewis acids suggest that one can select
from a wide range of aminyl radical activating agents.
Nevertheless, we note that high Lewis acid concentra-
tions were demonstrated to be deleterious to the overall
yields of cyclic products in several cases.
We note that Lewis acid activation of dialkylaminyl
radical cyclizations by BF3 has been shown to be compat-
ible with some other radical functionalization protocols.
For example, when PTOC carbamate 3 was allowed to
react in the presence of BF3 and diphenyl diselenide, the
phenylselenomethyl product 10 was obtained in 80%
yield.5 Similarly, the dialkylaminyl radical 11, produced
from its corresponding PTOC carbamate precursor, re-
acted in a tandem cyclization sequence to give, ulti-
mately, pyrrolizidine 12 as a mixture of diastereomers
in 70% yield.5
Ra d ica ls in THF a t 20 °C in th e P r esen ce of Acid sa
acid
noneb
10-4 × kC for 6 10-4 × kC for 7 10-4 × kC for 8
<0.5
8.1
9.3
56
0.7
16.5
15.9
677
3800
32
560
200
LiBF4
MgBr2
BF3
1120
CF3CO2Hc
2000
1000000d
Rate constants in units of s-1 calculated from eq 8 using
a
b
weighted average values for KC as constants. Values from ref 4.
c Values from ref 3; note that the dialkylaminium cation radical
d
reactions are highly sensitive to solvent identity. Rate-limiting
protonation occurs; estimated value.
constants for KC in eq 8 gave the rate constants listed in
Table 6. For comparison purposes, we have included in
Table 6 the rate constants for reactions of the uncom-
plexed dialkylaminyl radicals4 and the protonated di-
alkylaminium cation radicals.3
As in the case of the product studies with radical 1,
Lewis acid acceleration for reactions of 6-8 occurred for
all three Lewis acids studied by LFP. The extent of
activation by the Lewis acids correlated roughly for the
three dialkylaminyl radical reactions, and the kinetic
accelerations observed by LFP are consistent with the
qualitative observations with radical 1. BF3 resulted in
the greatest acceleration, and LiBF4 and MgBr2 gave
similar accelerations.
The consistency in the values for the equilibrium
constants for complexation by each Lewis acid with the
three dialkylaminyl radicals 6-8 clearly suggests that
these binding constants will also apply for other dialkyl-
aminyl radicals. Another question to address is whether
or not the catalytic effects of the Lewis acids observed
for reactions of radicals 6-8 can be expected to be similar
for other dialkylaminyl radical reactions. Some experi-
mental evidence suggests that this will be the case for a
given reaction type. For example, the qualitative results
of the 5-exo cyclizations of radical 1 generally correspond
to the quantitative results for the 5-exo cyclization of
radical 8 in that BF3 appeared to be more efficacious than
MgBr2 or LiBF4 on the basis of the yields of 5 for reactions
of 1 conducted under similar conditions. In a more
quantitative evaluation, the rate constant for 5-exo
cyclization of the neutral dialkylaminyl radical 13 at 20
°C is accelerated by about two orders of magnitude upon
BF3 complexation to give radical 14,19 an effect similar
to the acceleration by BF3 observed with 8.
The LFP kinetic studies permit a quantitative evalu-
ation of Lewis acid accelerations of dialkylaminyl radical
reactions. The simultaneous solution of saturation ki-
netic data provides both the equilibrium (or pseudoequi-
librium) binding constant for the first step in the reaction
sequence and the rate constant for the second step, and
these values are correlated. The three dialkylaminyl
radicals 6-8 are similarly substituted about nitrogen,
and the three apparent equilibrium constants for each
particular Lewis acid were similar to one another.
Therefore, we calculated the weighted average equilib-
rium constants for dialkylaminyl radical complexation by
each Lewis acid at 20 °C and then used these values as
constants in the regression analyses of kobs to recalculate
the rate constants of the reactions. The important point
for such data treatment is that rapid equilibria for Lewis
acid complexation must be established in all cases or else
the apparent equilibrium constant from the initial re-
gression analyses (listed in Table 4) will actually be the
pseudoequilibrium constant described by eq 11 where kf
is the rate constant for formation of the complex, kd is
the rate constant for dissociation of the complex, and kC
is the rate constant for the radical reaction of the
complex. If kC was competitive with kd for the fastest
radical reaction, cyclization of radical 8, then the appar-
ent equilibrium constants for complexation of this radical
resulting from the initial regression analyses would be
smaller than those for radicals 6 and 7; inspection of the
data in Table 4 shows that they are not.
One point that deserves note is the relatively small
kinetic accelerations observed in the 5-exo cyclizations
of radical 8 in comparison to the 6-exo cyclizations of
radical 7. The neutral radical 8 cyclizes about 1.5 orders
of magnitude faster than radical 7, and, for the proto-
nated analogs of 7 and 8, the 5-exo cyclization is about
2.5 orders of magnitude faster than the 6-exo cyclization.
Such differences in rate constants of 5-exo and 6-exo
cyclizations of analogous radicals are typically found, and
we have observed in several LFP studies of 5-exo and
6-exo cyclizations of terminal diphenylethene-containing
radicals which are isostructural with 7 and 8 that the
5-exo cyclizations are about two orders of magnitude
Kapp ) kf/(kd + kC)
(11)
The weighted average equilibrium constants for com-
plexation in THF at 20 °C were as follows: (56 ( 5) M-1
for BF3, (43.4 ( 0.9) M-1 for MgBr2, and (5.3 ( 0.3) M-1
for LiBF4. Regression analyses using these values as

Catalysis of Dialkylaminyl Radical Reactions
J . Org. Chem., Vol. 62, No. 9, 1997 2709
faster than the corresponding 6-exo cyclizations.3,4,23,24
One might rationalize that the cyclization kinetics of the
LiBF4 complexes of 7 and 8 are consistent with the
kinetic behavior of other matched pairs, but the kinetics
of the MgBr2 and BF3 complexes clearly are not. That
the 5-exo cyclization of the BF3 complex of 8 is only twice
as fast as the 6-exo cyclization of the BF3 complex of 7 is
remarkable.
The unusual kinetic behavior of the BF3 complex of 8
is apparent in the Arrhenius functions for cyclization of
the BF3 complexes of 7 and 8 (eqs 9, 10). Generally, the
log A terms, the entropic terms, for 5-exo and 6-exo
cyclizations of a matched pair of radicals are similar, and
the Ea value for the 6-exo cyclization is greater than that
for the 5-exo cyclization. However, the 5-exo cyclization
of the BF3 complex of 8 has a greater activation energy
than the 6-exo cyclization of the complex of 7. This
suggests that the Lewis acid imposes a considerable steric
barrier in the transition state of the 5-exo cyclization that
is not present in the transition state of the 6-exo cycliza-
tion.
Elucidation of the origins of a barrier imposed by the
Lewis acid BF3 in the 5-exo cyclization reaction of 8 might
require extensive computational work, but it is important
to note that a kinetic barrier for 5-exo cyclization cannot
be simply a function of a tertiary radical center per se.
For example, the rate constants for 5-exo cyclizations of
tertiary alkyl radicals 15 are similar to those of the
corresponding secondary alkyl radicals.21 On the other
hand, the tertiary ethoxycarbonyl-substituted radicals 16
cyclize considerably less rapidly than the isostructural
tertiary alkyl radicals 15 and the analogous secondary
ethoxycarbonyl-substituted radicals.24,25 We ascribed the
slow reactions of tertiary radicals 16 to the planarity of
the radical center which results in a steric interaction
in the cyclization that is not present in a pyramidalized
tertiary alkyl radical.24
tive.3 Therefore, one should expect that the kinetics of
reactions of Lewis acid complexes of aminyl radicals will
also demonstrate considerable solvent sensitivity.
Finally, we comment on the utility of the LFP kinetic
method for studies of Lewis acid effects in radical
reactions, a topic of considerable current research inter-
est. For example, focusing only on recent reports, Lewis
acids have been employed to effect highly diastereo-
selective26-35 and even enantioselective36-39 radical reac-
tions. In some cases, the Lewis acid might simply serve
to control the conformation of a radical or molecule in
the reaction and has little effect on the kinetics, but there
is good evidence that Lewis acid catalysis is involved in
selected cases.26-30,37,38 The LFP method we employed
in this work should be generally applicable for measuring
both the binding constants and catalytic rate constants
for the Lewis acids in these types of reactions because
absolute kinetics are measured directly. An alternative
approach, involving product yields from competing reac-
tions, would be considerably more difficult to execute and
less informative because the Lewis acid could affect the
kinetics of both of the competing reactions.
Con clu sion
Lewis acid catalysis of dialkylaminyl radical reactions
is demonstrated both qualitatively and quantitatively by
the studies reported here. The LFP kinetic method
provides both equilibrium binding constants and catalytic
rate constants for complexes of the aminyl radicals with
Lewis acids. The degree of kinetic activation by the
Lewis acids is much less than that observed upon
protonation of dialkylaminyl radicals, but the accelera-
tions in most cases are adequate for successful synthetic
applications involving simple 4-pentenaminyl radical
cyclizations. For example, the seemingly poor 6.2-fold
acceleration observed in the 5-exo cyclization of radical
8 upon complexation with MgBr2 suggests that the MgBr2
complex of the simple 4-pentenaminyl radical 1 cyclizes
with a rate constant of about 1 × 105 s-1 at ambient
temperature; this rate constant is similar to that for the
archetypal radical cyclization, 5-exo cyclization of the
5-hexenyl radical, one of the most important radical
reactions from the perspective of organic synthesis.
Exp er im en ta l Section
All Lewis acids with the exception of MgBr2 were purchased
from Aldrich Chemical Co. and used as received. Solutions
Another point that deserves comment involves the
acceleration observed with LiBF4. Anodic oxidation of
lithium dialkylamides containing the 4-pentenaminyl
moiety has been reported to give good to excellent yields
of pyrrolidine products17 in contrast to the modest yields
typically observed when similar dialkylaminyl radicals
are produced in chain reactions under neutral conditions.
In the electrochemical reactions, which were run in THF,
the supporting electrolyte was lithium perchlorate, and
the lithium cation almost certainly served to catalyze the
aminyl radical cyclizations.
One should note that the kinetic measurements re-
ported in this work were performed in one solvent, THF.
Whereas the kinetics of most radical reactions are
relatively insensitive to solvent, those for reactions of
dialkylaminium cation radicals are highly solvent sensi-
(26) Guindon, Y.; Guerin, B.; Rancourt, J .; Chabot, C.; Mackintosh,
N.; Ogilvie, W. W. Pure Appl. Chem. 1996, 68, 89-96.
(27) Guindon, Y.; Guerin, B.; Chabot, C.; Ogilvie, W. J . Am. Chem.
Soc. 1996, 118, 12528-12535.
(28) Nagano, H.; Kuno, Y.; Omori, Y.; Iguchi, M. J . Chem. Soc.,
Perkin Trans. 1 1996, 389-394.
(29) Nishida, M.; Nishida, A.; Kawahara, N. J . Org. Chem. 1996,
61, 3574-3575.
(30) Asao, N.; Liu, J . X.; Sudoh, T.; Yamamoto, Y. J . Org. Chem.
1996, 61, 4568-4571.
(31) Gerster, M.; Audergon, L.; Moufid, N.; Renaud, P. Tetrahedron
Lett. 1996, 37, 6335-6338.
(32) Sibi, M. P.; J asperse, C. P.; J i, J . G. J . Am. Chem. Soc. 1995,
117, 10779-10780.
(33) Sibi, M. P.; J i, J . G. J . Org. Chem. 1996, 61, 6090-6091.
(34) Nagano, H.; Azuma, Y. Chem. Lett. 1996, 845-846.
(35) Toru, T.; Watanabe, Y.; Tsusaka, M.; Ueno, Y. J . Am. Chem.
Soc. 1993, 115, 10464-10465.
(36) Murakata, M.; Tsutsui, H.; Hoshino, O. J . Chem. Soc., Chem.
Commun. 1995, 481-482.
(23) J ohnson, C. C.; Horner, J . H.; Tronche, C.; Newcomb, M. J . Am.
Chem. Soc. 1995, 117, 1684-1687.
(24) Newcomb, M.; Horner, J . H.; Filipkowski, M. A.; Ha, C.; Park,
S. U. J . Am. Chem. Soc. 1995, 117, 3674-3684.
(25) Newcomb, M.; Filipkowski, M. A.; J ohnson, C. C. Tetrahedron
Lett. 1995, 36, 3643-3646.
(37) Urabe, H.; Yamashita, K.; Suzuki, K.; Kobayashi, K.; Sato, F.
J . Org. Chem. 1995, 60, 3576-3577. Addition: Ibid. 1995, 60, 6641.
(38) Sibi, M. P.; J i, J . G.; Wu, J . H.; Gurtler, S.; Porter, N. A. J .
Am. Chem. Soc. 1996, 118, 9200-9201.
(39) Nishida, M.; Hayashi, H.; Nishida, A.; Kamahara, N. Chem.
Commun. 1996, 579-580.

2710 J . Org. Chem., Vol. 62, No. 9, 1997
Ha et al.
of MgBr2 were prepared by allowing dibromoethane to react
with magnesium metal. The mixed titanium salts were
prepared from appropriate amounts of Ti(O-i-Pr)4 and TiCl4.
The PTOC carbamate 3 and the PTOC carbamate precursors
to aminyl radicals 6-8 were prepared as previously re-
ported.3,12
P r od u ct Stu d ies. Solutions of PTOC carbamate 3 (0.05
M) and the appropriate Lewis acids were prepared under
nitrogen in flasks shielded from light and equilibrated at the
desired reaction temperature. Reactions were initiated by
irradiation with a 150 W tungsten filament bulb placed ca.
0.6 m from the reaction vessels. Reaction progress was
monitored by TLC. Upon consumption of 3, the reaction
mixtures were extracted with water, dried (MgSO4), and
analyzed by GC employing a predetermined response factor
for the known product 5.12
appropriate PTOC carbamate3 in THF in the presence of a
Lewis acid were sparged with He and thermally equilibrated
in a jacketed addition funnel. The samples were allowed to
flow through a flow cell placed in a thermostated well in an
Applied Photophysics LK50 kinetic spectrometer. Tempera-
tures were measured with a thermocouple placed in the
flowing stream approximately 1 cm above the irradiation zone.
The samples were irradiated with a 7 ns pulse of 355 nm light
from a Nd-YAG laser, and the absorbance at ca. 330 nm was
monitored. For studies conducted with various concentrations
of Lewis acids at 20 °C and 4 °C, the observed rate constants
as a function of Lewis acid concentration were fit to eq 8 by
nonlinear regression analysis to give the equilibrium binding
constants and rate constants in Table 4.
Ack n ow led gm en t. We are grateful to the National
Science Foundation for financial support.
LF P Kin etic Stu d ies. The method employed was the same
as that previously used in kinetic studies of dialkylaminyl
radicals and dialkylaminium cation radicals.3,4 Samples of the
J O962396W