Stereocontrol at the steady state in radical cyclizations of acyclic dihalides

Article abstract of DOI:10.1021/jo011056u

The first examples of manipulating stereocontrol solely by reaction topography in radical cyclizations starting from acyclic precursors are reported. The kinetic model for acyclic compound stereoselection is verified experimentally by conducting a series of radical cyclizations of 1,3-dihalo-2-(1-phenyl-3-butynyl)propanes with triphenyltin hydride and measuring the ratios of the products. Monohalide intermediates are observed for the first time, and evidence that bromide- and iodide-substituted radicals have different cyclization rate constants is provided.

Full text of DOI:10.1021/jo011056u

2982
J . Org. Chem. 2002, 67, 2982-2988
Ster eocon tr ol a t th e Stea d y Sta te in Ra d ica l Cycliza tion s of
Acyclic Dih a lid es†
Krzysztof Stalinski and Dennis P. Curran*
Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260
Received November 6, 2001
The first examples of manipulating stereocontrol solely by reaction topography in radical cyclizations
starting from acyclic precursors are reported. The kinetic model for acyclic compound stereoselection
is verified experimentally by conducting a series of radical cyclizations of 1,3-dihalo-2-(1-phenyl-
3-butynyl)propanes with triphenyltin hydride and measuring the ratios of the products. Monohalide
intermediates are observed for the first time, and evidence that bromide- and iodide-substituted
radicals have different cyclization rate constants is provided.
1. In tr od u ction
The study of stereoselection in radical reactions has
been a major focus of the synthetic radical community
over the past decade or so.1 Like their ionic and pericyclic
counterparts, stereoselective radical transformations of-
ten involve reactions of stereoheterotopic2 faces of sp2-
hybridized radicals or radical acceptors. Stereoselection
can be dictated by nearby stereocenters (substrate con-
trol), by chiral auxiliaries, by chiral additives, and even
by chiral catalysts.1 A few examples of traditional ste-
reotopic group selective radical cyclizations are also
known.3 In the example shown in Figure 13a (upper part),
two alkenes compete directly against each other (through
diastereomeric transition states) for a single radical. As
in many radical cyclizations, the major product of this
reaction is readily predicted by the Beckwith-Houk
model.1e,4
Beyond offering new options for traditional types of
face and group selective reactions, the transiency of
radicals also offers unique opportunities for stereocontrol
in nonequilibrium situations. For example, Rychnovsky
has shown that reactions of tetrahydropyranyl radicals
can be faster than ring flip,5 and we have shown that
aryl radical cyclizations of acrylanilides can be faster
than rotation of the N-aryl bond.6 Giese has found that
closures of diradicals can be faster than standard σ bond
F igu r e 1. Traditional and steady-state stereoselection in
radical cyclizations to make bicyclic rings.
rotations.7 In each of these reactions, the short lifetimes
and high reactivity of radicals avoid racemizations or
conformational changes that many other types of inter-
mediates would undergo.
† Dedicated to the memory of Dr. Emmanuil Troyansky.
(1) (a) Sibi, M. P.; Porter, N. A. Acc. Chem. Res. 1999, 32, 163-171.
(b) Guindon, Y.; Gue´rin, B.; Rancourt, J .; Chabot, C.; Mackintosh, N.;
Ogilvie, W. W. Pure Appl. Chem. 1996, 68 (1), 89-96. (c) Renaud, P.;
Gerster, M. Angew. Chem., Int. Ed. 1998, 37, 2563-2579. (d) Curran,
D. P.; Giese, B.; Porter, N. A. Acc. Chem. Res. 1991, 24, 296-304. (e)
Curran, D. P.; Giese, B.; Porter, N. A. Stereochemistry of Radical
Reactions: Concepts, Guidelines, and Synthetic Applications; VCH:
Weinheim, Germany, 1996.
(2) Eliel, L.; Wilen, S. Stereochemistry of Organic Compounds; Wiley-
Interscience: New York, 1994.
(3) (a) Curran, D. P.; Qi, H. Y.; Demello, N. C.; Lin, C. H. J . Am.
Chem. Soc. 1994, 116, 8430. (b) Qi, H. Ph.D. Thesis, University of
Pittsburgh, Pennsylvania, 1994. (c) Villar, F.; Equey, O.; Renaud, P.
Org. Lett. 2000, 2, 1061.
Recently, we have put forth a new type of stereoselec-
tive process founded on the transiency of radicals and
called “stereoselection at the steady state”.4,8 As il-
lustrated by Figure 1b (lower part), existing examples of
this process include competition of two radical precursors
for a single radical acceptor in a multistep process. We
have advanced the notion that stereoselection at the
steady state is fundamentally different from existing
multistep stereoselective processes, which are “compos-
ites” of individual steps.8a Stereoselection at the steady
state is not a composite process, and unlike other existing
stereoselective processes, stereoselection can be accom-
plished without ever pitting stereoisomeric transition
(4) (a) Spellmeyer, D. C.; Houk, K. N. J . Org. Chem. 1987, 52, 959.
(b) Beckwith, A. L. J .; Lawrence, T.; Serelis, A. K. J . Chem. Soc., Chem.
Commun. 1980, 482. (c) Beckwith, A. L. J .; Easton, C. J .; Serelis, A.
K. J . Chem. Soc., Chem. Commun. 1980, 484.
(5) (a) Buckmelter, A. J .; Kim, A. I.; Rychnovsky, S. D. J . Am. Chem.
Soc. 2000, 122, 386. (b) Buckmelter, A. J .; Powers, J . P.; Rychnovsky,
S. D. J . Am. Chem. Soc. 1998, 120, 5589.
(6) Curran, D. P.; Liu, W. D.; Chen, C. H.-T. J . Am. Chem. Soc. 1999,
121, 11012.
(7) Giese, B.; Wettstein, P.; Stahelin, C.; Barbosa, F.; Neuburger,
M.; Zehnder, M.; Wessig, P. Angew. Chem., Int. Ed. 1999, 38, 2586.
(8) (a) Curran, D. P.; DeMello, N. C. J . Am. Chem. Soc. 1998, 120,
329-341. (b) Curran, D. P.; DeMello, N. C.; J unggebauer, J .; Lin, C.-
H. J . Am. Chem. Soc. 1998, 120, 342-351.
10.1021/jo011056u CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/28/2002

Radical Cyclizations of Acyclic Dihalides
J . Org. Chem., Vol. 67, No. 9, 2002 2983
Sch em e 1. Ster eoselection a t th e Stea d y Sta tea
a
Bold arrows represent faster reactions. Dashed arrows represent slower reactions. Standard arros represent nonselective reactions.
2 suggest a substituent effect on the initial radical
cyclization.
2. Resu lts a n d Discu ssion
Dihalides 1 and 2 were selected for this study on the
basis of a number of criteria. First and foremost, on the
basis of the Beckwith-Houk model,1e,4 the two diaster-
eomeric radicals 3a and 3b shown in Figure 2 were
expected to cyclize at different rates. Stereoselection at
the steady state is a kinetic resolution that is predicated
on this rate difference. In addition to providing the bias
needed for stereoselection, the phenyl group adds mo-
lecular weight so that products trans/cis-4 are not volatile
and provides for a ready assignment of the relative
configuration of the products due to anisotropic shielding
of the methyl protons of the cis isomer by the phenyl
group.
The proposed kinetic framework for stereoselection at
the steady state in the radical cyclizations of precursors
1 and 2 is shown in Scheme 1.10 The abstraction of the
first halide from 1 or 2 produces two diastereomeric
radicals, 3a and 3b. This step irreversibly partitions
these and all subsequent intermediates into two different
branches, hereafter called A (upper paths) and B (lower
paths). Nevertheless, the radicals are not irreversibly
committed to any product since the two branches con-
verge on the final cyclic (4t/4c) and reduced (7) products.
The first key step that helps to determine the final ratio
of stereoisomers is reduction of radicals 3a ,b to monore-
duced products 5a ,b, which occurs in competition with
cyclization to 8t and 8c. Stereoselection at the steady
F igu r e 2. Designing substrates after the Beckwith-Houk
model.
states in direct competition with each other. Instead,
stereoselection is achieved when stereoisomeric interme-
diates at the steady state partition selectively between
two different chemical (not stereochemical) pathways9 to
the same stereoisomeric product.
All existing examples of stereoselection at the steady
state are diastereoselective and involve bicyclic sys-
tems.3,8 These systems have provided a rigidity to reduce
options for intermediate radicals, but there is no reason
bicyclic systems are needed. The goal of this work was
to identify and study an acyclic radical precursor that
would be subject to stereoselection at the steady state.
We report herein the design, synthesis, and detailed
study of the cyclizations of radicals derived from 1 and 2
(Figure 2). Results consistent with stereoselection at the
steady state are observed, and postulated monohalide
intermediates are identified for the first time. Differences
in the results between the dibromide 1 and the dihalide
(10) Bold and dotted arrows represent fast and slow stereoselective
processes, respectively. Standard arrows represent nonselective reac-
tions. The major and minor convergences are shown in the bold and
dotted boxes.
(9) Kagan, H. Tetrahedron 2001, 47, 2449.

2984 J . Org. Chem., Vol. 67, No. 9, 2002
Stalinski and Curran
Sch em e 2a
state capitalizes on different partitioning of 3a and 3b
between these two competing chemical transformations
(reduction and cyclization). Under suitable conditions, the
majority of both initial radicals 3a and 3b can be
partitioned into the major convergence toward 4t.
Even though reduction of intermediate radicals 3a ,b
with tin hydride is not a selective process, the overall
stereoselection depends crucially on the tin hydride
concentration. At low tin hydride concentrations, reduc-
tion of radicals 3a ,b does not compete with cyclization
and the ratio of 4t to 4c depends only on the ratio of the
rate constants for halogen abstraction to form 3b (which
partitions exclusively through the major convergence to
4t) and 3a (which partitions exclusively through the
minor convergence to 4c).11 As the tin hydride concentra-
tion increases, reduction of 3a to 5a starts to compete
with cyclization and radicals 3a set to enter the minor
convergence are derailed from the path to 4c and sent
into the major convergence on the path to 4t. A similar
partitioning occurs for 3b. However, at the steady state,
the concentration of the slower cyclizing radical 3a is
higher than that of its faster cyclizing diastereomer 3b
so more of radical 3a is rerouted to the major convergence
than 3b to the minor convergence. This results in
stereoselection by enhancement of the yield of the major
product 4t. In contrast, multistep composite processes
generally provide increased stereoselection by decreasing
the yield of the minor product. A similar partitioning
between reduction and cyclization occurs at radicals 6a
and 6b; however, now the radicals are irrevocably com-
mitted toward one or the other cyclized product, and
enhanced stereoselection here arises because the slower
cyclizing radical 6b is derailed to the doubly reduced
product 7 more efficiently than the faster cyclizing radical
6a .
a
Reagents and conditions: (a) Al, HgCl2, THF, ∼30 °C for 2
days, 50%; (b) MsCl/Et3N, CH2Cl2, 0 °C for 10 min, ∼100%; (c)
NaH/CH2(CO2Et)2, THF, 70 °C for 2 days, 75%; (d) LAH, Et2O, 0
°C for 4 h, 60%; (e) MsCl/Et3N, CH2Cl2, 0 °C for 2 h, ∼100%; (f)
LiBr or NaI, acetone, reflux for 12 h, 95-99%; (g) LAH, Et2O, 0
°C for 2 h, 74%.
Ta ble 1. P r elim in a r y Cycliza tion s of Dibr om id e 1 w ith
P h 3Sn H
[Ph3SnH]
(M)
4t
4c
7
(%)
4t + 4c
entry
(%)
(%)
(%)
4t/4c
1a
1
2
3
4
0.05
0.05
0.10
0.20
0.25
0.40
57
53
44
48
41
32
36
37
32
20
22
27
0
0
0
0
0
0
93
90
76
68
63
59
1.6
1.4
1.4
2.4
1.9
1.2
5
a
With Bu3SnH.
Preparative cyclization of 1 with triphenyltin hydride
(5 equiv, 0.05 M) provided a mixture of 4t and 4c in over
74% isolated yield. These isomeric products could not be
separated by chromatography to give the individual pure
isomers, so they were characterized as a mixture. The
assignment of the relative configuration of these products
After radical generation, there is no point in the
process where two stereoisomeric transition states com-
pete directly with each other to influence the product
ratio. Instead, the stereoselection arises from the reaction
topography, which offers multiple routes to the same
products. Despite the absence of a direct competition,
differences in the rate constants for cyclizations are
crucial since these establish the concentration gradient
of isomeric radicals 3a ,b at the steady state; higher rate
constant ratios (kfast1/kslow1) enforce higher gradients, and
that allows higher yields of the major stereoisomer 4t.
The addition of the tin hydride harvests the concentration
gradient by forcing radicals 3a present in higher con-
centrations down a reduction pathway while radicals 3b
present in lower concentration can still cyclize.
To experimentally verify the expected behavior for this
type of topological group selective process, compounds 1
and 2 were designed as models having two radical
precursors and one radical acceptor. The compounds are
readily available (Scheme 2) by mesylation of 9 followed
by standard malonate alkylation to give 10.12 Reduction
and mesylation gave 11, which was readily converted to
the bromide 1 or iodide 2. The synthesis can be carried
out on a multigram scale. Reduction of the dimesylate
11 with LAH provided a standard sample of the doubly
reduced product 7 in high yield.
1
was made by H NMR spectroscopy on the basis of the
chemical shifts of a methyl group; the resonance from
the cis isomer 4c resonates at higher field than that of
the trans isomer 4t due to anisotropic shielding of the
phenyl ring.
A preliminary series of cyclizations with bromide 1 was
first conducted with 10 equiv of triphenyltin hydride at
80 °C. All reactions were performed in benzene with a
catalytic amount of AIBN to ensure initiation. Biphenyl
was added as an internal standard for GC/MS analyses,
and each of the products was calibrated against this
standard. The results of this series of experiments are
shown in Table 1.
The expected cyclized products 4t/4c were produced in
up to 93% yield. Disconcertingly, we did not detect any
of the doubly reduced product 7 in any of the experiments
by comparison of retention times with that of the
authentic sample. Also, the ratio of 4t/4c varied in an
unexpected manner as the concentration of Ph3SnH was
changed. Initially, the trans/cis ratio increased as the
concentration increased up to 0.2 M Ph3SnH, but then it
began to decrease on further concentration increase.
These results are inconsistent with the model in Scheme
1, which predicts that the yield of 7 should increase as
the yield of 4 decreases and that the 4t/4c ratio should
(11) However, in practice this ratio might not be exactly 50/50
because one of the diastereomeric radicals should cyclize faster than
the second one.
(12) Katzenellenbogen, J . A.; Sofia, M. J . J . Med. Chem. 1986, 29,
230-238.

Radical Cyclizations of Acyclic Dihalides
J . Org. Chem., Vol. 67, No. 9, 2002 2985
Ta ble 2. Red u ction of Dibr om id e 1 w ith P h 3Sn H a t
Differ en t Tem p er a tu r es
continue to increase even as the total yield of these two
products falls (because the yield of 4c falls faster than
that of 4t).
The absence of the doubly reduced product 7 suggested
that hydrostannylation of its triple bond might have
taken place.13 To verify this, we treated an authentic
sample of 7 with excess (10 equiv) Ph3SnH in benzene
at 80 °C. The reaction was monitored by GC/MS. After
36 h, 7 was consumed, and the reaction was worked up.
The crude product was chromatographed to provide a
single compound (60%) identified as alkene 13 (eq 1). This
presumably results from protodestannylation of the hy-
drostannylated product 12, so hydrostannylation does
indeed explain the absence of 7.
t
4t
(%)
4c
(%)
7
(%)
4t + 4c
entry
(°C)
(%)
4t/4c
1
2
3
4
80
70
65
60
48
63
70
59
20
22
24
31
0
3
5
2
68
85
94
90
2.4
2.9
2.9
1.9
Ta ble 3. Red u ction s of Dibr om id e 1 w ith P h 3Sn H
a t 65 °C
[Ph3SnH]
(M)
4t
4c
4t + 4c
(%)
7
(%)
total
(%)
entry
4t/4c
(%) (%)
1
2
3
4
5
6
7
8
9
a
1.1
1.4
1.6
1.8
2.1
2.4
2.8
3.2
3.7
3.9
4.2
4.5
51
57
60
62
65
67
70
67
59
49
40
18
48
41
37
35
31
27
24
21
16
13
10
4
99
98
97
97
96
94
94
88
75
61
50
22
0
1
2
2
3
4
5
10
14
11
8
99
99
99
99
99
98
99
98
89
72
58
28
0.025
0.05
0.075
0.10
0.15
0.20
0.30
0.40
0.50
0.60
1.00
That the change in the 4t/4c ratio did not follow the
kinetic model also suggests that a side reaction, possibly
again hydrostannylation, consumes 4t faster than 4c. To
identify conditions that minimized side reactions, four
experiments were conducted at different temperatures
with the same concentration of the tin hydride (0.2 M).
The results of these experiments are summarized in
Table 2. The optimum temperature was 65 °C. At that
temperature, the combined yield of the products, includ-
ing the doubly reduced 7, was almost quantitative (99%).
Having identified conditions that minimized secondary
product formation, we next performed a series of experi-
ments at increasing Ph3SnH concentrations following a
standard protocol. A benzene solution containing 1,
Ph3SnH (10 equiv), and a catalytic amount of AIBN was
heated at 65 °C until the reaction was complete, and the
products were analyzed by GC. Reaction times varied
from 30 min for the highest concentration experiment to
>8 h for the lowest. The results of these experiments are
summarized in Table 3, entries 2-12. A syringe pump
addition experiment was also conducted (entry 1), and
this presumably has the lowest tin hydride concentration
even though the nature of the experiment precludes an
accurate measure of that concentration.
In the syringe pump addition experiment (entry 1), the
ratio 4t/4c is close to 1/1. This confirms that the selectiv-
ity in the initial abstraction of bromine by the tin radical
is negligible. As the tin hydride concentration increases,
the yield of 4t begins to increase and the yield of 4c
declines. After reaching a maximum in the vicinity of 70%
(entries 6 and 7), the yield of 4t/4c begins to decline, but
the yield of 4c declines more steeply, so the 4t/4c ratio
increases as the combined yield of 4 decreases. Concomi-
tantly, the doubly reduced product 7 begins to grow in
(entries 2-9). However, the yields of this product only
increase to 14% (entry 9), and then slowly decrease
(entries 10-12). At very high tin hydride concentrations
the total yield of 4 and 7 begins to fall from quantitative
as well. Apparently, hydrostannylation of the products
cannot be avoided at high concentrations even at 65 °C.
10
11
12
6
a
Syringe pump addition.
Interestingly, careful monitoring of the radical cycliza-
tions of 1 by GC/MS revealed the presence of several
intermediates with retention times between those of the
dibromide 1 and the collection of hydrocarbons 4t, 4c,
and 7 (see the experimental details in the Supporting
Information) which grew in and then disappeared during
the course of the reaction. Two overlapping, shorter-
retained peaks were assigned as acyclic monobromides
(5a ,b, X ) Br, Scheme 1) by comparison with authentic
samples (see below). The remaining two longer-retained
peaks were tentatively assigned as cyclic monobromides
(8t,c, X ) Br, Scheme 1) on the basis of GC/MS data.
The ratio 8t/8c at intermediate states of the experiments
was very similar to the final 4t/4c ratio, providing further
support for this assignment.
This discovery gave us an opportunity to confirm a
heretofore untested feature of stereoselection at the
steady state. According to the model in Scheme 1, the
rates of cyclization of the first pair of radicals (3a ,b)
should be different on the basis of the Beckwith-Houk
model;4 radical 3b cyclizes preferentially while radical
3a is preferentially reduced. This predicts that the
structure of the major reduced monobromide should be
5a (X ) Br). To prove this hypothesis, we first needed
authentic samples of 5 for structure assignment. These
were prepared in a straightforward, nonstereoselective
fashion, as shown in Scheme 3. Methylation of 10 gave
14, which was decarboxylated to give a mixture of 15a ,b.
Reduction with LAH gave alcohols 16a ,b. These alcohols
were separable, and the isomers were carried on to 5a ,b.
The isomers of 5 were readily assigned by conducting
(13) Nozaki, K.; Oshima, K.; Utimoto, K. J . Am. Chem. Soc. 1987,
109, 2547-2549.

2986 J . Org. Chem., Vol. 67, No. 9, 2002
Stalinski and Curran
Sch em e 3a
Ta ble 5. Red u ction of 2 a t Differ en t Con cen tr a tion s
[2]
(M)
4t
(%)
4c
(%)
4t + 4c
(%)
7
(%)
total
(%)
entry
4t/4c
1
2
3
4
0.007
0.02
0.03
0.06
5.8
5.6
5.4
5.9
63
63
64
68
11
11
12
12
74
74
76
80
4
21
13
9
78
95
89
88
a
Reagents and conditions: (a) NaH/MeI, THF, 50 °C for 6 h,
85%; (b) NaCl/H2O, DMSO, 160-170 °C for 2 days, 92%; (c) LAH,
Et2O, 0 °C for 2 h, 60-75%; (d) MsCl/Et3N, CH2Cl2, 0 °C for 2 h,
∼100%; (e) LiBr or NaI, acetone, reflux for 12 h, 85-96%.
reactions of 1,3-diiodopropanes, but the high total yield
of products (96-99%) rule out such a competition in this
case.14
Ta ble 4. Red u ction s of Diiod id e 2 w ith P h 3Sn H a t 65 °C
Surprisingly, the results with the iodide (Table 4) did
not match those with the bromide (Table 3). With the
exception of the initial syringe pump experiment (entry
1, 50/50 ratio), the ratios of the cyclic products as a
function of starting tin hydride concentration are ap-
proximately twice as high as those of dibromide 1 at any
given concentration. We considered three possible expla-
nations for this behavior: (1) equilibration of intermedi-
ate radicals, (2) reactions of intermediate radicals 3a ,b
(X ) Br) and 3a ,b (X ) I) with Ph3SnH at different rates
(Br versus I), or (3) cyclizations of the intermediate
bromide- and iodide-substituted radicals 3a ,b with dif-
ferent rate constants. The first of these possible explana-
tions is a radical precursor effect, while the second and
third are radical substituent effects. The third explana-
tion seems most likely, but is also most difficult to test
because the ratios of monoreduced products as a function
of time cannot be measured.
[Ph3SnH]
(M)
4a
(%)
4b
(%)
4a + 4b
(%)
7
(%)
total
(%)
entry
4a /4b
1
2
3
4
5
6
7
a
1.0
1.3
2.0
2.6
3.3
5.5
5.8
50
50
99
99
99
96
92
81
75
0
0
0
3
7
99
99
99
99
99
97
96
0.025
0.05
0.075
0.10
0.15
0.20
56.5 42.5
66
70
71
69
64
33
26
21
12
11
16
21
a
Syringe pump addition.
cyclizations: 5a gave 4t, and 5b gave 4c (see the
experimental details in the Supporting Information).
Equilibration by iodine transfer seems unlikely since
the relative rates of iodine transfer (kI[RI]) are small
compared to those of cyclization (kc; see below for values)
and reduction.15 Iodine transfer is a bimolecular process,
and cyclization is unimolecular, so experiments with
different ratios of 2 to Ph3SnH at a set Ph3SnH concen-
tration (0.2 M) were undertaken. As shown in Table 5,
the product ratio 4t/4c was constant within experimental
error, thus further suggesting that reversible iodine
transfer does not play a role when diiodides are used as
precursors.
To rule out any other unusual radical precursor effects,
we measured the rate constants for cyclization of the
diastereomeric radicals derived from monohalides 5a ,b
(X ) Br or I). The same radical 6a or 6b is generated
from both the iodide and the bromide, and these are the
second set of radicals that partition between reduction
and cyclization in Scheme 1. Accordingly, these experi-
ments directly measure kfast2 and kslow2 in Scheme 1.
Kinetic experiments using the standard protocol were
conducted on diastereomerically pure compounds, and
the results of these experiments are shown in Table 6.
Rate constants were calculated from the raw data in the
usual way.16 The iodide and bromide precursors provided
Radical cyclizations of 1 with a small amount of
Ph3SnH (2.2 equiv) provided a complex mixture of un-
reacted dibromide 7 (about one-third), monobromides
5a ,b (about one-third) and 8t ,c (about one-third), and
hydrocarbons 4t,c and 7. Co-injection of a sample from
the reaction mixture with the authentic samples of
monobromides 5a and 5b showed that 5a is the major
diastereoisomer. Quantitation is difficult because the
peaks are close together (see the GC chromatograms in
the Supporting Information), but we estimate that the
ratio 5a /5b was at least 90/10. This verifies the hypoth-
esis that 3a is reduced to 5a faster than 3b is reduced to
5b. This must be due to the increased concentration of
3b of the steady state.
We next studied the tin hydride mediated cyclization
of the diiodide 2. In limited prior studies, diiodides and
dibromides have provided similar results even though the
first pairs of radicals 3a ,b are different (X ) Br or I).8
Similar results are expected if the first pairs of radicals
3a ,b bearing iodine or bromine have identical or similar
rate constants for cyclization.
Using the standard protocol, we cyclized diiodide 2 at
different Ph3SnH concentrations, and the data for these
experiments are summarized in Table 4. The reaction
times for 2 (<15 min) were much shorter, presumably
because iodine is a better radical precursor than bromine.
We were not able to detect any monoiodide intermediates
even with careful monitoring by GC/MS. The short
reaction times also minimized the problem of hydrostan-
nylation, and high total yields (95-100%) were observed
in all experiments. Cyclopropane formation can occur in
(14) Curran, D. P.; Gabarda, A. E. Tetrahedron 1999, 55, 3327.
(15) Rate constants for iodine transfer between primary alkyl
radicals are more than an order of magnitude lower than those from
Ph3SnH, and Ph3SnH is also used in large excess. See: (a) Newcomb,
M.; Sanchez, R. M.; Kaplan, J . J . Am. Chem. Soc. 1987, 109, 1195-
1199. (b) Newcomb, M.; Curran, D. P. Acc. Chem. Res. 1988, 21, 206-
214. (c) Drury, R. F.; Kaplan, L. J . Am. Chem. Soc. 1972, 94, 3982-
3986. (d) Chatgilialoglu, C.; Newcomb, M. In Advances In Organometallic
Chemistry; West, R., Hill, A. F., Eds.; Academic Press: San Diego, 1999;
Vol. 44; p 67.

Radical Cyclizations of Acyclic Dihalides
J . Org. Chem., Vol. 67, No. 9, 2002 2987
Ta ble 6. Ra te Con sta n t Mea su r em en ts of Ra d ica ls
Ta ble 8. Tem p er a tu r e Effects on Dih a lid e Cycliza tion s
Der ived fr om 5a ,b (X ) Br , I)
t
4t
(%)
4c
(%)
4t + 4c
(%)
7
(%)
yield
(%)
[Bu3SnH]
(M)
4
7
(%)
kc
X
(°C)
4t/4c
compd
X
(%)a
(M-1 s-1
)
Br, 1
Br, 1
I, 2
0
-20
0
6.9
10.0
11.1
17.0
44
30
56
34
6
3
5
2
50
33
61
36
0
0
11
27
50
33
72
63
5a
Br
0.7
1.0
0.2
0.5
0.7
1.0
0.5
85.7
77.0
94.8
88.1
31.4
25.0
38.0
14.3
23.0
5.2
11.9
68.6
75.0
61.9
2.52 × 107
2.09 × 107
2.19 × 107
2.21 × 107
1.92 × 106
2.00 × 106
1.84 × 106
I
I, 2
-20
5b
Br
I
Thus, while the conclusion is not definitive, we feel that
the most likely reason for the differences in product ratios
between the iodide and the bromide is that the kfast/kslow
ratio for cyclization of radical 3a ,b with X ) I is about 2
times higher than that for radical 3a ,b with X ) Br.
It was also of interest to check the influence of
temperature on the 4t/4c ratio from dibromide 1 and
diiodide 2. Experiments were performed at two different
temperatures, 0 and -20 °C, using Et3B/O2 to initiate
radical chains. The results of these experiments are
summarized in Table 8. Experiments were conducted at
a 0.2 M concentration of Ph3SnH so the data can be
compared with the relevant experiments conducted at 65
°C (Tables 3 and 4, entry 7).
The temperature-dependent selectivities range from
7/1 and 11/1 at 0 °C to 10/1 and 17/1 at -20 °C for 1 and
2, respectively. Again the ratios 4t/4c resulting from 2
were higher than those resulting from 1. The modest
overall yields (33-72%) are apparently connected to the
hydrostannylation of the triple bond, which competes
effectively due to the long reaction times at lower
temperatures. Furthermore, the reduced yields of 4 (33-
61%) indicate that tin hydride reduction competes more
favorably with radical cyclization as the temperature is
lowered and that better preparative results could be
obtained by reducing the tin hydride concentration.
Nevertheless, the highest 4a /4b ratio (17) suggests that
the process could be important not only from a theoretical
but also from a preparative point of view.
a
5a provided 4t, and 5b provided 4c.
Ta ble 7. Red u ction of 1 w ith F ou r Differ en t Hyd r id es a t
a Con sta n t H-Tr a n sfer Ra te
[hydride]
(M)
4t
4c 4t + 4c
7
entry
hydride
4t/4c (%) (%)
(%)
(%)
1
2
Ph3SnH
0.025
0.088
1.4
1.6
1.6
1.4
1.4
1.4
1.4
57 41
61 38
61 37
57 42
58 41
52 38
53 37
98
99
98
99
99
90
90
1
0
0
0
0
0
0
nBu3SnH
3
4
(Me3Si)3SiH
nBu3GeH
0.42
1.42
the same rate constants for the slow and fast cyclizations.
This strongly suggests that the differences between the
dibromide 1 and the diiodide 2 emanate from radical
substituent effects, not radical precursor effects. The
cyclization of the radical derived from 5a (2 × 107 m-1
s-1) is about 10 times faster than that of the radical
derived from its diastereomer 5b (2 × 106 m-1s-1).
Turning to differential effects of bromine and iodine
as substituents in one of the component reactions, we
postulated that substituent effects on hydrogen-transfer
reactions were unlikely because most primary radicals
react with a given metal hydride with similar rate
constants16 and because the bromine/iodine substituent
is three atoms removed from the reacting radical. To
probe the involvement of differential hydrogen transfer,
we performed a series of experiments in which the
concentrations of a series of metal hydrides were selected
so kH[hydride] would be the same within experimental
error (Table 7). This was done by adjusting the concen-
tration of every reaction to offset the difference in relative
reactivity between that hydride and the others.16 The
reactivity of the tested hydrides follows the order Ph3-
SnH > Bu3SnH > (Me3Si)3SiH > Bu3GeH. The experi-
ments afforded very similar ratios of the products in all
the investigated cases, and this suggests that there are
no usual substituent effects on hydrogen transfer.
3. Con clu sion s
The results described in this paper expand and solidify
our understanding of stereoselection at the steady state
with radical reactions. It should be generally possible to
design such processes by extending the well established
Beckwith-Houk model for stereoselection to encompass
substrates with two radical precursors and one radical
acceptor, as illustrated in Figure 2. Intermediate mono-
halide products have been observed for the first time. And
while their behavior was difficult to quantitate because
they appear and disappear, it was consistent with the
kinetic model for steady-state stereoselection. Clear, if
small, differences in the results as a function of the
radical precursor (diiodide better than dibromide) have
been observed for the first time, and these have been
tentatively attributed to differing rates of cyclization at
the monohalo stage. In the big picture, kinetic theory8a
in this area still far outpaces experiments. With a solid
understanding on one type of steady-state stereoselection
process in hand, the challenge now becomes discovering
examples of other types of radical-based processes that
have been modeled as well as nonradical examples.
(16) The rate constants for hydrogen abstraction from these hydrides
by primary radicals at 25 °C are reported as 5 × 106, 2.4 × 106, 3.8 ×
105, and 1 × 105 M-1 s-1, respectively. The rate constants for hydrogen
abstraction from the hydrides at 65 °C by primary radicals were
calculated from the reported temperature-dependent Arrhenius equa-
tion. See: Newcomb, M. Tetrahedron 1993, 49, 1151-1176 and ref 14d
therein.

2988 J . Org. Chem., Vol. 67, No. 9, 2002
Stalinski and Curran
compounds, and copies of GC chromatograms from experi-
ments to identify dibromides 5 and 8. This material is available
free of charge via the Internet at http://pubs.acs.org.
Ack n ow led gm en t. We thank the National Science
Foundation for funding this work.
Su p p or tin g In for m a tion Ava ila ble: Complete experi-
1
mental details, copies of H and13C NMR spectra for all new
J O011056U