Synthesis and reactivity of new chiral bicyclic phospholanes as acyl-transfer catalysts

Article abstract of DOI:10.1021/jo0519155

Synthesis of chiral phosphines 1a, 14a, and 18a as nucleophilic catalysts for anhydride activation and kinetic resolution of alcohols is described. Radical cyclization of alkenylphosphines produced the phosphabicyclooctane (PBO) core of catalysts 1a and 14a, while 18a was made by quenching a metallocycle precursor with dichlorophenylphosphine. Catalysts 1a and 14a are less reactive, while 18a is comparable to the most reactive catalysts in the PBO family. The preferred ground-state geometries of phosphine-boranes were identified using computational methods, and were correlated with the catalytic reactivity of the corresponding free phosphines.

Full text of DOI:10.1021/jo0519155

Synthesis and Reactivity of New Chiral Bicyclic Phospholanes as
Acyl-Transfer Catalysts
James A. MacKay and Edwin Vedejs*
Department of Chemistry, UniVersity of Michigan, Ann Arbor, Michigan 48109
edVed@umich.edu
ReceiVed September 13, 2005
Synthesis of chiral phosphines 1a, 14a, and 18a as nucleophilic catalysts for anhydride activation and
kinetic resolution of alcohols is described. Radical cyclization of alkenylphosphines produced the
phosphabicyclooctane (PBO) core of catalysts 1a and 14a, while 18a was made by quenching a
metallocycle precursor with dichlorophenylphosphine. Catalysts 1a and 14a are less reactive, while 18a
is comparable to the most reactive catalysts in the PBO family. The preferred ground-state geometries of
phosphine-boranes were identified using computational methods, and were correlated with the catalytic
reactivity of the corresponding free phosphines.
Introduction
nucleophilic phosphide displacement of bismesylates used to
prepare 1-4.2,7 Furthermore, the relationship between catalyst
structure and reactivity is not understood sufficiently to identify
potentially useful catalysts. In an effort to address both issues,
we have targeted ring-fused phospholanes that can be accessed
using radical cyclization and carbometalation approaches, as
described below.
Recently, we reported that bicyclic phosphabicyclo[3.3.0]-
octane (PBO) catalysts 1-4 demonstrate impressive reactivity
for anhydride activation and good to excellent enantioselectivity
in the kinetic resolution of arylalkylcarbinols1-4 and allylic
alcohols.5 While the PBO catalysts can be very effective in
Results and Discussion
Phosphinyl radical addition to olefins is well-known,8 and
there are several examples of analogous intramolecular additions
to make five- and six-membered cyclic phosphines.9-13 To test
(6) For a recent alternative approach, see: Pakulski, Z.; Kwiatosz, R.;
Pietrusiewicz, K. Tetrahedron 2005, 61, 1481.
(7) Burk, M. J.; Feaster, J. E.; Harlow, R. L. Tetrahedron: Asymmetry
1991, 2, 569.
(8) (a) Stacey, F. W.; Harris, J. F. Organic Reactions; Wiley: New York,
1963; Vol. 13, pp 150-376. (b) Emsley, J.; Hall, D. The Chemistry of
Phosphorus; Wiley: New York, 1976; pp 352-378. (c) Bentrude, W. G.
In Free radicals; Kochi, J. K., Ed.; Wiley: New York, 1973; Vol. 2, pp
595-663.
(9) Davies, J. H.; Downer, J. D.; Kirby, P. J. Chem. Soc. C 1966, 245.
(10) Krech, F.; Issleib, K. Z. Anorg. Allg. Chem. 1976, 425, 209. Krech,
F.; Issleib, K.; Zschunke, A. Z. Anorg. Allg. Chem. 1991, 600, 195. Krech,
F.; Issleib, K.; Zschunke, A.; Muegge, C.; Skvorcov, S. Z. Anorg. Allg.
Chem. 1991, 594, 66. Krech, F.; Zschunke, A.; Issleib, K. Z. Anorg. Allg.
Chem. 1993, 619, 989. Krauss, B.; Mugge, C.; Ziemer, B.; Zschunke, A.;
Krech, F. Z. Anorg. Allg. Chem. 2001, 627, 1542. Krech, F.; Krauss, B.;
Zschunke, A.; Mugge, C. Z. Anorg. Allg. Chem. 2003, 629, 1475.
(11) Brandt, P. F.; Schubert, D. M.; Norman, A. D. Inorg. Chem. 1997,
36, 1728.
kinetic resolutions, several issues remain to be addressed. So
far, little has been done to evaluate alternative methodology to
access the phosphabicyclo[3.3.0]octane subunit4,6 beyond the
(1) Vedejs, E.; Daugulis, O.; MacKay, J. A.; Rozners, E. Synlett 2001,
1499.
(2) (a) Vedejs, E.; Daugulis, O. J. Am. Chem. Soc. 1999, 121, 5813. (b)
Vedejs, E.; Daugulis, O. J. Am. Chem. Soc. 2003, 125, 4166.
(3) Vedejs, E.; Daugulis, O.; Harper, L. A.; MacKay, J. A.; Powell, D.
R. J. Org. Chem. 2003, 68, 5020.
(4) Vedejs, E.; Daugulis, O. LatV. Kim. Z. 1999, 1, 31.
(5) Vedejs, E.; MacKay, J. A. Org. Lett. 2001, 3, 535.
10.1021/jo0519155 CCC: $33.50 © 2006 American Chemical Society
Published on Web 12/22/2005
498
J. Org. Chem. 2006, 71, 498-503

Synthesis of New Chiral Bicyclic Phospholanes
SCHEME 1
SCHEME 2
the potential of the radical cyclization approach, a synthesis of
the known catalyst 1a was designed from substituted cyclo-
pentene precursors. Previously, the same catalyst had been
obtained by a nucleophilic displacement strategy for ring closure,
but the earlier route encountered difficult purifications at several
stages.3 The radical cyclization approach has the advantage that
the intermediate alcohol 5 is easily prepared and purified.14 In
the key reaction, the known (racemic) mesylate 6 was displaced
with lithium phenylphosphide to give 7, followed by AIBN-
induced radical cyclization (Scheme 1). A mixture of P-epimers
8a and 1a was formed, as evidenced by the isolation of the
corresponding borane complexes 8b and 1b (10:1 ratio, 62%
combined yield) after addition of THF-borane. The major
product proved to be the exo-phenyl stereoisomer 8a having
the undesired, less reactive phosphorus configuration.15 The
same diastereomers 8b and 1b had been obtained earlier from
a bismesylate precursor and P-lithiophenylphosphine,3 with 8b
favored by a ratio of 6.7:1.
Following a literature precedent for the oxidative inversion
of configuration at phosphorus,4,16 the phosphine 8a was treated
with iodine in aqueous acetonitrile to give the phosphine oxide
9. Next, 9 was deoxygenated with retention of configuration
using phenylsilane.17 Temporary protection of the resulting
phosphine 1a as the borane adduct afforded 1b, corresponding
to the more reactive phosphorus configuration as established
in our prior study using the bismesylate displacement route.3
By comparison, the new sequence allows easier product isolation
and scaleup. Furthermore, the feasibility of radical cyclization
for preparation of the phosphabicyclo[3.3.0]octane core is
demonstrated.
With the radical cyclization strategy established, a concise
synthesis of a benzo-fused analogue, 14a, was envisioned using
a similar approach, as shown in Scheme 2. Indenyllithium (10)
was generated by lithiation of indene with n-butyllithium,
followed by the sequential addition of the cyclic sulfate 1118 at
-78 °C followed by 1.5 equiv of LiPHPh to effect formation
of the C-P bond. The resulting secondary phosphine intermedi-
ate was not isolated, but was treated with AIBN (benzene,
reflux) to induce radical cyclization. Subsequent addition of
borane-THF afforded a 4:1 ratio of P-epimers 13b and 14b,
isolated in 44% combined yield after chromatographic separa-
tion. The phosphine obtained from the major borane adduct by
decomplexation with pyrrolidine did not activate isobutyric
anhydride for acyl transfer at room temperature, while the minor
phosphine diastereomer was reactive. In all of the other bicyclic
phosphines studied to date, the more reactive diastereomer has
proven to be the endo-phenyl isomer, having the more accessible
unshared electron pair at phosphorus as required for nucleophilic
catalysis. By that argument, the major radical cyclization product
was assigned the stereochemistry as shown in 13a, while the
minor product was identified as the desired phosphine 14a. This
assignment is also consistent with the stereochemical outcome
of the radical cyclization from 7 to 8a described in Scheme 1.
The major phosphine-borane 13b from radical cyclization
was subjected to the same procedure for inversion of phosphorus
configuration as used previously. The sequence of decomplex-
ation with pyrrolidine, oxidation of 13a to 15 with aqueous
iodine, and deoxygenation with phenylsilane converted 13b into
14b in 78% yield.4,16,17 Finally, resolution of enantiomers by
HPLC was used to obtain enantioenriched 14a to test enanti-
oselectivity in kinetic resolutions. Considering the one-pot nature
(12) Robertson, A.; Bradaric, C.; Frampton, C. S.; McNulty, J.; Capretta,
A. Tetrahedron Lett. 2001, 42, 2609.
(13) For a related nonradical approach for the formation of five-
membered cyclic phosphines, see: Douglas, M. R.; Stern, C. L.; Marks, T.
J. J. Am. Chem. Soc. 2001, 123, 10221. Kawaoka, A. M.; Douglass, M. R.;
Marks, T. J. Organometallics 2003, 22, 4630.
(14) Balme, G.; Bouyssi, D.; Faure, R.; Gore, J.; Vanhemelryck, B.
Tetrahedron 1992, 48, 3891.
(15) (a) We have not determined whether 8a is formed under kinetic or
thermodynamic control. Formation of the more stable phosphorus config-
uration 8a would be no surprise in a kinetically controlled cyclization.
However, reversible phosphinyl radical additions are precedented15b and
would also be expected to lead to 8a. (b) Pellon, J. J. Am. Chem. Soc.
1961, 83, 1915.
(16) Horner, L.; Winkler, H. Tetrahedron Lett. 1964, 455.
(17) Marsi, K. L. J. Org. Chem. 1974, 39, 265.
(18) Fries, G.; Wolf, J.; Pfeiffer, M.; Stalke, D.; Werner, H. Angew.
Chem., Int. Ed. 2000, 39, 564.
J. Org. Chem, Vol. 71, No. 2, 2006 499

MacKay and Vedejs
TABLE 1. Isobutyroylation of 20 with Chiral Phosphines 14a, 18a,
and 2a
SCHEME 3
cat.
(loading, %)
time
(h)
conv
(%)
rel
rate
entry
temp
rt
rt
rt
s
1
2
3
4
14a (8)
2a (19)
18a (2.6)
18a (2.5)
112
6
20.5
93
41
44
29
63
1
1.5
9
5
5.5
6.6
1.7
-25 °C
6
TABLE 2. Benzoylations with 18a
of the conversion from 10 to 13b + 14b, this sequence is short
and relatively convenient compared to other known synthetic
approaches to racemic phosphabicyclo[3.3.0]octanes, despite the
need to invert the phosphorus configuration.
entry
alcohol
Ar
R
cat.
temp
rt
rt
rt
35 °C
35 °C
-25 °C
-25 °C
rt
rt
rt
rt
s
1
2
3
4
5
6
7
8
9
22
22
20
20
20
20
20
23
23
24
24
Ph
Ph
tBu
tBu
Me
Me
Me
Me
Me
nBu
nBu
iPr
18a
2a
18a
18a
2a
18a
2a
18a
3a
23
14a
12
1-naphthyl
1-naphthyl
1-naphthyl
1-naphthyl
1-naphthyl
Ph
10
An alternative sequence to related structures containing the
phosphabicyclo[3.3.0]octane core was considered, following
reports by Dzhemilev19 and Walther20 detailing the carbometa-
lation of norbornene using zirconocene dichloride together with
triethylalane or ethylmagnesium chloride. The triethylalane
procedure is catalytic in zirconium and affords the metallacycle
16, while the EtMgCl method uses a stoichiometric amount of
the zirconocene reagent and produces 17. Quenching either 16
or 17 with phenyldichlorophosphine afforded diastereomeric
mixtures of the norbornane-fused phospholanes that were
isolated as the corresponding borane complexes 18b and 19b
(Scheme 3). Interestingly, the diastereomer ratios obtained from
16 and 17 were different and complementary. The zirconacycle
17 gave the desired endo-phenyl epimer 18b as the major
product in a 3:1 ratio and in variable yield (15-35% combined
for both isomers). On the other hand, the cyclic alane 16
obtained using catalytic Cp2ZrCl2 consistently gave ca. 50% of
cyclic phosphine products favoring 19b. In either case, the
purified undesired exo-phenyl epimer 19b could be inverted in
high yield using the standard sequence of deprotection, oxidative
inversion at phosphorus (I2/H2O), deoxygenation of PdO with
phenylsilane, and reprotection to afford 18b.4,16,17 Separation
of enantiomers was achieved by HPLC as before, and depro-
tection of the borane complex by warming with pyrrolidine
afforded the free phosphine 18a as needed for testing enantio-
selectivity in acyl-transfer reactions.
10a
19
28a
10
Ph
Ph
Ph
11b
10
10
11
18a
3a
iPr
15b
a Ref 3 data. b Ref 2 data.
somewhat lower, and little improvement in selectivity was
observed at lower temperatures (entry 4).
More promising results were obtained when the more reactive
catalyst 18a was used for the activation of benzoic anhydride
(Table 2). A ca. 2-fold rate increase was observed in the
benzoylation of 22 using 18a compared to 2a, and selectivity
was improved (entry 1, s ) 23 for 18a; entry 2, s ) 14 for 2a).
With the less hindered alcohol 20, no significant difference
between 18a and 2a was observed from room temperature to
35 °C (entries 3-5), although a smaller improvement in
enantioselectivity was found at -25 °C for 18a compared to
2a (entries 6 and 7). Similar enantioselectivities were also
determined when substrates 23 and 24 were benzoylated using
catalyst 18a. Marginally better results had been observed earlier
using 3a as the catalyst (entries 9 and 11 vs entries 8 and 10),
but 18a is a viable catalyst overall in terms of reactivity as well
as enantioselectivity.
The origins of reactivity and selectivity for PBO catalysts
are not well understood, but a possible correlation between
catalyst reactivity and ground-state conformational preferences
has been noted.3 Simple P-phenylphospholanes are relatively
unreactive, apparently because they prefer a geometry having
the lone pair (lp) eclipsed by the P-Ph group (Figure 1, 26;
dihedral angle θ near 0°) over alternatives such as structure 25
where P-Ph eclipses one of the ring CH2 groups (θ near 60°).
The latter geometry has the less hindered unshared electron pair
required for effective nucleophilic catalysis, but the increased
The new catalysts 14a and 18a were compared with the PBO
catalyst 2a in the kinetic resolution of 1-(1-naphthyl)ethanol
(20), the standard benzylic alcohol test substrate, following
procedures developed earlier. Catalyst 14a proved to have
modest reactivity (5-6-fold lower compared to that of 2a) and
poor enantioselectivity in the isobutyroylation of 20 (Table 1,
entry 1, s ) 1.5, vs entry 2). In contrast, the reactivity of catalyst
18a was comparable to that of 2a if the relative rates were
adjusted for the difference in catalyst loading (entry 2 vs entry
3). However, the enantioselectivity observed with 18a was
(19) Dzhemilev, U. M.; Abragimov, A. G.; Zolotarev, A. P.; Khalilov,
L. M.; Muslukhov, R. R. Bull. Russ. Acad. Sci., DiV. Chem. Sci. 1992, 41,
300.
(20) Fischer, R.; Walther, D.; Gebhardt, P.; Gorls, H. Organometallics
2000, 19, 2532.
500 J. Org. Chem., Vol. 71, No. 2, 2006

Synthesis of New Chiral Bicyclic Phospholanes
having substantial P-phenyl dihedral angles θ relative to the
unshared electron pair at phosphorus in 2a, or the BH3 group
in 2b.
The same computational approach (HF/3-21G* level) and
search for local energy minima in 3° increments3 was used to
evaluate conformational preferences for borane complexes 14b
and 18b corresponding to the new catalysts 14a and 18a. In
the case of the ineffective benzo-fused catalyst 14a, two energy
minima were obtained (Figure 2). The more stable conformer
27 has a relatively small dihedral angle, θ ) -9 ( 3°, between
the P-Ph group and the P-B bond, a result that correlates with
lower reactivity due to increased steric hindrance near the
unshared electron pair of the catalyst 14a. The less stable (by
2 kcal/mol) conformer 28 has a larger dihedral angle, θ ) -20
( 3°. If the dihedral angle/reactivity correlation holds, then the
conformer of 14a that resembles 28 would be more reactive
than a geometry similar to 27. However, the less reactive
phosphine geometry resembling 27 would be dominant at
equilibrium because it is more stable.
FIGURE 1. CH2‚‚‚P-Ph eclipsed (25) and lp‚‚‚P-Ph eclipsed (26)
conformers of 1-phenylphospholane.
Computational modeling of 18a by the same approach
revealed two local minima, 29 and 30 (Figure 3). Conformer
30 has the lower calculated energy by 1 kcal/mol and the larger
dihedral angle (θ ) -40 ( 3°), consistent with high catalytic
reactivity for the corresponding conformer of catalyst 18a. The
phospholane ring geometry of 30 is virtually identical to that
found in one of the energy minima for 2a/2b (conformer 31;
θ ) -42 ( 3°) consistent with the proposition that the high
reactivity of 2a as well as 18a may be due to access to similar
transition-state geometries for acyl transfer. Because of the fused
norbornane subunit, structure 18a is far more rigid than is 2a
and has relatively few conformational options. Therefore, the
similar reactivities and good enantioselectivities of 18a and 2a
can be taken as evidence that the preferred transition-state
geometries for acyl transfer resemble conformers 30 and 31,
respectively. According to this analysis, reactive catalysts should
have low-energy conformers with θ in the range of 35-45° to
provide a less hindered environment near the unshared electron
pair at phosphorus. This conclusion may be helpful in the design
of new conformationally constrained catalyst families containing
the phosphole ring and having high nucleophilic reactivity.
FIGURE 2. Energy minima (27, 28) for 14b (3-21G*; selected H
atoms erased for clarity).
Ph‚‚‚CH2 interaction destabilizes the transition state for acyl
transfer. However, geometries intermediate between the limiting
cases 25 and 26 should have improved reactivity and fewer steric
repulsions. This conformational argument was supported by
computational modeling (HF/3-21G*) of ground-state geom-
etries for monocyclic and bicyclic phosphines.3 Borane com-
plexes of the catalytically active phosphines were also modeled,
and were used as deliberately oversimplified surrogates for the
transition states corresponding to acyl transfer. Of course, the
phosphine-boranes were not expected to closely mimic the
transition states for acyl transfer, but they did provide an
opportunity to develop a confidence level in the HF/3-21G*
geometries by comparison with three solid-state conformers
observed in the X-ray crystal structure of the borane complex
2b, and could be modeled without encountering unknowns that
arise in any detailed description of the acyl-transfer process.
Using this indirect approach to probe conformational prefer-
ences, a tentative correlation was made between the high
reactivity of catalyst 2a and access to low-energy conformers
Summary
As the first member of a new family of chiral phosphabi-
cyclooctanes, 18a is noteworthy because it has good reactivity
FIGURE 3. Geometries of energy minima (3-21G*) for 18b (29, 30) and 2b (31).
J. Org. Chem, Vol. 71, No. 2, 2006 501

MacKay and Vedejs
(14b). To a flame-dried, N2-purged flask were added indene (0.59
mL, 5 mmol; freshly distilled) and THF (5 mL). The solution was
cooled to -78 °C, and nBuLi (3.8 mL of a 1.44 M solution in
hexanes, 5.5 mmol) was added dropwise. After the resulting solution
was stirred at -78 °C for 5 min, the cooling bath was removed
and the reaction was allowed to stir at room temperature for 20
min. The resulting solution was added to a solution of 1,3,2-
dioxathiolane 2,2-dioxide (11) (620 mg, 5.0 mmol) in THF (8 mL)
at -78 °C via cannula. Stirring was continued for 2 h before the
solution was warmed to room temperature. In a separate flask,
lithiophenylphosphide was prepared by addition of nBuLi (5.6 mL
of a 1.44 M solution in hexanes, 8.0 mmol) to phenylphosphine
(0.75 mL, 7.5 mmol) in toluene (30 mL) at 0 °C followed by stirring
for 30 min. The reaction mixture was recooled to -78 °C, and to
it was added the lithiophosphide via cannula. After the addition,
the mixture was stirred at -78 °C for 3 h and then warmed to
room temperature. Benzene (50 mL, degassed) was added, a short
path condenser was attached, and THF was removed by distillation
as well as enantioselectivity in the kinetic resolution of alcohols
by benzoylation.21 The synthesis of 18a via the metallacycle
16 is short compared to the nucleophilic displacement or radical
cyclization routes used to prepare 1a. The radical cyclization
approach to 14a is also short, but 14a has poor enantioselectivity
and reactivity. Conformational modeling studies support the
correlation between catalyst reactivity and P-phenyl dihedral
angle preferences. Future efforts to develop chiral phospholane
acylation catalysts will need to keep these geometrical prefer-
ences in mind.
Experimental Section
exo-2-(Phenylphospha)bicyclo[3.3.0]octane-Borane Complex
(8b) and endo-2-(Phenylphospha)bicyclo[3.3.0.]octane-Borane
Complex (1b). To a flame-dried, N2-purged flask were added
phenylphosphine (0.42 mL, 4.18 mmol; caution! stench!) and
toluene (16 mL). The solution was cooled to 0 °C, and nBuLi (2.8
mL of a 1.54 M solution in hexanes, 4.3 mmol) was added
dropwise. A solution of mesylate 614 in toluene (16 mL) was added,
and the solution was warmed to room temperature and stirred for
1 h. Assay by 31P NMR (unlocked, crude reaction mixture) showed
the presence of a secondary phosphine (δP ) -51 ppm). A reflux
condenser was attached, and the crude solution was heated to 85
°C with the slow addition of AIBN (161 mg, 0.98 mmol) in toluene
(10 mL) via syringe pump over 4 h. Stirring was continued for 8
h. Assay by 31P NMR (unlocked, crude reaction mixture) showed
the presence of two tertiary phosphines in a 10:1 ratio (δP ) 6.9
and -1.5 ppm). After the addition of borane-THF complex (4.6
mL of a 1 M solution in THF), the colorless solution was stirred
for 1 h, the solvent was evaporated (N2 stream), and HCl (25 mL,
10% solution in water) was added. Next, the white residue was
extracted with CH2Cl2 (3 × 30 mL), dried (MgSO4), and evaporated
(aspirator), and the residue was purified by flash chromatography
on silica gel (16 × 4 cm), 4:1 hexanes/toluene f 1:1 hexanes/
toluene eluent.
until the temperature of the distillate reached 80 °C. Assay by 31
P
NMR (unlocked, crude reaction mixture) showed the presence of
a secondary phosphine (δP ) -51 ppm). A reflux condenser was
attached, and the crude solution was heated to reflux with the slow
addition of AIBN (200 mg, 0.24 mmol) in benzene (40 mL) via a
syringe pump over 6 h. Stirring continued for 8 h. Assay by 31P
NMR (unlocked, crude reaction mixture) showed the presence of
a 4:1 ratio of two tertiary phosphines (δP ) 13.2 and -4.7 ppm).
After the addition of borane-THF complex (9.0 mL of a 1 M
solution in THF), the colorless solution was stirred for 1 h, the
solvent was evaporated (N2 stream), and HCl (25 mL, 10% solution
in water) was added. Next, the white residue was extracted with
CH2Cl2 (3 × 30 mL), dried (MgSO4), and evaporated (aspirator),
and the residue was purified by flash chromatography on silica gel
(16 × 4 cm), 2:1 hexanes/toluene eluent. The first 400 mL was
blank, then the next 550 mL contained impurities, and the next 50
mL contained the minor diastereomer 14b. The eluent was switched
to 1:1 hexanes/toluene, and the elution was continued, collecting a
mixture of the two P-epimers in the next 100 mL, followed by
pure 13b, the major diastereomer, in the next 575 mL.
Fractions containing pure 14b (endo-isomer) were evaporated
to give 68 mg (5%) of colorless crystals. Analytical TLC, 1:1
hexanes/toluene, Rf ) 0.25. Pure material was obtained by
crystallization from ethanol, mp 75-76 °C. Separation into enan-
tiomers was effected using a CHIRALCELL OJ analytical HPLC
column (10% ethanol/hexanes, 1 mL/min), retention times of
enantiomers, 8.9 min, 14.0 min. ESMS: C17H20BP (M - 1), m/z
) 265.1311, base peak 252.1 amu. IR (neat, cm-1): 2370, B-H;
2335, B-H. 1H NMR (500 MHz, CDCl3, ppm): δ 7.43 (1 H, dddd,
J ) 7.8, 7.5, 3.0, 1.6 Hz), 7.31-7.26 (4 H, m), 7.20-7.16 (3 H,
m), 6.91 (1 H, d, J ) 7.6 Hz), 4.11 (1 H, ddd, J ) 20.0, 6.6, 6.6
Hz), 3.18-3.09 (1 H, m), 3.03 (1 H, t, J ) 8.1 Hz), 2.68-2.58 (1
H, m), 2.48-2.34 (2 H, m), 2.12 (1 H, dd, J ) 13.7, 5.9 Hz), 1.93-
1.83 (1 H, m), 1.2-0.4 (3 H, br m). 13C NMR (125.7 MHz, CDCl3,
ppm): δ 143.2, 143.1, 133.1 (d, J ) 9.2 Hz), 131.5, 131.5, 128.3
(d, J ) 9.9 Hz), 127.4 (d, J ) 28.2 Hz), 125.9 (d, J ) 45.8 Hz),
124.6, 123.2, 52.0 (d, J ) 4.6 Hz), 41.4 (d, 33.6 Hz), 34.1 (d, J )
4.6 Hz), 29.2 (d, J ) 6.1 Hz), 23.4 (d, J ) 36.6 Hz). 31P NMR
(161.9 MHz {H},CDCl3, ppm): δ 38.6 (br m).
Fractions containing 8b (exo-isomer) were combined to give 452
mg (53%) of a thick oil which solidified on standing. All physical
properties were identical to previously reported data for this
compound.5 Analytical TLC, 1:1 hexanes: Rf ) 0.19. MS: no
parent ion for C13H20BP; M - BH3, m/z ) 204.1060, error 4 ppm,
1
base peak 204 amu. IR (neat, cm-1): 2373, B-H. H NMR (300
MHz, CDCl3, ppm): δ 7.75-7.65 (2 H, m), 7.49-7.40 (3 H, m),
2.98-2.82 (1 H, m), 2.74-2.61 (1 H, m), 2.21-1.73 (8 H, m),
1.69-1.54 (1 H, m), 1.50-1.36 (1 H, m), 1.32-0.16 (3 H, br m).
31P NMR (21.4 MHz {H}, CDCl3, ppm): δ 25.7-23.6 (br m).
More polar fractions containing 1b (endo-isomer) were evapo-
rated to give 79 mg (9%) of an oil. All physical properties were
identical to previously reported data for this compound.5 Analytical
TLC, 1:1 hexanes/toluene: Rf ) 0.09. Separation into enantiomers
was effected using a CHIRALCELL OJ analytical HPLC column
(11% ethanol/hexanes, 1 mL/min), retention times of enantiomers,
9.3 min, 11.2 min. MS: no parent ion for C13H20BP; M - BH3,
m/z ) 204.1061, error 4 ppm, base peak 204 amu. IR (neat, cm-1):
2369, B-H. 1H NMR (300 MHz, CDCl3, ppm): δ 7.70-7.60 (2
H, m), 7.52-7.41 (3 H, m), 2.95-2.79 (1 H, m), 2.64 (1 H, dq, J
) 9.1, 3.7 Hz), 2.46-2.21 (2 H, m), 2.13-2.00 (1 H, m), 1.92-
1.44 (5 H, m), 1.43-1.29 (1 H, m), 1.4-0.2 (3 H, br m), 1.21-
1.03 (1 H, m). 31P NMR (121.4 MHz {H}, CDCl3, ppm): δ 41.2-
38.9 (br m).
More polar fractions containing pure 13b (exo-isomer) were
combined to yield 520 mg of a colorless solid (39%). Analytical
TLC, 1:1 hexanes/toluene: Rf ) 0.20. Pure material was obtained
by crystallization from ethanol, mp 96.0-96.8 °C. ESMS: C17H20-
BP (M - 2), m/z ) 264.1246, base peak 252.1 amu. IR (neat,
exo-1-Phenyl-1,2,3,3a,8,8a-hexahydro-1-phosphacyclopenta-
[a]indene-Borane Complex (13b) and endo-1-Phenyl-1,2,3,3a,8,-
8a-hexahydro-1-phosphacyclopenta[a]indene-Borane Complex
1
cm-1): 2374, B-H; 2339, B-H. H NMR (500 MHz, CDCl3,
ppm): δ 7.77-7.72 (2 H, m), 7.52-7.46 (3 H, m), 7.28-7.17 (4
H, m), 4.17 (1 H, dddd, J ) 7.6, 7.6, 3.8, 3.8 Hz), 3.67 (1 H, ddd,
J ) 17.6, 17.6, 3.3 Hz), 3.37-3.28 (1 H, m), 3.20 (1 H, dddd, J )
8.1, 8.1, 8.1, 3.3 Hz), 2.49-2.40 (1 H, m), 2.38-2.28 (1 H, m),
2.20-2.14 (1 H, m), 1.84-1.76 (1 H, m), 1.1-0.2 (3 H, br m).
13C NMR (100.57 MHz, CDCl3, ppm): δ 143.2 (d, J ) 29.0 Hz),
(21) For recent reviews on the kinetic resolution of alcohols using chiral
nucleophilic catalysts, see: (a) Fu, G. C. Acc. Chem Res. 2000, 33, 412.
(b) Jarvo, E. R.; Miller, S. J. Tetrahedron 2002, 58, 2481. (c) Vedejs, E.;
Jure, M. Angew. Chem., Int. Ed. 2005, 44, 3974.
502 J. Org. Chem., Vol. 71, No. 2, 2006

Synthesis of New Chiral Bicyclic Phospholanes
131.5, 131.5 (d, J ) 47.3 Hz), 131.4, 131.2, 129.2 (d, J ) 10.7
Hz), 127.1, 127.8, 125.0, 53.0, 41.3 (d, J ) 33.6 Hz), 34.1 (d, J )
4.6 Hz), 31.9, 26.6. 31P NMR (161.9 MHz {H}, CDCl3, ppm): δ
43.6 (br m).
by crystallization from hexane, mp 79-81 °C. Analytical TLC,
1:1 ether/toluene: Rf ) 0.17. Separation into enantiomers was
effected using a CHIRALPAK AS semiprep HPLC column (3%
ethanol/hexanes, 4 mL/min), retention times of enantiomers (not
baseline separation), 13.2 min, 13.9 min. Analytical HPLC,
CHIRALCEL OJ (10% 2-propanol/hexanes, 1 mL/min): retention
times of enantiomers, 9.5 min, 11.4 min. HRMS: C14H22BP (M -
BH3), m/z ) 230.1221, base peak 230 amu. IR (neat, cm-1): 2374,
B-H; 2351, B-H; 2328, B-H. 1H NMR (500 MHz, CDCl3,
ppm): δ 7.65-7.59 (2 H, m), 7.52-7.42 (3 H, m), 2.52-2.31 (3
H, m), 2.24-2.16 (1 H, m), 2.11-2.03 (1 H, br s), 1.94-1.83 (1
H, m), 1.69-1.58 (1 H, m), 1.55-1.40 (2 H, m), 1.3-0.4 (3 H, br
m), 1.24-1.13 (2 H, m), 0.76-0.67 (2 H, m). 13C NMR (100.6
MHz {H}, CDCl3, ppm): δ 131.7 (d, J ) 8.4 Hz), 130.8 (d, J )
2.3 Hz), 128.6 (d, J ) 9.9 Hz), 127.1 (d, J ) 43.5 Hz), 51.0, 47.2
(d, J ) 31.3 Hz), 40.9, 37.9 (d, J ) 3.8 Hz), 34.0, 32.2, 30.2 (d,
J ) 11.4 Hz), 28.2, 23.9 (d, J ) 37.4 Hz). 31P NMR (161.9 MHz
{H}, CDCl3, ppm): δ 45.6 (br m).
An alternative preparation of 18b and 19b was a modification
of a literature procedure:20 To a flame-dried, Ar-flushed round-
bottom flask was added Cp2ZrCl2 (250 mg, 0.86 mmol). A glass
stopcock was attached and wrapped with Teflon tape, and the flask
was flushed with Ar for 5 min. Next, THF (2 mL) was added, and
the suspension was cooled to -78 °C. Ethylmagnesium chloride
(0.84 mL, 2.0 M in THF, 1.67 mmol) was added dropwise via
syringe, and the mixture was stirred for 1 h, during which the
solution turned pale yellow. Norbornylene (81 mg, 0.855 mmol;
sublimed) was added as a solution in 0.5 mL of THF. The mixture
was warmed to room temperature and stirred for 13 h, during which
the solution turned red. The solution was cooled to -78 °C, and
dichlorophenylphosphine (0.23 mL, 1.71 mmol; freshly distilled)
was added over 5 min via syringe. The reaction was stirred for 11
h, during which the cooling bath warmed to room temperature.
Borane-THF (3 mL, 1 M solution in THF, 3 mmol) was added
via syringe, the pale yellow mixture was stirred for 5 h at room
temperature, and then the solvent was removed (N2 stream). Careful
addition of water (5 mL) quenched the excess BH3, and the mixture
was extracted with Et2O (3 × 10 mL). The organic layer was dried
(MgSO4) and concentrated (aspirator). Assay of the crude mixture
by 31P NMR showed a ratio of ca. 3:1 18b:19b. The residue was
adsorbed on silica and purified as above to afford ca. 50 mg of
18b (24%) and 25-30 mg of impure material containing 19b. All
spectroscopic properties were identical to those reported in the
previous synthesis of 18b and 19b.
exo-3-Phenyl-3-phosphatricyclo[5.2.1.02,6]decane-Borane Com-
plex (19b) and endo-3-Phenyl-3-phosphatricyclo[5.2.1.02,6]-
decane-Borane Complex (18b). The cyclic alane intermediate
16 was prepared according to a literature procedure.19 To a flame-
dried, Ar-flushed round-bottom flask were added Cp2ZrCl2 (50 mg,
0.17 mmol), norbornylene (250 mg, 2.66 mmol; sublimed), and a
magnetic stir bar. A glass stopcock was attached and wrapped with
Teflon tape, and the flask was flushed with argon for 5 min (note:
norbornylene is volatile, and flushing with argon for too long will
evaporate the reagent). The flask was cooled to 0 °C, and neat
triethylaluminum (0.39 mL, 2.8 mmol; caution! highly pyrophoric!)
was added via syringe. The flask was flushed briefly with argon,
and the stopcock was closed. The reaction was stirred for 11 h,
during which the cooling bath warmed to room temperature.
Benzene (2 mL) was added via syringe to the yellow mixture, and
the reaction was recooled to 0 °C. Dichlorophenylphosphine (0.79
mL, 5.85 mmol; freshly distilled) was then added dropwise over
ca. 8 min via syringe. The cooling bath was removed, and the
mixture was allowed to stir for 24 h. The mixture was cooled to 0
°C, and borane-THF (7 mL, 1 M solution in THF, 7 mmol) was
added via syringe to quench the reaction. The mixture was stirred
for 5 h at room temperature, the solvent was evaporated (N2 stream),
and HCl (10 mL of a 5% solution in water; caution! foaming!) and
Et2O (20 mL) were added, followed by stirring for 15 min. The
mixture was extracted with Et2O (3 × 20 mL). The extracts were
combined and dried (MgSO4). After removal of the solvent
(aspirator), the residue was adsorbed on silica and purified by flash
chromatography (15 × 5 cm), eluting with 600 mL of 3:1 hexanes/
toluene and then switching to 2:1 hexanes/toluene. The first 1350
mL was discarded followed by 500 mL of eluent containing the
major diastereomer 19b, 500 mL containing a ca. 1:1 mixture of
diastereomers, and then 1 L containing the minor diastereomer 18b.
The mixed fractions were resubmitted to flash chromatography.
Fractions containing the major diastereomer (exo-isomer) were
combined to give 246 mg (37%) of 19b. Pure material was obtained
by crystallization from hexane, mp 62.0-62.8 °C. Analytical TLC,
1:1 ether/toluene: Rf ) 0.24; HRMS: C14H22BP (M - BH3 + H),
m/z ) 231.1309, base peak 231 amu. IR (neat, cm-1): 2370, B-H;
1
2359, B-H; 2327, B-H. H NMR (500 MHz, CDCl3, ppm): δ
7.74-7.69 (2 H, m), 7.50-7.42 (3 H, m), 2.68-2.64 (1 H, m),
2.44-2.36 (1 H, m), 2.35-2.22 (1 H, m), 2.19-2.12 (2 H, m),
2.10-2.03 (1 H, m), 2.01-1.90 (1 H, m), 1.82-1.72 (1 H, m),
1.71-1.67 (1 H, m), 1.63-1.52 (2 H, m), 1.24-1.10 (3 H, m),
1.1-0.4 (3 H, br m). 13C NMR (125.7 MHz, CDCl3, ppm): δ 131.7
(d, J ) 47.6 Hz), 131.5 (d, J ) 8.7 Hz), 130.9 (d, J ) 2.3 Hz),
128.7 (d, J ) 9.6 Hz), 51.1 (d, J ) 1.4 Hz), 47.2 (d, J ) 53 Hz),
41.7, 38.0 (d, J ) 5.0 Hz), 35.3, 31.1 (d, J ) 6.9 Hz), 30.2 (d, J
) 12.4 Hz), 28.5, 27.4 (d, J ) 36.6 Hz). 31P NMR (161.9 MHz
{H}, CDCl3, ppm): δ 39.8 (m).
Acknowledgment. This work was supported by the National
Science Foundation.
Supporting Information Available: General procedures for
catalyst manipulations, acylations, and assay and spectra of new
compounds. This material is available free of charge via the Internet
at http://pubs.acs.org.
More polar fractions containing 18b (endo-isomer) were com-
bined to give 87 mg (13%) of product. Pure material was obtained
JO0519155
J. Org. Chem, Vol. 71, No. 2, 2006 503