Conformational memory in enantioselective radical reductions and a new radical clock reaction

Article abstract of DOI:10.1021/ja002068k

The radical decarboxylation of acid 1 led to tetrahydropyran 3 with significant optical activity. This transfer of chirality is an example of a conformational memory effect and derives from the slow ring inversion of the atropisomer 2 and its enantiomer 2

Full text of DOI:10.1021/ja002068k

9386
J. Am. Chem. Soc. 2000, 122, 9386-9390
Conformational Memory in Enantioselective Radical Reductions and a
New Radical Clock Reaction
Alexandre J. Buckmelter, Angie I. Kim, and Scott D. Rychnovsky*
Contribution from the Department of Chemistry, UniVersity of California, IrVine, California 92697-2025
ReceiVed June 9, 2000
Abstract: The radical decarboxylation of acid 1 led to tetrahydropyran 3 with significant optical activity. This
transfer of chirality is an example of a conformational memory effect and derives from the slow ring inversion
of the atropisomer 2 and its enantiomer 2′. A similar conformational memory effect was observed in the
reductive decyanation and reductive lithiation of cyanohydrin 4. We propose that the racemization of radicals
with chiral conformations such as 2 could be used as new radical clocks. The rate of racemization of 2 was
evaluated and used to determine the rate of H atom transfer to 2 by different H atom donors. The rate of
racemization of 2 is 5.7 × 108 s-1 at 22 °C, which is about five times faster than ring opening of a
cyclopropylmethyl radical.
The field of enantioselective radical reactions is rapidly
to complete equilibration. The 9-decalyl radical is the most
prominent example of this type.8 The slow conformational
interconversion of cyclohexyl radicals has also been invoked
to explain the stereoselectivity in radical additions to substituted
cyclohexenes.7,9 The present investigation examines a system
of the later type, where a slow conformational interconversion
was found to impose a barrier to equilibration despite a very
low barrier to radical inversion.
The reaction of an optically pure substrate with a single
stereogenic center was envisioned to proceed through a radical
intermediate in which the only stereogenic center was destroyed.
Could an optically active product be produced in such a
reaction? The radical decarboxylation of carboxylic acid 1 to
give radical 2 (Scheme 1) is an example. Radical 2 does not
contain a stereogenic center, and a cursory inspection might
classify it as achiral. However, radical 2 is chiral by virtue of
its three-dimensional structure and only behaves as an achiral
molecule when in equilibrium with its enantiomer, 2′. Inter-
conversion of 2 and 2′ involves a radical inversion with a very
low inversion barrier estimated at e0.5 kcal/mol10 and a chair-
chair ring inversion with a more substantial barrier estimated
to be between 5 and 10 kcal/mol.11 Because the key ring
inversions of 2 and 2′ involves rotations around single bonds,
they can be classified as rapidly interconverting atropisomers.
When the reaction of the chiral radical 2 is fast compared with
the ring inversion and when the addition reaction of 2 is
stereoselective, an optically active product will be produced.
advancing with the emergence of both auxiliary-based and chiral
Lewis acid-based strategies.1 Our recent investigation of non-
equilibrium radical reactions made us aware of a third strategy
for introducing chirality in radical reactions: the retention of
stereochemical information due to slow equilibration of chiral
conformations.2 A similar effect has been observed in the
alkylation of certain enolates and has been described as a
“memory of chirality”.3 A demonstration of memory of chirality
in radical reactions is described herein.
The reactions of radicals are very rapid, and it has long been
recognized that in certain situations these reactions may be
competitive with conformational isomerizations. The recombi-
nation of radical pairs often shows a memory effect. Our primary
interest, however, is in the reaction of free radicals with bulk
reagents. Although most alkyl radicals have very low barriers
to inversion, there are a few structures such as cyclopropyl
radicals,4 1,3-dioxolan-2-yl radicals,5 and some vinyl radicals6
that have modest barriers to inversion.7 These radicals can react
with partial or complete retention of configuration, depending
upon the conditions and trapping reagents used in the reaction.
Less common are systems where the radical inverts rapidly but
the conformation of the rest of the molecule imposes a barrier
(1) (a) Sibi, M. P.; Porter, N. A. Acc. Chem. Res. 1999, 32, 163-171.
(b) Curran, D. P.; Porter, N. A.; Giese, B. Stereochemistry of Radical
Reactions; VCH: New York, 1996.
(2) Buckmelter, A. J.; Powers, J. P.; Rychnovsky, S. D. J. Am. Chem.
Soc. 1998, 120, 5589-5590.
(3) (a) Fuji, K.; Kawabata, T. Chem. Eur. J. 1998, 4, 373-376. (b) Giese,
B.; Wettstein, P.; Stahelin, C.; Barbosa, F.; Neuburger, M.; Zehnder, M.;
Wessig, P. Angew. Chem., Int. Ed. 1999, 38, 2586-2587. (c) Sauer, S.;
Schumacher, A.; Barbosa, F.; Giese, B. Tetrahedron Lett. 1998, 39, 3685-
3688. (d) Schmalz, H. G.; de Koning, C. B.; Bernicke, D.; Siegel, S.;
Pfletschinger, A. Angew. Chem., Int. Ed. 1999, 38, 1620-1623. (e) Vedejs,
E.; Fields, S. C.; Hayashi, R.; Hitchcock, S. R.; Powell, D. R.; Schrimpf,
M. R. J. Am. Chem. Soc. 1999, 121, 2460-2470.
(8) (a) Bartlett, P. D.; Pinock, R. E.; Rolston, J. H.; Schindel, W. G.;
Singer, L. A. J. Am. Chem. Soc. 1965, 87, 2590-2596. (b) Greene, F. D.;
Lowry, N. N. J. Org. Chem. 1967, 32, 882-885. (c) Greene, F. D.; Lowry,
N. N. J. Org. Chem. 1967, 32, 875-883. (d) Roberts, B. P.; Steel, A. J.
Tetrahedron Lett. 1991, 32, 5385-5388.
(9) (a) Goering, H. L.; Sims, L. L. J. Am. Chem. Soc. 1955, 77, 3465-
3469. (b) Goering, H. L.; Relyea, D. I.; Larsen, D. W. J. Am. Chem. Soc.
1956, 78, 348-353.
(4) Gawronska, K.; Gawronska, J.; Walborsky, H. M. J. Org. Chem.
1991, 56, 2193-2197.
(10) Griller, D.; Ingold, K. U.; Krusic, P. J.; Fischer, H. J. Am. Chem.
Soc. 1978, 100, 6750-6752.
(5) (a) Kobayashi, S. O.; Simamura, O. Chem. Lett. 1973, 695-8. (b)
Kobayashi, S. O.; Simamura, O. Chem. Lett. 1973, 699-702.
(6) (a) Fry, A. J.; Mitnick, M. A. J. Am. Chem. Soc. 1969, 91, 6207-
6208. (b) Ohnuki, T.; Yoshida, M.; Simamura, O. Chem. Lett. 1972, 797-
801.
(11) Cyclohexane and tetrahydropyran have ring-inversion barriers of
about 10 kcal/mol, but the cyclohexyl radical has a ring-inversion barrier
of only 5.5 kcal/mol. (a) Eliel, E. L.; Wilen, S. H. Stereochemistry of
Organic Compounds; John Wiley & Sons: New York, 1994; pp 740-742.
(b) Roberts, B. P.; Steel, A. J. J. Chem. Soc., Perkin Trans. 2 1992, 2025-
2029.
(7) Simamura, O. Top. Stereochem. 1969, 4, 1-37.
10.1021/ja002068k CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/16/2000

Conformational Memory in Radical Reductions
J. Am. Chem. Soc., Vol. 122, No. 39, 2000 9387
Table 2. Enantioselectivity of the Reduction of 5 in PhSH As a
Function of Concentration and Temperature
Scheme 1
entry
temp (°C)
[X-H] (M)
% yield
ratio 3:3′
% ee
1
2
3
4
5
6
7
8
9
-40
-40
-40
0
0
0
22
22
22
0.5
50
46
41
77
79
54
72
62
60
69.9:30.1
59.7:40.3
50.6:49.4
57.6:42.4
53.0:47.0
50.3:49.7
54.2:45.8
51.2:48.8
50.7:49.3
39.8
19.4
1.2
15.2
6.0
0.5
8.4
0.15
0.01
0.5
0.15
0.01
0.5
0.15
0.01
2.4
1.4
Table 1. Enantioselective Reduction of 5 in the Presence of
Different H Atom Donors
a nearly racemic product (entries 1-3) at each concentration,
as expected from its relatively slow reaction with electron-rich
radicals such as 2/2′. The electron-deficient H atom donor
t-BuSH reacts rapidly with 2/2′ to give optically active product
3 (entries 4-7). The optical purity of the product decreases with
decreasing t-BuSH concentration. The reaction of t-BuSH should
be first order in radical and first order in thiol; reduced thiol
concentrations allow more time for the racemization of radical
2. Benzenethiol is a more efficient H atom donor than t-BuSH
and leads to tetrahydropyran 3 in 86% ee when present at 1 M
concentration (entry 8). Benzeneselenol is an even more efficient
H atom donor,14 and it produces 3 in 32% ee when present at
a nominal concentration15 of 0.05 M, compared with 23% ee
when PhSH was present at the same concentration (entries 10
and 12). The optical purity of 3 also depends on the temperature
of the decarboxylation reaction. Table 2 shows the results from
decarboxylation of 5 at different temperatures and with different
concentrations of PhSH. As the temperature increases, the
maximum enantiomeric excess decreases until the enantiomeric
excess of 3 is only 8% with 0.5 M PhSH at 22 °C. The efficiency
of the chiral memory effect in the decarboxylation of 5 depends
on the concentration, temperature, and identity of the H atom
donor.
Cyanohydrins can be reduced to give configurationally stable
alkyllithium reagents.16 The reduction of cyanohydrin 4 by
lithium di-tert-butylbiphenylide (LiDBB) is outlined in Scheme
2 and Table 3. Reduction of 4 to alkyllithium 6, followed by
protonation with methanol, gave reduced tetrahydropyran 3 of
low to moderate optical purity. Inverse addition led to products
with low optical purity, whereas direct addition of the substrate
to 0.63 M LiDBB gave products with ca. 40% ee. Optically
active alkylithium reagent 6, prepared using 0.63 M LiDBB,
was trapped with CO2 and esterified with CH2N2 to give the
methyl ester 7 in 42% yield and 49% ee. Trapping of
R-alkoxylithium reagents with CO2 proceeds with retention of
configuration.17 Ester 7 was shown by direct correlation to have
the same absolute configuration as 4. In addition, the same major
enantiomer of tetrahydropyran 3 was produced by each of the
entry
donor
[X-H] (M)
% yield
ratio 3:3′
% ee
1
2
3
4
5
6
7
8
9
Bu3SnH
Bu3SnH
Bu3SnH
t-BuSH
t-BuSH
t-BuSH
t-BuSH
PhSH
PhSH
PhSH
PhSH
PhSeH
1.0
0.5
0.05
2.0
1.0
0.5
0.05
1.0
70
78
77
29
83
31
30
92
75
58
67
28
51.4:48.6
50.7:49.3
50.0:50.0
67.1:32.9
62.9:37.1
61.7:38.3
51.1:48.9
93.2:6.8
84.8:15.2
61.5:38.5
52.2:47.8
67.3:32.7
2.6
1.5
0.0
34.3
25.8
23.0
2.3
86.5
69.7
23.0
4.3
0.5
10
11
12
0.05
0.01
0.05
34.6
In such a reaction, the memory of the chirality is held by the
conformation of the radical, even though the original stereogenic
center has been destroyed. Radical decarboxylation of 1 is an
example of memory of chirality in a radical reaction.
Carboxylic acid 1 (Scheme 1) was prepared in optically pure
form12 and converted to its corresponding N-hydroxypyridine-
2-thione ester 5. Photolysis of the ester in toluene at -78 °C
with 1 M PhSH as a hydrogen atom donor led to the expected
product 3 and its enantiomer 3′ in a 93:7 ratio. Thus, in the
presence of a very efficient hydrogen atom trap, reduction of 2
is competitive with ring inversion and an optically active product
is produced. The radical reduction of 1 is an unusual example
of a self-regenerating stereocenters (SRS) reaction as designated
by Seebach,13 although in this case the added chiral element is
the ring conformation rather than an added stereogenic center.
The efficiency of the chiral memory effect will depend on
the relative rates of racemization and productive coupling. Table
1 shows the results of decarboxylations of 5 with different H
atom donors at -78 °C in toluene. Tributyltin hydride leads to
(14) Newcomb, M.; Choi, S. Y.; Horner, J. H. J. Org. Chem. 1999, 64,
1225-1231.
(15) Benzeneselenol was prepared in situ from Bu3SnH and PhSeSePh:
Crich, D.; Yao, Q. J. Org. Chem. 1995, 60, 84-88.
(16) (a) Skalitzky, D. Ph.D. Thesis, University of Minnesota, 1992. (b)
Rychnovsky, S. D.; Swenson, S. S. J. Org. Chem. 1997, 62, 1333-1340.
(c) Guijarro, D.; Yus, M. Tetrahedron 1994, 50, 3447-3452.
(17) Still, W. C. J. Am. Chem. Soc. 1978, 100, 1481-1487.
(12) Details for the preparation and resolution of 1 are provided in the
Supporting Information.
(13) Seebach, D.; Sting, A. R.; Hoffmann, M. Angew. Chem., Int. Ed.
Engl. 1996, 35, 2708-2748.

9388 J. Am. Chem. Soc., Vol. 122, No. 39, 2000
Buckmelter et al.
Scheme 2
Scheme 3
for ring inversion. The most intriguing example is the reduction
of 4 by high concentrations of lithium in NH3. The optical
purity of the product is not a simple function of Li concentration.
Above 4 M, Li solutions in ammonia have a metallic gold
appearance, and these gold-colored solutions lead to a rapid
increase in the % ee of the product with further increases in Li
concentration. At 6.4 M, the product 3 was obtained in 90% ee
(entry 8). Evidently, the highly concentrated solution of lithium
reduces the radical 2 much faster than the ring inversion.
Although reductive decyanations are sensitive to reaction
conditions, they can show a nearly complete memory of
chirality.
Table 3. Reductive Decyanations of 4 with LiDBB in THF
conditionsa
optical purity of 3
0.12 M LiDBB, -78 °C
0.31 M LiDBB, -78 °C
0.47 M LiDBB, -78 °C
0.63 M LiDBB, -78 °C
Inverse addition, -78 °C
26% ee
30% ee
39% ee
40% ee
9% ee
a The nitrile in THF was added to the solution of Li in NH3 unless
otherwise noted.
Table 4. Results of the Reduction of 4 with Li/NH3
An alternative explanation for the optical activity of the
products at high lithium concentrations is that the reaction
mechanism has changed and a two-electron reduction of the
nitrile is taking place with retention of configuration. This
possibility can be excluded by the following result (Scheme
3): reduction of a single diastereomer of the tetrahydrofuran
nitrile19 at -78 °C gave essentially the same 1.15:1 ratio of
tetrahydrofuran products at lithium concentrations of 0.8 and
6.4 M. Ring inversion of the five-membered ring radical is much
faster than with the six-membered ring radical, and the chiral
ring conformation cannot be intercepted under these conditions.
Under conditions where the six-membered ring shows nearly
complete retention of configuration, Vide supra, the five-
membered ring shows pronounced stereochemical scrambling.
Assuming both substrates react through a common mechanism,
stereospecific two-electron reductions can be excluded.
entry
conditionsa
optical purity of 3
1
2
3
4
5
6
7
8
Li (ca. 1 M)/NH3, -78 °Cb
Li (ca. 1 M)/NH3, -33 °Cc
Na (1.8 M)/NH3, -78 °Cc
Li (0.8 M)/NH3, -78 °Cc
Li (3.5 M)/NH3, -78 °Cc
Li (4.3 M)/NH3, -78 °Cd
Li (5.4 M)/NH3, -78 °Cd
Li (6.4 M)/NH3, -78 °Cd
7% ee
16% ee
16% ee
30% ee
30% ee
10% ee
54% ee
90% ee
a The nitrile in THF was added to a solution of metal in NH3.
b Inverse addition: lithium was added to the nitrile in NH3. c The Li/
NH3 solution was blue in color. d The Li/NH3 solution had a metallic
gold appearance.
routes described, so we conclude that all of the radical reductions
discussed herein proceed with retention of configuration.
Retention of configuration is expected for the memory of
chirality phenomena outlined in Scheme 1.
Results and Discussion
Racemization as a Radical Clock Reactions. In the decar-
boxylation of 5, the optical purity of the product is a reflection
of the lifetime of the radical. The first-order racemization
reaction competes with other radical processes, much like the
5-hexenyl radical cyclization or the opening of cyclopropyl-
methyl radicals. The rate of racemization (kR) and the rate of H
atom transfer (kH) are related to the enantiomeric ratio of the
products 3 and 3′ by eq 1.20
A memory effect also was observed in the reductive decya-
nation of optically active nitrile 4 (Table 4). Reductive decya-
nations of nitriles are stepwise reactions that proceed via a
radical intermediate.18 Reduction of 4 by addition to a 0.8 M
Li/NH3 solution at -78 °C gave tetrahydropyran 3 in 30% ee
(entry 4). However, the optical purity of the product is very
sensitive to conditions. Addition of lithium metal to the nitrile
(inverse addition) gave the product in only 7% ee (entry 1).
Presumably, the concentration of the lithium in the reaction was
low and thus the lifetime of the intermediate radical 2/2′ was
long enough to allow significant racemization. When the
reduction was carried out with lithium at -33 °C, the product
was produced in only 16% ee because ring inversion takes place
faster at higher temperatures (entry 2). Reduction of 4 with
sodium at -78 °C gave a product with 16% ee (entry 3). Sodium
is a less powerful reducing agent than lithium, and the slower
reduction of the radicals 2/2′ with this reagent allows more time
kR ) kH[PhSH]2[3′]/([3] - [3′])
(1)
One can calculate the approximate rate of H atom transfer to
2 by extrapolating Arrhenius data for the rate of H atom transfer
(19) Rychnovsky, S. D.; Dahanukar, V. H. J. Org. Chem. 1996, 61,
7648-7649.
(20) Eq 1 was derived by assuming that 2 and 2′ react with perfect axial
selectivity and that the product ratios are controlled by simple partitioning
of radical 2 between the optically pure product [3] and the racemic product
[3rac]. These assumptions leads to the following equation, which simplifies
to eq 1 when evaluated in terms of the observed products [3] and [3′]:
[3]/[3rac] ) kH/kR[RSH].
(18) Rychnovsky, S. D.; Powers, J. P.; Lepage, T. J. J. Am. Chem. Soc.
1992, 114, 8375-8384.

Conformational Memory in Radical Reductions
J. Am. Chem. Soc., Vol. 122, No. 39, 2000 9389
from t-BuSH to a secondary R-alkoxy radical.21 The estimated
rate constant kH is 1.1 × 106 M-1 s-1 at -78 °C. From the
enantiomeric ratio of 3 produced at different concentrations of
t-BuSH (Table 1, entries 4-7), one can calculate that the rate
of racemization (kR) for 2 as 1.9-4.3 × 106 s-1 with an average
value of 3.9 × 106 s-1 at -78 °C, based on 12 experiments.22
The rate of racemization of 2 should be relatively invariant to
most reaction parameters except temperature. We suggest that
the racemization of 2 and related conformationally chiral radicals
could be used as radical clocks. The absolute rate for racem-
ization of 2 is much faster than for the 5-hexenyl radical and
significantly faster than for the cyclopropyl methyl radical but
slower than for the trans-(2-phenylcyclopropyl)carbinyl radi-
cal.23 Racemization reactions have several advantages over more
conventional radical-clock reactions. First, a racemization
reaction is by its very nature irreversible. Second, the down-
stream radical reactions of the initial radical 2 and the rearranged
radical 2′ will have identical rates because they are enanti-
omers.24 Finally, the enantiomeric ratio of 3 can be measured
directly by GC using a chiral column, and the ratio of 3/3′ is
unlikely to be affected by isolation and purification procedures.
The rate of racemization of radical 2 suggests that it could
function as a very rapid radical-clock reaction with unique
advantages over more conventional radical clocks.
The rates of H atom transfer from different H atom donors
can be determined using the racemization of 2 as a radical clock
reaction. In the case of PhSH, the enantiomeric ratios of reduced
product 3 as a function of concentration (Table 1, entries 8-11)
lead to H atom transfer rates of 1.8-2.5 × 107 M-1 s-1 at -78
°C with an average value of 2.0 × 107 M-1 s-1. These values
compare favorably with the rate of 1.8 × 107 M-1 s-1 calculated
from Arrhenius parameters for the PhSH reduction of a
secondary radical.25 The rate of H atom trapping with Bu3SnH
can be calculated from the data in Table 1, although the low
optical purity of the product compromises the precision of the
determination. The rate of H atom transfer from Bu3SnH comes
to 1.1 × 105 M-1 s-1 at -78 °C based on entries 1 and 2. This
rate is faster than the estimated rate of 1.4 × 104 M-1 s-1 at
-78 °C that comes from analysis of the reported Arrhenius
parameters for reduction of an R-alkoxy radical.26 The difference
could be due to the difference in substrates but is in part a
consequence of the disparity in rates between the racemization
and the rate of the H atom transfer. Measuring precise H atom
transfer rates requires products with significant optical enrich-
ment, and these arise from similar rates of racemization and H
atom transfer.
The enantiomeric enrichment of 3 found in the reduction of
5 shows a strong temperature dependence. Table 2 lists the
optical purity of product 3 as a function of temperature and
PhSH concentration. One can calculate the kR/kH ratio for each
example; the average value of kR/kH at each temperature is 0.7
at -40 °C, 2.5 at 0 °C, and 6.5 at 22 °C. Thus, the rate of
racemization increases faster than the rate of H atom transfer
with increasing temperature. In theory one could use these data
to calculate Arrhenius parameters for the racemization of 2, but
unfortunately we are not aware of reliable H atom transfer rates
for the reaction of PhSH with a radical similar to 2. Arrhenius
parameters have been reported for the reaction of secondary
radicals with PhSH, and one can use these parameters to estimate
an approximate rate of racemization for radical 2 of 5.7 × 108
s-1 at 22 °C. Thus, the racemization of 2 is about five times
faster than the ring opening of a cyclopropylmethyl radical.23,27
The rate of racemization for radical 2 should be established with
higher precision, but it is apparent from the estimated rates
reported above that it can function as a very fast radical clock
reaction.
It must be emphasized that in each of the enantioselective
radical reactions described, the radicals themselves are con-
figurationally unstable. Two features are necessary for retention
of configuration: slow ring inversion and a pronounced prefer-
ence for reactions at one face of the radical. All of the present
examples make use of 2-tetrahydropyranyl radicals because the
radicals have a significant facial bias and there is a significant
barrier to ring inversions in six-membered rings. The racem-
ization of radical 2 can be used as a fast radical clock reaction,
with an estimated rate of 5.7 × 108 s-1 at 22 °C. Conformational
interconversions can be very fast, and the use of conformational
interconversions in other ring systems has the potential to
produce new ultrafast radical clock reactions.
Experimental Section28
Procedure for Preparation and Reduction of N-Hydroxypyridine-
2-thione Ester 5 in the Presence of an H Atom Donor: 2-Benz-
yltetrahydropyran (3). All reactions were performed under an inert
atmosphere. To a 0.01 M solution of optically pure 2-benzyltetrahy-
dropyran-2-carboxylic acid (1 equiv) in freshly distilled solvent were
added DMAP (cat.), 2-mercaptopyridine N-oxide (1.5 equiv), and
dodecane (1.0 equiv). The mixture was shielded from light then cooled
to 0 °C. Diisopropylcarbodiimide (1.5 equiv) was then slowly added
to the reaction flask. The reaction was allowed to warm to room
temperature and stirred for 2.5 h. Then a 1-2 mL aliquot from this
stock solution was added to individual reaction flasks, which were
shielded from light. These reaction flasks were then cooled to the
appropriate temperature under argon and the appropriate H atom transfer
agent was added. The mixture was exposed to a 120 V (60 Hz) lamp
at a distance of 30-50 cm from the reaction flask. The reaction mixture
was then photolyzed for 3.5 h at the appropriate temperature. A
precooled (0 °C) solution of potassium trimethylsiloxide (0.5 equiv)
was added to destroy the ester 5, followed by the addition of a saturated
aqueous NaHCO3 solution. The organic layer was extracted with Et2O
and washed with brine then dried with MgSO4. The solution was filtered
through a silica gel plug and then analyzed by GC using a CHIRAL-
DEX â-cyclodextrin permethylated hydroxylpropyl (B-PH) chiral
column (20 m × 0.25 mm) to determine % yield and % ee. The yields
were determined relative to a dodecane internal standard. The GC oven
began at 100 °C and was ramped to 150 °C at 0.3 deg/min. The two
enantiomers 3 and 3′ showed retention times of 37.3 and 38.4 min,
respectively.
(21) The Arrhenius function for the reaction of t-BuSH with a secondary
R-alkoxy radical was determined to be log(ktMS) ) (8.4 ( 0.3) - (2.1 (
0.4)/2.3RT. Johnson, C. C.; Horner, J. H.; Tronche, C.; Newcomb, M. J.
Am. Chem. Soc. 1995, 117, 1684-7.
(22) The decarboxylations of 5 in the presence of t-BuSH at -78 °C
were repeated three times at each concentration. The average value of for
kR was 3.9 × 106 with a standard deviation of 2.2 × 106. The standard
deviation was smaller for reactions run simultaneously in the same cooling
bath, suggesting that some of the scatter can be attributed to imprecise
temperature control.
(23) (a) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13, 317-323.
(b) Newcomb, M. Tetrahedron 1993, 49, 1151-1176. (c) Newcomb, M.;
Manek, M. B. J. Am. Chem. Soc. 1990, 112, 9662-9663.
(24) This assumes that the trapping agents are achiral. Chiral trapping
agents would lead to diastereomeric transition states. However, most trapping
agent of interest are achiral.
(25) The Arrhenius function for the reaction of PhSH with a secondary
radical was determined to be log(ktMS) ) (8.4 ( 0.6) - (3.8 ( 0.9)/2.3RT.
Franz, J. A.; Bushaw, B. A.; Alnajjar, M. S. J. Am. Chem. Soc. 1989, 111,
268-275.
(26) The Arrhenius function for the reaction of Bu3SnH with a secondary
R-alkoxy radical was determined to be log(ktMS) ) (9.26 ( 0.13) - (1.78
( 0.26)/2.3RT. See ref 21.
Procedure for Reductive Decyanation of 4 Using Li/NH3:
2-Benzyltetrahydropyran (3). Anhydrous ammonia (5.0 mL, distilled
(27) Newcomb, M.; Glenn, A. G. J. Am. Chem. Soc. 1989, 111, 275-
277.
(28) The general experimental section is included in the Supporting
Information.

9390 J. Am. Chem. Soc., Vol. 122, No. 39, 2000
Buckmelter et al.
from sodium) was condensed into a flame-dried two-neck flask
equipped with a coldfinger and glass stirbar under Ar at -78 °C. To
make a concentrated (6.4 M Li/NH3) solution, Li wire (221 mg, 31.8
mmol) was cut directly into the ammonia, forming the characteristic
metallic gold color of a Li/NH3 solution (Li/NH3 solutions of con-
centrations e4.0 M generated a deep blue color). This solution was
stirred for 30 min at -78 °C, and 4 (typically 10 mg) was added via
cannula as a solution in THF (1.0 mL). The mixture was stirred for 15
min and then NH4Cl(s) was added through the coldfinger until the
reaction was quenched (as evidenced by a complete dissipation of color).
The ammonia was allowed to evaporate, Et2O and water were added,
and the mixture was extracted twice with Et2O. The crude product was
dried (MgSO4), filtered, concentrated, and taken up in benzene. A small
amount of DDQ (typically 10 mg) was then added, and the yellow-
orange reaction mixture was stirred overnight. The mixture was loaded
directly onto a column and flashed (SiO2, 10% EtOAc/hexanes) to
afford the product as a pale yellow oil (consistently isolated in >80%
to the THF solution, and the deep green mixture was stirred for 5 h.
The LiDBB solution was cooled to -78 °C, and (S)-(-)-4 (65.1 mg,
0.32 mmol, >95% ee) was added as a solution in THF (1.0 mL) via
cannula. The mixture was stirred for 10 min at -78 °C, and then CO2-
(g) was bubbled through the solution for 30 min (the solution color
goes from malachite green to orange). The reaction (and unreacted
lithium) was quenched with MeOH, warmed to room temperature,
basified with 2 N NaOH, and extracted with CH2Cl2 (2×). From the
basic extracts, 2-benzyltetrahydropyran (3) could be isolated (19.6 mg,
0.11 mmol, 34% yield, 41% ee). The aqueous layer was acidified with
6 N HCl, extracted with CH2Cl2 (3 ×), and dried (Na2SO4). The crude
acid was taken up in Et2O, cooled to 0 °C, and treated with excess of
an ethereal solution of CH2N2. The yellow mixture was quenched with
HOAc, neutralized with saturated NaHCO3, extracted with Et2O (3 ×),
and dried (MgSO4). Chromatography (SiO2, 10% EtOAc/hexanes)
afforded 7 as a yellow oil (31.9 mg, 0.14 mmol, 42% yield, 49% ee).
1
yield). IR (neat) 3027, 2936, 2843, 1454, 1088 cm-1; H NMR (400
Acknowledgment. Support has been provided by the
University of California, Irvine. S.D.R. acknowledges Pfizer and
Zeneca for support of his research program. A.J.B. thanks the
Department of Education and Abbott Laboratories for fellowship
support.
MHz, CDCl3) δ 7.34-7.24 (m, 5 H), 4.04-4.00 (m, 1 H), 3.53 (dddd,
J ) 10.8, 6.5, 6.5, 2.1 Hz, 1 H), 3.45 (td, J ) 11.7, 2.5 Hz, 1 H), 2.93
(dd, J ) 13.6, 6.6 Hz, 1 H), 2.69 (dd, J ) 13.6, 6.5 Hz, 1 H), 1.65-
1.61 (m, 1 H), 1.54-1.32 (m, 5 H); 13C NMR (125 MHz, CDCl3) δ
138.8, 129.3, 128.1, 126.0, 78.7, 68.6, 43.1, 31.4, 26.0, 23.4. HRMS
(EI-GC-MS) m/z calcd for C12H16O (M+) 176.1201, found 176.1196.
Procedure for Reductive Lithiation of 4 Using LiDBB, Trapping
with CO2, and Esterification with CH2N2: 2-Benzyl-2-carbomethoxy-
tetrahydropyran (7) and 2-Benzyltetrahydropyran (3). A 0.47 M
solution of LiDBB was made by dissolving 4,4′-di-tert-butylbiphenyl
(500 mg) in THF (4.0 mL) at 0 °C, adding a crystal of 1,10-
phenanthroline, and titrating with n-BuLi. Excess Li wire was added
Supporting Information Available: The synthesis and
resolution of 1, 4, and 7 are described as well as the absolute
configuration of these compounds by X-ray analysis of a
diastereomeric derivative. This material is available free of
charge via the Internet at http://pubs.acs.org.
JA002068K