The polyhomologation of 1-boraadamantane: Mapping the migration pathways of a propagating macrotricyclic trialkylborane

Article abstract of DOI:10.1021/ja0361291

Trialkyl and triaryl organoboranes undergo multiple, repetitive homologations upon reaction with dimethylsulfoxonium methylide (1). This multiple homologation reaction, or polyhomologation, produces polymethylene in a living reaction. Applying the polyhom

Full text of DOI:10.1021/ja0361291

Published on Web 09/17/2003
The Polyhomologation of 1-Boraadamantane: Mapping the
Migration Pathways of a Propagating Macrotricyclic
Trialkylborane
Carl E. Wagner,† Jang-Seob Kim,‡ and Kenneth J. Shea*,†
Contribution from the Department of Chemistry, UniVersity of California IrVine,
IrVine, California 92697, and Department of Chemistry, Daegu UniVersity,
Gyungsan, Gyunbuk 712-714, South Korea
Received May 14, 2003; E-mail: kjshea@uci.edu
Abstract: Trialkyl and triaryl organoboranes undergo multiple, repetitive homologations upon reaction with
dimethylsulfoxonium methylide (1). This multiple homologation reaction, or polyhomologation, produces
polymethylene in a living reaction. Applying the polyhomologation reaction to cyclic and polycyclic
organoboranes permits the construction of unique oligomeric and polymeric architectures that are not readily
accessible by standard olefin polymerization. The polyhomologation of 1-boraadamantane‚THF (2) by ylide
1 generates novel macrotricyclic trialkylboranes (3). The oxidation of these macrocyclic organoboranes
generates a three-armed star polymethylene polymer (4) incorporating a cis,cis-1,3,5-trisubstituted
cyclohexane core. Interestingly, only one-third of the initiators lead to product formation, resulting in an
observed degree of polymerization 3 times higher than expected. Close examination of the initial stages of
polymerization show that 1-boraadamantane‚THF reacts with 1 equiv of 1 to afford a monohomologated
product. Subsequent homologations were found to contain branch points leading to isomeric tricyclic products
after the third, fourth, and fifth methylene insertions. At these stages of homologation, all of the propagating
species result in tricyclic trialkylborane cages with collapsed, inverted pyramidal boron centers that are
substantially less reactive toward ylide. Approximately two-thirds of the species discontinue polymerization
at these stages. However, one-third of these species continue to propagate and eventually result in the
formation of giant macrotricyclic polymers of narrow polydispersity. Molecular modeling and kinetic simulation
have aided in the analysis of the probable pathways through which the reaction proceeds.
Introduction
methylsulfoxonium methylide (1) by trialkylboranes. In this
reaction, the alkyl groups of the trialkylboranes undergo
Hydrocarbon polymers constitute one of the world’s leading
sources of synthetically derived materials.1 An increasing
demand for high-performance polymers has necessitated the
development of new catalysts, monomers, and polymerization
techniques that permit greater control over factors that influence
the physical properties of the polymers.2 Despite advances in
the development of efficient catalysts, there are still many
synthetic challenges remaining to be solved. Studies in our
laboratory have concentrated on developing a new living
polymethylene synthesis based on the polymerization of di-
repetitive 1,2-migrations with ylide monomers to build the
carbon backbone of the developing polymethylene polymers one
carbon at a time. This process of repetitive 1,2-migration, or
homologation, has been termed polyhomologation.3 Indeed, the
polyhomologation reaction has demonstrated its versatility in
the construction of novel materials ranging from copolymers
of polymethylene and poly(dimethylsiloxane)s4 to the com-
mercially important commodity copolymer poly(ethylene-b-
styrene).5 By developing and combining new monomer sources,
the polyhomologation reaction has also yielded a route to a
material with the same chemical composition as an ethylene-
propylene copolymer.6
† University of California Irvine.
‡ Daegu University.
Additionally, the polyhomologation of cyclic and bicyclic
organoboranes has demonstrated routes to oligomeric and
polymeric boracyclanes and carbocyclics.7 Applying the poly-
homologation reaction to polycyclic organoboranes offers an
opportunity for the construction of novel topologies such as
(1) (a) Barghoorn, P.; Stebani, U.; Balsam, M. AdV. Mater. 1998, 10, 635. (b)
Morse, P. M. Chem. Eng. News 1999, 77, 11. (c) Younkin, T. R.; Connor,
E. F.; Henderson, J. I.; Friedrich, S. K.; Grubbs, R. H.; Bansleben, D. A.
Science 2000, 287, 460. (d) Wilson, E. Chem. Eng. News 2000, 78, 14.
(2) (a) Brintzinger, H. H.; Fisher, D.; Mu¨elhaupt, R.; Rieger, B.; Waymouth,
R. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143. (b) Fink, G., Mu¨elhaupt,
R., Brintzinger, H. H., Eds. ts: Recent Scientific InnoVations and
Technological ImproVement; Springer-Verlag: Berlin, 1995. (c) Coates,
G. W.; Waymouth, R. M. ComprehensiVe Organometallic Chemistry II;
Pergamon Press: Elmsford, NY, 1995; Vol. 12, p 1193. (d) Britovsek, G.
J. P.; Givson, V. C.; Wass, D. F. Angew. Chem., Int. Ed. 1999, 38, 428.
(e) Coates, G. W. Chem. ReV. 2000, 100, 1223. (f) Ittel, S. D.; Johnson, L.
K.; Brookhart, M. Chem. ReV. 2000, 100, 1169. (g)Mark, T. J.; Chen, E.
Y. Chem. ReV. 2000, 100, 1391. (h) Matyjaszeweski, K.; Xia, J. Chem.
ReV. 2001, 101, 2921.
(3) Shea, K. J.; Walker, J. W.; Zhu, H.; Paz, M.; Greaves, J. J. Am. Chem.
Soc. 1997, 119, 9049.
(4) Shea, K. J.; Staiger, C. L.; Lee, S. Y. Macromolecules 1999, 32, 3157.
(5) Zhou, X.-Z.; Shea, K. J. Macromolecules 2001, 34, 3114.
(6) Zhou, X.-Z.; Shea, K. J. J. Am. Chem. Soc. 2000, 122, 11515.
(7) (a) Shea, K. J.; Lee, S. Y.; Busch, B. B. J. Org. Chem. 1998, 63, 5746. (b)
Shea, K. J.; Busch, B. B.; Paz, M. Angew. Chem., Int. Ed. 1998, 37, 1391.
9
10.1021/ja0361291 CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 12179-12195
12179

A R T I C L E S
Wagner et al.
Scheme 1
Scheme 2
macrotricyclic organoboranes capped with a Lewis acid center.
These structures may be elaborated to hydrocarbon star polymer
architectures.8 1-Boraadamantane‚THF (2)9 was selected for
study as a tricyclic alkylborane initiator to access a three-armed
star polymethylene polymer, with the arms radiating from a
cyclohexane core. In this strategy, trialkylborane 2 would serve
as an initiator, while dimethylsulfoxonium methylide (1) would
serve as a methylene monomer source. One equivalent of DMSO
would be expelled for each equivalent of ylide 1 consumed.
Random insertion into all three carbon-boron bonds would
result in the formation of a hyperextended macrotricyclic
organoborane (3). The living nature of the polyhomologation
reaction would allow control of the length of each wall of 3.10
Providing each branch grows randomly, subsequent oxidation
of 3 would provide a regular star polymer 4 (Scheme 1).
The strategy thus outlined was undertaken, and the results
are discussed below.
Free 1-boraadamantane has been reported by Mikhailov and
co-workers to be highly reactive and predisposed to complex
formation with Lewis bases.13 However, an apparent glass
transition occurring at 183 K renders solid 1-boraadamantane
amorphous at room temperature.14 The geometry optimized
structure of free 1-boraadamantane at the RB3LYP/6-311**G
level of theory shows a B-C bond distance of 1.569 Å and
CBC ) 116.4°.15 These values are in accord with the values
of 1.574 Å and 116.3° calculated at the HF/3-21G* level of
theory used throughout this study. The θσπ parameter of the
POAV1 analysis measures the angle between a π-bond orbital
and the σ-orbitals for a distorted sp2 atom in an aromatic
system.16 This parameter was modified to measure the angle
(θσp) between the empty p-orbital on boron and the σ orbitals
to the carbon atoms to assess the degree of pyramidalization
for free 1-boraadamantane. The value of θσp - 90° is a
convenient expression of this parameter, and it ranges from 0°
for a trigonal planar sp2 atom to 19.5° for a tetrahedral sp3
atom.17 The θσp - 90° value for the HF/3-21G* geometry
optimized structure of free 1-boraadamantane is 11.1°, corre-
sponding to a 57% pyramidalization. The geometrically imposed
prehybridization about the boron center of free 1-boraadaman-
tane likely predisposes 1-boraadamantane to readily form
tetrahedral complexes with Lewis basic substrates. Indeed,
1-boraadamantane displays a remarkable capacity to form stable
complexes with Lewis basic substrates such as pyridine18 as
well as weakly Lewis basic substrates such as ether and THF.
The X-ray diffraction study shows that complex 2 adopts a
tetrahedral geometry with B-O ) 1.64 Å, B-C ) 1.61 Å,
O1B1C1 ) O1B1C2 ) 107°, and O1B1C3 ) 111°.
Results and Discussion
Thus, 1-boraadamantane‚THF (2) was synthesized by the
hydroboration of 3-methoxy-7-(methoxymethyl)-3-borabicyclo-
[3.3.1]non-6-ene (5). This compound was prepared by the
reaction of triallylborane (6) with methyl propargyl ether
followed by methanolysis, according to the procedure of
Mikhailov and co-workers (Scheme 2).11 Triallylborane (6) was
synthesized according to the method of Zakharkin and co-
workers.12
Complex 2, a white crystalline solid that slowly degrades in
air, was purified by vacuum sublimation, and its structure was
confirmed by X-ray crystallography (Figure 1).
(8) Wagner, C. E.; Shea, K. J. Org. Lett. 2001, 3, 3063.
(9) (a) Mikhailov, B. M. Pure Appl. Chem. 1980, 52, 691. (b) Mikhailov, B.
M. IzV. Akad. Nauk SSSR, Ser. Khim. 1984, 33, 225.
(10) Shea, K. J. Chem.-Eur. J. 2000, 6, 1113.
(11) (a) Mikhailov, B. M.; Baryshnikova, T. K.; Kiselev, V. G.; Shashkov, A.
S. IzV. Akad. Nauk SSSR, Ser. Khim. 1979, 28, 2544. (b) Mikhailov, B.
M.; Baryshnikova, T. K. IzV. Akad. Nauk SSSR, Ser. Khim. 1979, 28, 2541.
(12) Zakharkin, L. I.; Stanko, V. I. IzV. Akad. Nauk SSSR, Otd. Khim. Nauk
1960, 1896.
(13) (a) Mikhailov, B. M.; Smirnov, V. N. IzV. Akad. Nauk SSSR, Ser. Khim.
1973, 22, 2165. (b) Mikhailov, B. M.; Smirnov, V. N. IzV. Akad. Nauk
SSSR, Ser. Khim. 1974, 23, 1137. (c) Mikhailov, B. M.; Smirnov, V. N.;
Kasparov, V. A. IzV. Akad. Nauk SSSR, Ser. Khim. 1976, 25, 2302.
(14) Bukalov, S. S.; Leites, L. A.; Bubnov, Yu. N.; Gurskii, M. E.; Potapova,
T. V. IzV. Akad. Nauk SSSR, Ser. Khim. 1989, 38, 483.
(15) Wrackmeyer, B.; Wolfgang, M.; Tok, O. L.; Bubnov, Y. N. Chem.-Eur. J.
2002, 8, 1537.
(16) (a) Haddon, R. C.; Scott, L. T. Pure Appl. Chem. 1986, 58, 137. (b) Haddon,
R. C. J. Am. Chem. Soc. 1986, 108, 2837.
(17) Haddon, R. C. J. Am. Chem. Soc. 1990, 112, 3385.
(18) Vorontsova, L. G.; Chizhov, O. S.; Smirnov, V. N.; Mikhailov, B. M. IzV.
Akad. Nauk SSSR, Ser. Khim. 1981, 30, 595.
Figure 1. The X-ray crystal structure of 1-boraadamantane‚THF (2).
9
12180 J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003

The Polyhomologation of 1-Boraadamantane
A R T I C L E S
Figure 2. Representative 1H NMR spectrum of polymeric triol 4 (C6D6, 78 °C, 500 MHz). This spectrum is that of the polymeric product of trial 2 in Table
1. The doublet at 0.9 ppm arises from the incorporation of 1-2% ethylidene groups in the polymethylene backbone arising from dimethylsulfoxonium
ethylide, formed as a minor product in the synthesis of ylide 1.21 The broad singlet at 0.5 ppm has been assigned to the hydroxy-protons of the end groups.
Scheme 3
Table 1. Molecular Weight Data for Star Polymers 4
a
c
DP
M b
(g/mol)
DP
dead catalysts (%)c
1H NMR
N
1H NMR
PDIb
b
trial
(calcd)
DP
1
2
3
4
25
50
69
2403
5595
10 542
19 613
53
129
246
462
73
148
190
345
66
66
63
60
1.06
1.09
1.08
1.08
138
a The DP refers to n (Scheme 1). b Data refer to GPC in o-xylenes at
100 °C calibrated with linear polyethylene standards. c Data refer to 1H
NMR end group analysis (i.e., the ratio of the terminal hydroxymethylene
protons to the broad polymethylene peak of 4).
The homologation reaction involves a two-step process
consisting of addition of the nucleophilic ylide 1 to form an
“ate” complex followed by a [1,2]-migration. Related migrations
of bicyclic organoboranes proceed via an antiperiplanar transi-
tion state.19 To establish the compatibility of the 1-boraada-
mantane core with addition-migration, we first examined its
oxidation. Compound 2 is readily oxidized by trimethylamine
N-oxide dihydrate20 (TAO), to give the triol, cis,cis-1,3,5-tris-
(hydroxymethyl)cyclohexane (7), in 76% yield (Scheme 3).
1-Boraadamantane‚THF (2) was next examined as an initiator
for the polyhomologation reaction. Toluene solutions of ylide
1 were preheated to 80 °C and treated with an aliquot of a
toluene solution of 2. The ylide was rapidly consumed (5 min).
TAO was added directly to the solution to effect oxidation under
reflux conditions. Removal of solvents followed by filtering and
washing of the polymeric solids with methanol, water, and
hexanes gave near quantitative crude yields of polymeric triol
4. Samples were further purified by reprecipitation in toluene/
acetonitrile to afford an opaque white polymer in 88-94% yield.
The structural assignment of the star polymeric triols 4 is as
follows.
ppm (J ) 12.2 Hz) and a quartet at 0.59 ppm (J ) 11.8 Hz)
corresponding to the equatorial and axial methylene protons of
the cyclohexane ring of 4 (Figure 2).
Lightly branched polymers of the type of 4 comprised of
unsymmetrical branches (i.e., two long branches and one short
branch) would not be expected to exhibit apparent three-fold
symmetry of the axial and equatorial hydrogens of the cyclo-
hexane core. The clear three-fold pseudosymmetry of the
equatorial and axial protons, as well as the multiplicity of the
terminal hydroxymethylene groups of 4, support the conclusion
that all three branches of the propagating species participate in
migration.
The degree of polymerization (DP) was established by GPC
and 1H NMR end group analysis (Table 1). Reactions were run
with initial ratios of (1/3)ylide:boraadamantane of 25, 50, 69,
and 138. These reactions should produce regular star polymers
4 with n ) 25, 50, 69, and 138, respectively (Scheme 1). In all
The 1H NMR spectra of the polymer 4 (C6D6, 78 °C)
produced by these reactions have features consistent with a
regular star architecture. The spectra include a triplet at 3.38
ppm (J ) 6.4 Hz) corresponding to the terminal hydroxymeth-
ylene protons of 4. Also, there is an apparent doublet at 1.85
1
cases, however, the DP by H NMR end group analysis was
consistently 3 times greater than the calculated DP.
While the GPC was calibrated with linear polyethylene
standards, corrections for star topology should result in observed
(GPC) molecular weights that are lower than the calculated
molecular weight on the basis of stoichiometry.22 Furthermore,
previous examples of regular three-armed star polymethylene
(19) Soderquist, J. A.; Najafi, M. R. J. Org. Chem. 1986, 51, 1330.
(20) Kabalka, G. W.; Hedgecock, H. C., Jr. J. Org. Chem. 1975, 40, 1776.
9
J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003 12181

A R T I C L E S
Wagner et al.
Table 2. Light Scattering, Viscosity, and GPC Data for High
Molecular Weight Star Polymers 4
carbinol polymers resulted in polymers whose theoretical and
observed molecular weights by GPC corresponded within 10%.10
Importantly, the PDIs of all trials in the present study are <1.09.
Such narrow PDIs are characteristic of living polymerization.
Additionally, two reactions were conducted to give star
polymers 4 in the molecular weight range appropriate to
characterize the nature of the chain branching. The difference
between polymers exhibiting long chain branching and linear
chain polymers is frequently expressed according to the Zimms
Stockmayer model23 in terms of the square of the radius of
gyration of a branched polymer as compared to that of a linear
polymer of the same molecular weight. This ratio, g, is given
below.
a
b
DP
MW
[η]c
(dL/g)
gd
b
trial
(calcd)
(g/mol)
DP
g′ c
(∈ ) 0.5)
PDIe
1
2
296
593
24 270
39 650
573
938
0.567
0.805
0.887
0.878
0.787
0.771
1.12
1.12
a The DP refer to n (Scheme 1). b Data refer to light scattering according
to the technique of deGroot and co-workers.27 c Data refer to viscometry
according to the technique of deGroot and co-workers.27 d Calculated from
the experimental value of g′. e Data refer to GPC in o-xylenes at 100 °C.
intrinsic viscosity by viscometry. The data for the star polymers
4, including the experimental values of g′ and g, are summarized
in Table 2.
The close correlation between the experimental values of g
for the star polymers 4 in Table 2 (0.787 and 0.771) and the
theoretical value of g ) 0.778 for a perfectly regular star
polymer supports the assignment of a regular star architecture
for the star polymers 4. For the reactions in Table 2, ap-
proximately one-half as opposed to one-third of the initiators
led to the formation of star polymers 4 according to the
molecular weight discrepancy. Again, the PDIs of the reactions
to produce high molecular weight star polymers 4 were low
(1.12). Indeed, both 1H NMR and viscometry data for the high
molecular weight star polymers 4 provide strong evidence that
the star polymers (4) produced by the polyhomologation of
1-boraadamantane‚THF (3) are regular star polymers.
Because extreme care was taken to quantitate the stoichio-
metric ratios of catalyst 2 to monomer 1, it appeared that a
significant fraction of propagating species terminated and did
not result in polymer formation. This presented us with a
mechanistic quandary. Following initiation by 2, approximately
two-thirds of the propagating species must effectively terminate
(dead catalysts in Table 1) to account for the molecular weight
discrepancy. The demise of these propagating species must be
rather sudden, because random termination reactions during the
course of polymerization would result in a higher polydispersity.
Additionally, we speculated that the terminated species occurred
at the early stages of polymerization, before a significant fraction
of ylide was consumed. Thus, we directed our efforts at
examining the products and product distributions at the initial
stages of polymerization.
2
Rg
branched
g )
2
Rg
linear
For a three-armed A2B star polymer, the ZimmsStockmayer
model23 gives the following predicted value for g.
6s
(s + 2)3
g
pred ) 1 -
where s ) MA/MB is the ratio of the A arm molecular weight
(MA) to the B arm molecular weight (MB). For a regular star
polymer, MA ) MB, and gpred ) 0.778.
Another common characterization of long chain branching
in star polymers arises from the comparison of the intrinsic
viscosity (η) of a given star polymer to the viscosity of a linear
polymer of the same molecular weight.24 This ratio, g′, is given
below.
[η]branched
g′ )
[η]linear
The relation between g′ and g is expressed in an equation with
the exponent, ∈, as shown below.
g′ ) g∈
For a three-armed regular star polymer, g′ can be approximated
as the square root of g,25 although experimentally it has been
found that ∈ ≈ 0.6 for stars and other branched comb
polymers.26
The technique of triple detector GPC involves a combination
of GPC, light scattering, and viscometry. Triple detector GPC
according to the method of deGroot and co-workers27 was
applied to the high molecular weight star polymers 4 to
determine absolute molecular weights by light scattering and
Controlling the first step in this reaction to achieve mono-
homologation28 was explored by slowly adding 1 equiv of ylide
1 to a solution of 1-boraadamantane‚THF (2) in toluene at -78
°C. The resulting white crystalline precipitate was filtered cold,
washed with hexanes, and dried under vacuum (0.01 Torr, -60
°C). Upon warming to room temperature under vacuum, the
white crystalline solid rapidly melted, exothermically, forming
a colorless liquid. The liquid crystallized upon cooling. These
observations are consistent with the initial formation and
isolation of a solid 1-boraadamantane‚ylide complex (8) that
subsequently undergoes a [1,2]-migration upon warming to room
temperature. The decomposition of complex 8 proceeds under
ambient conditions, and the expelled DMSO is captured to form
1-borahomoadamantane‚DMSO (9) (Scheme 4).
(21) Busch, B. B.; Paz, M. M.; Shea, K. J.; Staiger, C. L.; Stoddard, J. M.;
Walker, J. R.; Zhou, X. Z.; Zhu, H. J. Am. Chem. Soc. 2002, 124, 3636.
(22) Burchard, W. AdV. Polym. Sci. 1999, 143, 113.
(23) (a) Zimm, B. H.; Stockmayer, W. H. J. Chem. Phys. 1949, 17, 1301. (b)
Berry, G. C.; Orofino, T. A. J. Chem. Phys. 1964, 46, 1614.
(24) (a) Drott, E. E.; Mendelson, R. A. J. Polym. Sci., Part A-2 1970, 8, 1361.
(b) Drott, E. E.; Mendelson, R. A. J. Polym. Sci., Part A-2 1970, 8, 1373.
(25) Zimm, B. H.; Kilb, R. W. J. Polym. Sci. 1959, 37, 19.
(26) Hadjichristidis, N.; Xenidou, M.; Iatrou, H.; Pitsikalis, M.; Poulos, Y.;
Avgeropoulos, A.; Sioula, S.; Paraskeva, S.; Velis, G.; Lohse, D. J.; Schulz,
D. N.; Fetters, L. J.; Wright, P. J.; Mendelson, R. A.; Garc´ıa-Franco, C.
A.; Sun, T.; Ruff, C. J. Macromolecules 2000, 33, 2424.
(27) Wood-Adams, P. M.; Dealy, J. M.; deGroot, A. W.; Redwine, O. D.
Macromolecules 2000, 33, 7489.
Complex 9 sublimes (0.01 mmHg, 75 °C) and was isolated
in 65% yield. The structure of 9 was confirmed by X-ray
crystallography (Figure 3).
(28) Stoddard, J. M.; Shea, K. J. Organometallics 2003, 22, 1124.
9
12182 J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003

The Polyhomologation of 1-Boraadamantane
A R T I C L E S
Scheme 4
Scheme 5
Figure 3. The X-ray crystal structure of 1-borahomoadamantane‚DMSO
(9).
There have been earlier efforts to homologate 1-boraada-
mantane. The monohomologation of 1-boraadamantane was pre-
viously attempted with trimethylamine methylide. The primary
product, 1-borahomoadamantane‚trimethylamine, proved dif-
ficult to separate from the nonhomologated 1-boraadamantane‚
trimethylamine formed as a byproduct of the reaction.29 Interest-
ingly, Mikhailov and co-workers previously prepared a 1-bora-
adamantane‚methylenetriphenylphosphorane complex which
was stable to 180 °C.30 However, the monohomologation of
1-boraadamantane by triphenylphosphine methylide could not
be accomplished because of thermal decomposition prior to
migratory insertion.31 Thus, the successful monohomologation
of 1-boraadamantane depends not only on the facility with which
the leaving group departs, but also on the degree to which the
ylide remains bound to boron under a given set of conditions
for [1,2]-migration. The monohomologation of 1-boraadaman-
tane by dimethylsulfoxonium methylide (1) is a rather unique
reaction because it readily transpires from the solid state, and,
most importantly, it results in the controlled monohomologation
of a trialkylborane32 by ylide 1.
example, in an attempt to effect the monohomologation of
1-borahomoadamantane, 1 equiv of ylide 1 was added to a
solution of 1-borahomoadamantane‚DMSO (9) in toluene at -50
°C. The reaction mixture was allowed to warm to room
temperature with stirring. Following consumption of ylide 1,
oxidation of the products was accomplished by refluxing with
TAO. After separation by perparative TLC, triols 10, 13a, and
14a were identified as the products of the reaction, and their
1
structures were determined by H NMR and 13C NMR (D2O).
The relative distribution of the three triols 10, 13a, and 14a
was obtained by GC analysis of the triacetates 12, 13b, and
14b (Scheme 6).
Triol 10 arises from the oxidation of the remaining 1-bora-
homoadamantane 9, and triols 13a and 14a arise from the oxi-
dation of the triorganoboranes 15 and 17, respectively (Figure
4).
Compound 9 is readily oxidized with TAO, providing cis-
1-hydroxyethyl-3,5-bis(hydroxymethyl)cyclohexane (10) in 87%
isolated yield (Scheme 5).
The low molecular weight oxidized products were acetylated
prior to being assayed by GC. Acetylation of triols 7 and 10
provide the acetylated triols 11 and 12 in 83% and 86% yield,
respectively.
In the first homologation of the 1-borahomoadamantane core,
formation of the organoborane species 15 is strongly favored
over the formation of 16. Similarly, in the subsequent homolo-
gation of 15, species 17 is formed exclusively over 18.
The GC spectrum of the triacetates 12, 13b, and 14b from
the reaction depicted in Scheme 6 is shown in Figure 5.
The triacetates 12, 13b, and 14b appear at retention times of
18.5, 19.4, and 20.3 min in 26%, 43%, and 31% composition,
respectively. GC-MS under chemical ionization (CI) conditions
provided the corresponding masses (M + NH4)+ of the
triacetates 12, 13b, and 14b as 332, 346, and 360 m/z,
respectively. The minor peak at the retention time of 19.5 min
has been tentatively assigned to the triacetate formed upon
oxidation and acetylation of the trialkylborane (16), because it
also has a mass (CI) of 346 m/z. Because the side peak at 19.5
min accounts for less than 2% of the total reaction products,
16 is the disfavored product of the second [1,2]-migration.
Triacetates 13b and 14b were separated by column chroma-
With the GC standards 11 and 12 in hand, identification of
the products from multiple homologations was facilitated. For
(29) Mikhailov, B. M.; Govorov, N. N.; Angelyuk, Ya. A.; Kiselev, V. G.;
Struchkova, M. I. IzV. Akad. Nauk SSSR, Ser. Khim. 1980, 29, 1621.
(30) Sergeeva, M. V.; Yanovskii, A. I.; Struchkov, Yu. T.; Mikhailov, B. M.;
Gurskii, M. E.; Pershin, D. G. IzV. Akad. Nauk SSSR, Ser. Khim. 1985, 34,
2483.
(31) Gurskii, M. E.; Pershin, D. G.; Mikhailov, B. M. J. Organomet. Chem.
1984, 260, 17.
(32) (a) Tufariello, J. J.; Lee, L. T. C. J. Am. Chem. Soc. 1966, 88, 4757. (b)
Goddard, J.; Le Gall, T.; Mioskowski, C. Org. Lett. 2000, 2, 1455.
1
tography and were characterized by H NMR and 13C NMR
(CDCl3).
The 1H NMR spectra of triols 7, 10, 13a, and 14a are shown
in Figure 6.
9
J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003 12183

A R T I C L E S
Wagner et al.
Scheme 6
The pseudosymmetry of the equatorial and axial protons of
the triols at 1.80 and 0.60 ppm, respectively, degrades as the
asymmetry of the triols increases. In the 1H NMR spectrum of
polymeric triol 4 (Figure 2), the return of the equatorial and
axial protons to pseudosymmetry and the observed triplet of
the equivalent hydroxymethylene terminal groups are features
consistent with similar polymer arm lengths and the proposed
regular star architecture.
and remaining the 1-borahomoadamantane 9 strongly suggest
that the reaction pathways have substantially different activation
energies with ylide resulting in a path-selective sequence of
homologation. Molecular modeling of the trialkylborane‚ylide
complexes of 9, 15, and 17, and their associated [1,2]-migration
transition state energies leading to isomeric tricyclic organoboron
compounds, provides useful insight into the observed pathway
and will be subsequently examined in detail.
Both the monohomologation of 1-boraadamantane to give
1-borahomoadamantane 9 and the highly selective subsequent
two homologations of 9 reveal a change in the reactivity of the
tricyclic organoborane compounds. Nevertheless, the first three
generations of tricyclic organoboron compounds formed by
homologation of 1-boraadamantane contain active Lewis acidic
centers for polymerization, and, as yet, no evidence for
terminated species was found. Thus, we continued to search
for evidence for termination at the early stages of polymeriza-
tion.
The experimental outcome of the addition of 1 equiv of ylide
1 to 1-borahomoadamantane‚DMSO (9) is remarkable for two
reasons. First, it was anticipated that the analysis of products
resulting from the reaction of 9 with 1 equiv of ylide 1 would
show a distribution of triacetates arising primarily from the
oxidation and acetylation of species 15 and 16. Second, only
minor amounts of the triacetates arising from either the starting
1-borahomoadamantane 9 or species 17, 18, and 19 were
anticipated. The experimental results, however, showed that a
substantial amount of the 1-borahomoadamantane starting
material remained (26%) after the consumption of 1 equiv of
ylide. Further, the homologation of 1-borahomoadamantane
proceeds almost exclusively to form species 15 (41%) over
species 16 (2%), and the subsequent homologation of 15
proceeds exclusively to form organoborane 17 (31%). The
selective formation of 15 and 17 and the distribution of 15, 17,
Insight into the termination of certain intermediates came from
a polymerization (Scheme 1) conducted with an initial (1/3)-
ylide:boraadamantane ratio of 27. After oxidation and repre-
Figure 4. Products from the first three homologations of 1-boraadamantane
by ylide 1.
Figure 5. The GC chromatograph of triacetates 12, 13b, and 14b formed
from the reaction shown in Scheme 6.
9
12184 J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003

The Polyhomologation of 1-Boraadamantane
A R T I C L E S
1
Figure 6. The H NMR (500 MHz, D2O) of triols 7, 10, 13a, and 14a in descending order.
Figure 7. (A) The GC chromatograph of acetylated triols produced by the oxidation of terminated species. (B) The GC chromatograph of the “terminated”
trialkylboranes.
cipitation, polymer 4 was isolated in 93% yield. The polymer’s
DP was observed to be 100 by GPC with a polydispersity of
1.08. The filtrate from reprecipitation and washing of the
polymer was found by TLC to contain a distribution of triols
with Rf values similar to that of 10. The triols were acetylated
by heating with excess acetic anhydride at 100 °C, followed by
purification by column chromatography. The GC chromatograph
of the acetylated triols revealed a total of five compounds
(Figure 7A).
GC-MS reveals two triacetates at 20.9 and 21.0 min, both
with masses of 374 m/z (CI) corresponding to a total DP of 4
and accounting for 14% of the terminated species. In addition,
9
J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003 12185

A R T I C L E S
Wagner et al.
Scheme 7
the peak at 21.7 min contained two triacetates, both with masses
of 388 m/z (CI) corresponding to a total DP of 5. These species
amount to 51% of the terminated catalysts. Finally, there is one
triacetate appearing at 22.4 min that has a mass of 402 m/z (CI)
corresponding to a total DP of 6 and accounting for 35% of the
terminated species. The total mass of the acetylated triols
accounts for 63% of the initiator used in the polymeriza-
tion.
The triacetates 20, 21, and 22 arise from the oxidation and
acetylation of the triorganoboranes 23, 24, and 25, respectively.
The species with a retention time of 22.4 min in Figure 7A
was separated by column chromatography. The 1H NMR
(CDCl3) spectrum of this triacetate includes a triplet at 4.07
ppm (J ) 6.8 Hz, 4H) and a doublet at 3.89 ppm (J ) 6.4 Hz,
2H) corresponding to the ester methylenes. The 13C NMR
spectrum of this compound shows a total of 13 carbon peaks.
Direct spectroscopic evidence for the terminated triorganobo-
ranes 23, 24, and 25 was obtained by conducting a polymeri-
zation similar to trial 1 in Table 1, removing solvents in vacuo
and filtering off the polymer on a plug of silica. The GC
chromatograph of the filtrate reveals three low molecular weight
components (Figure 7B). EI-GC TOF-MS established the peaks
at 12.4 and 12.6 min to have masses of 204.2057 and 204.2039
m/z (EI, calcd for C14H25B: 204.2049) corresponding to
triorganoborane species with a total DP of 5. The compounds
1
The compound was thus identified as the triacetate 20 by H
and 13C NMR (CDCl3).
1
are present in roughly a 1:1 ratio as determined by H NMR.
The peak at 15.8 min was found to have a mass of 218.2202
m/z (EI, calcd for C14H25B: 218.2202) corresponding to a total
DP of 6. The minor species with a total DP of 4 were not
observed after filtering through silica.
Additional experiments were conducted to elucidate the
sequence of events at the early stages of polyhomologation.
After the addition of 3 equiv of ylide 1 to a solution of
1-boraadamantane‚THF (2), followed by oxidation and acety-
lation, a mixture of triacetates was obtained in 93% yield
(Scheme 7).
Additionally, the species appearing in the peak at 21.7 min
in Figure 7A were separated by column chromatography and
identified as a 1:1 mixture of triacetates 21 and 22 by 1H NMR
and 13C NMR.
The GC chromatograph of the triacetates produced in the
reaction shown in Scheme 7 reveals a distribution of five
products arising from one to four homologations, with no
products resulting from five homologations (Figure 8A).
Triacetates 12, 13b, and 14b were identified by their retention
times and associated masses from GC-MS. The products of four
9
12186 J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003

The Polyhomologation of 1-Boraadamantane
A R T I C L E S
Figure 8. (A) The GC chromatograph of the triacetates produced by the reaction of 3 equiv of 1 with 1-boraadamantane‚THF (2) followed by oxidation and
acetylation. (B) The GC chromatograph of the triacetates produced by the reaction of 4 equiv of 1 with 1-borahomoadamantane‚DMSO (9) followed by
oxidation and acetylation.
Scheme 8
homologations in Figure 8A appearing at retention times of 20.9
and 21.0 min were separated from triacetates 12, 13b, and 14b
by column chromatography and identified as a mixture of
triacetates 26 and 27 by 1H NMR and 13C NMR. The preceding
results offer a clear indication that an energetic barrier exists
between the fourth and fifth homologation steps.
oxidation and acetylation, a mixture of seven triacetates was
obtained in quantitative yield (Scheme 8).
The distribution contains products arising from one to five
homologations of the 1-boraadamantane core, with no products
resulting from six homologations (Figure 8B). This result
indicates that an energetic barrier exists between the fifth and
sixth homologation of 1-boraadamantane as well. All seven
triacetates were identified by their retention times. No evidence
In a similar experiment, this time adding 4 equiv of ylide 1
to a solution of 1-borahomoadamantane‚DMSO 9, followed by
9
J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003 12187

A R T I C L E S
Wagner et al.
Scheme 9
Computational Studies
Figure 9. Products (outlined) arising from two homologations of trialky-
lborane 17 with ylide 1.
Because of the complexity of the branching pathways that
are available to propagating species in the polyhomologation
of 1-boraadamantane, a modeling study of each trialkylborane
species and branching pathway was undertaken in an effort to
corroborate and extend the experimental findings. From 1-boro-
adamantane, which contains all six-membered rings, the tricyclic
boranes examined by modeling consist of 6- to 13-membered
rings. The size of the rings expands from the common to
medium to large rings. It is of interest that many of the tricyclic
boranes that experience severe transannular ring strains, such
as species 23-25, 28, and 29, fall in the medium-sized range
of rings. In the computational study, the ground-state conforma-
tions of the free trialkylborane species were first optimized.
Next, the energy of complex formation between the free
trialkylborane species with ylide 1 was calculated. Finally, the
energies of transition states leading to the different, isomeric
trialkylboranes were assessed to elucidate the lowest energy
pathways of polymerization.
was found for products with a total DP of 5 other than triacetates
21 and 22. The identification of the triacetates in the reaction
shown in Scheme 8 demonstrates that species 17 only gives
rise to the organoboranes 28 and 29, which give rise directly to
organoboranes 24 and 25, respectively. Thus, the pathways of
polymerization from three homologations to five homologations
of the 1-boraadamantane core have been experimentally deter-
mined (Figure 9).
Interestingly, by the fifth homologation stage of the polyho-
mologation of 1-boraadamantane, the only two organoborane
species with a total DP of 5 observed are the same species that
are observed to be the major component of the two-thirds
terminated catalysts. These results led to speculation that the
one-third propagating catalysts must arise from the organoborane
species 24, because species 25 leads directly to species 26 which
is also found in the residual two-thirds terminated catalysts.
Although acyclic trialkylboranes have a flat geometry utilizing
the sp2 hybrid orbitals of boron, cyclic trialkylboranes may
experience pyramidalization because of geometrical restrictions.
The boron centers in 1-boraadamantane, 1-borahomoadaman-
tane, and other smaller tricyclic boranes show a calculated
outward pyramidalization. Larger boracycles such as 23-25,
28, and 29, however, show a flat or inward pyramidalization at
the boron center. The inward distortion of bridgehead boron
centers for larger boracycles reduces the transannular and angle
strain present in medium and large multicyclic rings. The θσp
- 90° value for inward pyramidal distortions of the HF/3-21G*
geometry optimized structures ranges from 2.0° for 28 to 4.6°
for 23. The boron centers of these systems should be less Lewis
acidic, because the formation of a borate complex requires
outward pyramidalization. The outward rehybridization exac-
erbates transannular strain, which further destabilizes complex
formation for these inward pyramidal species.
As an initial survey of possible ylide borane constructs for
computational studies, trimethylborane ylide complexes were
examined. The lowest energy conformation of the trimethylbo-
rane-dimethylsulfoxonium methylide complex is shown sche-
matically as A in Scheme 9. The substituents at sulfur and boron
are slightly twisted from the staggered conformation. The other
rotational isomers B and C, although seemingly less sterically
To give experimental supporting evidence to the theory that
24 is the species through which polymerization proceeds, the
isolated trialkylboranes 23, 24, and 25 were resubjected to ylide
1 and subsequently oxidized under the normal polyhomologation
protocol. According to the observed DP (GPC) of the polymer
recovered, one-third of the two-thirds residual trialkylboranes
23, 24, and 25 react to give a star polymer 4 with a PDI of 1.04
upon this second exposure to ylide. This result implies that at
least one of the residual trialkylboranes 23, 24, or 25 is more
active than the remaining two. Examining the species resulting
from the oxidation and acetylation of the residual boranes after
the second exposure to ylide 1, we should expect to see a
predominance of triacetate 20 and 22, if the borane leading to
the triacetate 21 is largely consumed upon reexposure to ylide.
In the GC chromatograph of the triacetates produced by this
experiment, we do observe an enrichment of triacetate 20 (49%
composition) and a depletion of triacetates 21 and 22 (47% total
composition). Additionally, the ratio of triacetates 21:22 drops
from 1:1 after the first exposure to ylide to 1:3 after the second
exposure to ylide. These results strongly suggest that species
24 is the primary contributor to the active one-third catalytic
species that continue to propagate and result in polymer
formation.
9
12188 J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003

The Polyhomologation of 1-Boraadamantane
A R T I C L E S
hindered, were found to be less stable at all levels of
computational theory used in the study. The energy differences
between A, the most stable form, and B were 3.52, 4.68, and
5.02 kcal/mol at the 3-21G*/3-21G*, 6-31G*/6-31G*, and
B3LYP/6-31G* levels of theory, respectively. Conformation C
is 3.60 kcal/mol higher in energy than A at the 3-21G*/3-21G*
level of theory. The anti oriented S-O bond to the CH2-B
can effectively stabilize the negative charge associated with the
borate component. Conformations resembling A, therefore, were
used for the calculation of transition state energies and complex
formation energies for the trialkylborane‚ylide complexes of
1-boraadamantane and its homologated boranes.
Estimation of activation energies for the pathways of homolo-
gation is helpful to compare the time-dependent reaction
progress and rationalize the observed products and product
distributions. From the calculations, the reaction energy profiles
varied depending on the homologation stage as shown in Figure
12.
The reaction in which 1 equiv of ylide 1 is added to
1-boraadamantane (Scheme 4) resulting in a single product of
monohomologation is best described by the energy profile in
Figure 12a. The subsequent second, third, and fourth stages of
homologation have energy profiles best represented by the
profile in (b). Therefore, addition of 1 equiv of ylide 1 to
1-borahomoadamantane leads to product mixtures comprised
of unreacted, mono-, and polyhomologated products. The energy
profile in (b) is the same profile that applies to the parent
reaction with trimethylborane. The energy profiles for the fifth,
sixth, and seventh homologation stages resemble the profile
shown in (c). For the fifth through seventh homologation stages,
the energy of complex formation (EC) is positive, and the
activation energy for those systems is the sum of ETS and EC.
For the eighth stage of homologation and subsequent homolo-
gations, the energy profiles return to resemble the profile in
(b). Further stages in homologation will have energy profiles
of type (b), because the cyclic boranes containing very large
rings should behave like acyclic alkyl boranes.
Even after the structure of the complex was restricted to
conformation A, a total of six conformations could exist for
each complex formed with an asymmetric borane. As shown
by A(y) and A′(y) in Scheme 9, for asymmetric boranes there
are two possible conformational isomers with the sulfoxide
portion anti to each different alkyl group. All six conformations
were examined for each asymmetric borane to determine the
lowest energy conformation. A summary of the energies of
complex formation and the transition state energies for each
trialkylborane species produced during the first seven homolo-
gations of 1-boraadamantane is illustrated in Figure 10.
The calculations reveal that the transition state energies,
defined as ETS ) E(transition state) - E(complex), of the lowest
energy transition state at each homologation stage did not change
significantly whereas the energy of complex formation was
much more variable. For convenience, matrix notation is used
in addition to compound numbers when referring to a particular
trialkylborane to specify a given isomer at a given homologation
stage. In this system of notation, the 1-boraadamantane cage is
denoted as the [0.0.0] organoborane, and the 1-borahomoada-
mantane cage is denoted as the [1.0.0] species. A maximum
difference of 4 kcal/mol was found between the [0.0.0]
1-boraadamantane system (ETS ) 18.82 kcal/mol) and the [2.2.0]
system (29) (ETS ) 14.82 kcal/mol). The transition state energy
is found to decrease as the ring size increases, with minima
occurring at both the fifth and the sixth homologation stages.
This trend is opposite to the trend in complexation energy which
increases with the homologation stage, reaching a maximum at
the fifth and sixth homologation stages. Systems with low
transition state energies are aided by the partial relief of strain
from the tetravalent borate complex to the transition state that
has a flattened trialkylboranyl ring portion. An illustration of
the complexation of a strained tricyclic borane is shown in
Figure 11.
The calculated activation energies thus obtained were plotted
for the three energetically possible reaction paths in Figure 13
to elucidate which pathway is the probable pathway for active
catalysts.
While the activation energy barriers remain relatively steady
until the fourth homologation stage, the activation energies in
all pathways increase for the fifth, sixth, and seventh stages of
homologation. It is evident that the reaction via the [2.2.1]
species (32) leading to the [2.2.2] species cannot be an actual
polymerization path because of a calculated energy difference
of 10.3 kcal/mol from the lowest energy path via the [5.1.0]
organoborane (33). The path via the [3.3.0] species (23) also
has an energy difference 3.4 kcal/mol higher than that via the
[5.1.0] system (33). We believe this energy difference is high
enough to produce the observed selectivity. The polyhomolo-
gation with 1-boraadamantane, therefore, likely proceeds via
the [4.1.0] species (24) to the [5.1.0] organoborane (33) to the
[6.1.0] species (34), ultimately leading to the propagating
catalysts. The low activation energy for the homologation with
the [6.1.0] borane (34), 16.27 kcal/mol, suggests that strain is
relieved and further homologations will have similarly low
activation energies or, at most, energies close to those for the
homologation with trimethyl borane (18.4 kcal/mol).
The calculated values of transition state energies summarized
in Figure 10 enabled the mapping and comparison of the
different modes of the polyhomologation of 1-boraadamantane
until the eighth stage. Thick solid arrows in Figure 10 represent
the probable reaction flow where the competing isomeric
transition states have more than 2 kcal/mol higher energy. After
branching at the second and third stages of homologation, three
probable homologation paths were identified as energetically
feasible to lead to products at the fourth stage of homologation.
Experimentally, however, branching was not observed at the
second stage of homologation, although, in the experimentally
observed pathway of homologation, branching from the third
stage of homologation matched perfectly with one of the two
energetically reasonable pathways elucidated by modeling.
If a fraction of propagating species pass through the fifth to
seventh homologation stages, then the remaining unreacted
boranes would compete with the higher homologated boranes
having relatively low activation energies. If we assume an
activation energy of 18.4 kcal/mol for the polymerizing ho-
mologation reactions, the values of 24.29 kcal/mol for the [3.3.0]
system (23) are 5.9 kcal/mol higher. Kinetic simulation for a
simple competing polymerization reaction in which a 3 kcal/
mol difference in activation energy for competing reactants was
imposed shows that roughly 83% and 74% of unfavorable
reactants remain after reacting with 20 and 40 equiv of ylide,
respectively, at 298 K. Imposing a 5 kcal/mol difference in the
9
J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003 12189

A R T I C L E S
Wagner et al.
Figure 10. Results of a molecular modeling study of the initial stages of the polyhomologation of 1-boraadamantane‚THF.
simulation leads to 98% and 96% of unfavorable reactants
remaining after reaction with 40 and 100 equiv of ylide,
respectively, at 298 K. Thus, a significant fraction of the [3.3.0]
borane (23) and other five-carbon homologated boranes such
as [4.1.0] (24), [3.2.0] (25), and [2.2.1] (32) could plausibly
survive the conditions of the polyhomologation reaction.
It is of interest whether further homologations would also
proceed selectively past the eighth stage of homologation. One
9
12190 J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003

The Polyhomologation of 1-Boraadamantane
A R T I C L E S
Figure 11. An illustration of the [3.2.0] system (25), its complex, and transition state structures. (i) Tricyclic boranes at the fourth through sixth homologated
stages contain boron centers with inward pyramidalization to reduce transannular interaction and angle strain. (ii) The boracycles are destabilized due to the
complex formation with the ylide 1. (iii) Transannular strain energy is partially relieved at the transition state.
Figure 12. The changing energy profiles of complexation and subsequent homologation describing different stages among the first through eighth homologation
stages.
might expect polyhomologation to proceed only along one or
some selected paths considering the computational results for
the first through eighth stages of homologation. The homolo-
gation reactions with cyclic boranes containing large rings,
however, take place rather randomly, because large cyclic
boranes such as the [6.1.0] species (34) and the [6.1.1]
organoborane (35) have several conformational isomers with
ylide 1 similar in energy to the lowest energy isomer. Inevitably,
this causes the transition energies for the three regioisomeric
trialkylborane‚ylide complexes to be closer. The actual structure
of the polymer (4) is not highly symmetric even with the random
homologation after the eighth or ninth homologation stage,
because the probable borane formed after the eighth homolo-
gation step, the [6.1.1] borane (23), has one arm longer by five
carbons. However, in the limit of making large polymers, this
difference becomes insignificant, and a fairly symmetric, regular
star polymer 4 is expected.
structures 3. The macrotricyclic species 3 may be further
elaborated by oxidation to regular star polymers 4 of low PDI.
Discrepancies between the calculated and observed molecular
weight of 4 led to a detailed mechanistic study of the initial
stages of polyhomologation. In the mechanistic study, 1-bora-
homoadamantane‚DMSO (9) was synthesized by the controlled
monohomologation of 1-boraadamantane‚THF (2) with ylide
1. A high degree of selectivity in the second and third stages of
homologation was also observed.
After a polymerization experiment, the discrepancy between
the theoretical and observed polymer molecular weight was
established by isolating and identifying the two-thirds of
trialkylborane species that did not result in polymer formation.
Because the two-thirds residual trialkylboranes occurred at the
fifth and sixth stages of homologation, further homologation
experiments were conducted to probe the initial stages of
polyhomologation. These experiments showed that no additional
species were produced in addition to the ones that were isolated
and identified as the residual boranes. Molecular modeling
corroborated this observation because estimates of the lowest
Conclusions
1-Boraadamantane‚THF (2) can be used as an initiator for
polyhomologation to give rise to macrotricyclic trialkylborane
9
J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003 12191

A R T I C L E S
Wagner et al.
hexanes) were recorded using either a VG 7070e high-resolution mass
spectrometer or a Fisons Autospec mass spectrometer. Differential
scanning calorimetry (DSC) was conducted on a Dupont Thermal
Analyst 2000 with a 910 DSC module at a 10 °C/min heating rate
under a N2 atmosphere with an aluminum oxide reference. Melting
points were assayed on a Thomas-Hoover capillary melting point
apparatus.
General Procedures. Tetrahydrofuran, methylene chloride, diethyl
ether, and benzene were dried by filtration through alumina according
to the procedure described by Grubbs.33 All other solvents were distilled
from CaH2 prior to use. Boron trifluoride etherate (BF3‚Et2O, Aldrich)
was distilled over CaH2 under reduced pressure (20 mmHg, 65 °C).
Allyl bromide (Aldrich) was distilled over CaH2. All reactions were
run in flame-dried glassware under a positive N2 or argon atmosphere.
Removal of volatile solvents transpired under reduced pressure using
a Bu¨chi rotary evaporator and is referred to as removing solvents in
vacuo. Thin-layer chromatography was conducted on precoated (0.25
mm thickness) silica gel plates with 60F-254 indicator (Merck). Column
chromatography was conducted using 230-400 mesh silica gel (E.
Merck reagent silica gel 60).
Computational Methods. All calculations were performed by
Silicon Graphics workstations. Molecular mechanics calculations were
carried out using MacroModel (version 7.0). Ab initio calculations were
performed by Spartan34 (version 5.0) and Gaussian 9835 (revision A.11)
programs. Spartan software was used for structural modifications,
whereas Gaussian 98 was used for energy and frequency calculations.
The first step for finding the energies of conformational isomers for
the cyclic boranes and the ylide complexes was the conformational
search for the corresponding cycloalkanes. The molecular mechanics
conformational searching routine imbedded in MacroModel software
was utilized with the MM3 force field. Conformations within 2 kcal/
mol were transferred to Spartan software and fully optimized at the
HF/3-21G* level of theory. The structures of the lowest energy
conformations were essentially the same regardless of computational
levels with MM3 molecular mechanics or HF/3-21G*. The lowest
energy conformations thus obtained were used as the basic skeletons
for constructing the borate complexes. The ylide-borate complex had
mostly six different conformations depending on the orientations of
the ylide portion (see Discussion section). All six conformations were
optimized independently to find the lowest energy conformation. The
structures of transition states were optimized from the guessed structures
obtained either by gradually changing the distance of leaving group or
by optimizing the fixed geometry with the values for the optimized
transition state of similar structure. The optimized structures of transition
states were verified as having a single imaginary vibrational frequency
and a normal mode consistent with the anticipated reaction coordinate
from the calculated frequencies. Transition states also had six possible
conformations mostly for a single reaction. Thus, all six conformations
were tested for the calculations. All energy was optimized at the HF/
3-21G* level and corrected with unscaled zero point energy. All kinetic
simulations were performed using Chemical Kinetics Simulator36
(version 1.01).
Figure 13. A plot of activation energies for three different pathways at
the initial stages of polyhomologation. The data point labeled [1.0.0]
represents the transition state energy for the homologation of the [0.0.0]
species.
energy pathway for polymerization and kinetic simulation
showed that polymerization could proceed through the boranes
that were isolated as residual boranes after polyhomologation.
In accordance with the model, we hypothesized, tested, and
demonstrated that a large fraction of the one-third active
propagating species species that ultimately results in polymer
formation proceeds through the [4.1.0] trialkylborane (24).
According to the model, the lowest energy pathway for the
homologation of 24 transpires to give the [5.1.0] triorganoborane
(33) followed by the [6.1.0] (34) and finally the [6.1.1]
trialkylborane (35), after which each branch on the cyclohexane
core participates in polyhomologation. During each polymeri-
zation, it is curious that only one-third of the propagating species
result in polymer formation and, further, that those propagating
species largely proceed through 24 and yet there is a fraction
of 24 that remains intact after polymerization. However,
molecular modeling and kinetic simulation support this scenario
in an analysis of the pertinent geometrical, energetic, and kinetic
factors contributing to the loss of activity in the tricyclic boranes
of medium ring size. Additionally, the second exposure of ylide
to boranes 23, 24, and 25 showed a substantial depletion of
borane 24 relative to 23 and 25, strongly supporting the lowest
energy pathway elucidated by molecular modeling.
Experimental Section
(33) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.; Timmers,
F. J. Organometallics 1996, 15, 1518.
(34) Spartan version 5.0, Wavefunction, Inc.: Irvine, CA.
Instrumentation. All 1H NMR spectra were acquired at 400 or 500
MHz on Bruker spectrometers. Chemical shifts (δ) are listed in ppm
against deuterated solvent peaks as an internal reference. Coupling
constants (J) are reported in hertz, and the abbreviations for splitting
include: s, single; d, doublet; t, triplet; q, quartet; p, pentet; m, multiplet;
br, broad. All 13C NMR spectra were acquired on Bruker instruments
at 125.8 MHz. Chemical shifts (δ) are listed in ppm against solvent
carbon peaks as an internal reference. All 11B NMR spectra were
acquired at 237 MHz on a Bruker spectrometer. All chemical shifts
(δ) are listed in ppm against boron trifluoride etherate as an external
reference. Infrared spectra (IR) were assayed on a Perkin-Elmer 1600
Series FTIR spectrometer. GPC was conducted on a Waters 150-C Plus
Gel permeation chromatograph. High-resolution mass spectra (EI,
(35) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-
Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.11;
Gaussian, Inc.: Pittsburgh, PA, 1998.
(36) Chemical Kinetics Simulator was obtained from the IBM’s Almaden
Research Center Web site: http://www.almaden.ibm.com/st/msim/ckspa_n-
t.html.
9
12192 J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003

The Polyhomologation of 1-Boraadamantane
A R T I C L E S
Dimethylsulfoxonium Methylide 1. General Method. To a flame-
dried, three-necked flask fitted with a reflux condenser connected to a
nitrogen inlet was added a 60% dispersion of NaH (4.05 g, 101 mmol)
in mineral oil. The dispersion of NaH in mineral oil was washed with
dry hexanes (3 × 30 mL). Trimethylsulfoxonium chloride (9.5 g, 74
mmol) was added to the flask, and anhydrous toluene (90 mL) was
added. The reaction mixture was vigorously stirred and heated to 110
°C. The reaction solution was allowed to stir at reflux until the evolution
of H2 ceased and the color of the solution changed from gray to white.
The ylide solution in toluene was filtered through a Schlenk filter
containing dry Celite 545 under N2. An aliquot of the ylide solution
(0.250 mL) in water was treated with 1 drop of a solution of
phenolphthalein in EtOH and titrated with 0.1000 N HCl. The
concentration was 0.61 M. The yield was 76%. The ylide solution was
stored at -20 °C and remained stable for several weeks.
1-Boraadamantane THF 2. The procedure of Mikhailov and co-
workers was followed to synthesize 2.11 To a flame-dried, two-neck
50 mL flask with a stir bar and condenser was added 3-methoxy-7-
(methoxymethyl)-3-borabicyclo[3.3.1]non-6-ene (2.8698 g, 14.8 mmol).
A solution of borane THF (14.80 mL, 1 M, 14.8 mmol) was slowly
added with stirring. The reaction was exothermic and came to reflux
during addition of the borane THF. The reaction solution was then
lowered into an oil bath preheated to 80 °C and refluxed for 1 h. The
THF was removed by distillation at atmospheric pressure under
nitrogen, followed by vacuum distillation (20 mmHg). The white residue
was then sublimed under vacuum (0.100 mmHg, 70 °C) onto an ice-
water chilled coldfinger to yield 1-boraadamantane THF 2 (1.6140 g,
53%) as a solid white crystalline compound, mp 89-91 °C. 1H NMR
(500 MHz, C6D6): δ 3.34 (t, J ) 6.8, 4H), 2.74 (br s, 3H), 2.00 (m,
6H), 1.01 (m, 10H). 13C NMR (125.8 MHz, C6D6): δ 68.4, 40.7, 34.6,
29.7 (br s), 24.3. 11B NMR (237 MHz, CDCl3): δ 5.10 (br s). IR
(KBr): 2868, 1438, 1350, 1299, 1252, 1226, 1192, 1119, 1078, 1025,
982, 930, 918, 855, 728, 614 cm-1. HRMS (CI): m/z calcd for C13H24-
BO (M + H)+ 207.1923, found 207.1929.
Star Polymer 4. General Method. Toluene solutions of ylide 1
were preheated to 80 °C and treated with an aliquot of a toluene solution
of 1-boraadamantane‚THF (2). The ylide was rapidly consumed (5 min),
and trimethylamine-N-oxide dihydrate (3.1 equiv) was added to effect
oxidation under reflux conditions. Removal of solvents followed by
filtering and washing of the polymeric solids with methanol, water,
and hexanes gave near quantitative crude yields of polymeric triol 4.
Purification by reprecipitation in toluene/acetonitrile afforded a purified
white polymer 4 in high yield, mp 125 °C (spectroscopic data for trial
2 of Table 1). 1H NMR (500 MHz, C6D6, 78 °C): δ 3.38 ppm (t, J )
6.4 Hz, 6H), 1.85 ppm (d, J ) 12.2, 3H), 1.35 (m (br), 890H), 0.94 (d,
J ) 6.4, 3H), 0.59 ppm (q, J ) 11.8, 3H), 0.50 (s, 3H). 13C NMR
(125.8 MHz, C6D6, 78 °C): δ 62.8, 40.9, 38.2, 38.1, 37.6, 33.3, 30.6,
30.5, 30.1, 30.0, 29.9, 27.6, 27.5, 26.2, 20.0. IR: 3422 (br), 2918, 1473,
1057, 719 cm-1. MN ) 5595 (GPC).
2.49 (br s, 1H), 2.43 (br s, 1H), 2.25 (d, J ) 17.9, 1H), 1.74 (d, J )
12.7, 1H), 1.70 (d, J ) 17.8, 1H), 1.60 (d, J ) 12.7, 1H), 1.04 (d, J )
17.4, 1H), 0.96 (dd, J ) 17, 7.5, 1H), 0.86 (d, J ) 17.4, 1H), 0.83 (dd,
J ) 17, 5.3, 1H). 13C NMR (125.8 MHz, CDCl3): δ 131.7, 130.8,
77.0, 56.9, 53.0, 34.9, 32.5, 29.0, 26.9, 25.4 (br s), 24.1 (br s). 11B
NMR (237 MHz, CDCl3): δ 55.4 (br s). IR (thin film): 3397, 2909,
1465, 1431, 1371, 1298, 1167, 1101, 1038, 1019, 996, 955, 914, 880,
847, 831, 678 cm-1. HRMS (EI/hexanes): m/z calcd for C11H19BO2
(M+) 194.1478, found 194.1481.
Triallylborane 6. The protocol of Zhakarkin and co-workers was
followed to synthesize 6.12 To a flame-dried 250 mL round-bottom
flask sealed with rubber septum and equipped with a magnetic stir bar,
reflux condenser, and thermometer were transferred solid aluminum
granules (12.03 g, 446 mmol) and HgCl2 (s) (88 mg, 0.32 mmol). Dry,
N2 purged diethyl ether (60 mL) was added, and the solution was stirred
and heated to 35 °C. Allyl bromide (52.1 mL, 602 mmol) was added
dropwise under a nitrogen atmosphere. Care must be taken to add the
allyl bromide slowly, because the reaction with aluminum is exothermic.
The reaction mixture was stirred at reflux in an oil bath set at 67 °C
for 3 h. The reaction mixture was then transferred via syringe into a
septum sealed flame-dried 250 mL round-bottom flask containing BF3‚
Et2O (20.0 mL, 142 mmol) under nitrogen. This solution was refluxed
at 70 °C for 3 h. The ether was removed by short-path distillation under
nitrogen. The residue was then distilled under vacuum (20 mmHg) with
a short-path distillation head at a head temperature of 65 °C to give
crude triallylborane as a clear, colorless liquid (17.00 g, 78%). 11B NMR
(237 MHz, CDCl3) of this distillate showed a major peak, δ 80.7 (br
s) 83%, and minor peaks at δ 50.6 (br s) 11%, and δ 32.0 (br s) 6%.
The peak at δ 80.7 (br s) has been assigned to triallylborane. On the
basis of chemical shift, the peak at δ 50.6 has been tentatively assigned
to a borinic ester of some type, most likely the product of the
triallylborane’s reaction with oxygen. A second vacuum distillation (20
mmHg) yielded pure triallylborane as a clear, colorless liquid (13.07
g, 60%). 1H NMR (500 MHz, CDCl3): δ 2.2 (br s, 6H), 4.9 (br s, 6H),
5.89-5.98 (p, J ) 10.6, 3H). 13C NMR (125.8 MHz, CDCl3): δ 134.8,
114.7 (br s), 34.4 (br s). 11B NMR (237 MHz, CDCl3): δ 80.6 (br s).
IR (neat): 3076, 2999, 2973, 2914, 1807, 1635, 1420, 1370, 1270,
1170, 993, 900 cm-1
.
cis-1,3,5-Trihydroxymethylcyclohexane 7. To a solution of 1-bo-
raadamantane THF 2 (1.47 g, 7.1 mmol) in toluene (18 mL) was added
trimethylamine N-oxide dihydrate (2.39 g, 21.4 mmol), and the solution
was refluxed for 16 h. After the solution cooled to 100 °C, 1 mL of
H2O was added. The toluene, trimethylamine, and H2O were removed
in vacuo, and the residue was purified by flash chromatography (9:1
EtOAc:MeOH) and recrystallized in acetone to give the triol 7 (0.944
g, 76%) as a white crystalline solid, mp 96-98 °C.9 1H NMR (500
MHz, D2O): δ 3.36 (d, J ) 6.3, 6H), 1.68 (d, J ) 12.5, 3H), 1.52 (m,
3H), 0.50 (q, J ) 12.2, 3H). 13C NMR (125.8 MHz, D2O): δ 31.8,
35.8, 67.3. IR: 3356, 2915, 1064, 1024 cm-1. HRMS (CI): m/z calcd
for C9H19O3 (M + H)+ 175.1334, found 175.1338.
3-Methoxy-7-(methoxymethyl)-3-borabicyclo[3.3.1]non-6-ene 5.
The method of Mikhailov and co-workers was followed to synthesize
5.11 To a flame-dried, two-neck 50 mL flask with a stir bar and
condenser was added distilled triallylborane (6) (8.5 g, 63 mmol) under
a nitrogen atmosphere. The flask was heated with stirring under nitrogen
at 130 °C in a preheated oil bath (10 min). Methyl propargyl ether
(4.5 mL, 64 mmol) was added via syringe over the course of 10 min
to the stirring triallylborane. The reaction was heated at 130 °C for 1
h. The reaction solution was cooled to room temperature, and methanol
(2.9 mL, 72 mmol) was added dropwise under nitrogen with stirring.
The reaction was exothermic and accompanied by the evolution of
propylene gas. After an additional 5 min of stirring, the reaction mixture
was distilled under vacuum (0.10 mmHg) with a short-path distillation
head at a head temperature of 79 °C to give NMR-pure 3-methoxy-
7-(methoxymethyl)-3-borabicyclo[3.3.1]non-6-ene 5 (10.75 g, 87%) as
1-Borahomoadamantane DMSO 9. To a solution of 1-boraada-
mantane THF 2 (1.67 g, 8.0 mmol) in toluene (12 mL) at -78 °C was
slowly added a solution of dimethylsulfoxonium methylide in toluene
(17.0 mL, 0.49 M, 8.3 mmol). The reaction was maintained at -78 °C
overnight. The resulting white crystals that precipitated were filtered,
washed with hexanes, transferred to a 20 dram vial, and dried under
high vacuum (1 mmHg) at -60 °C for 3 h. Upon warming to room
temperature under high vacuum (1 mmHg), the white crystals melted
to a liquid, releasing heat, and resolidified upon cooling. The white
residue was then sublimed under vacuum (0.100 mmHg, 73 °C) onto
an ice-water chilled coldfinger to yield 1-borahomoadamantane DMSO
9 (1.18 g, 65%) as a solid white crystalline compound, mp 73-75 °C.
1H NMR (500 MHz, DMSO-d6): δ 2.52 (s, 6H), 1.90 (m, 2H), 1.86
(m, 1H), 1.70 (m, 2H), 1.48 (m, 2H), 1.26 (m, 4H), 0.50 (m, 6H). 13
C
1
a clear, viscous liquid. H NMR (500 MHz, CDCl3): δ 5.66 (br s,
NMR (125.8 MHz, DMSO-d6): 40.4, 39.1, 39.0, 33.0, 32.8, 31.5, 29.2,
23.7. 11B NMR (237 MHz, DMSO-d6): δ 5.7 (br s). IR (KBr): 2892,
1H), 3.74 (d, J ) 11.7, 1H), 3.64 (d, 11.8), 3.58 (s, 3H), 3.19 (s, 3H),
9
J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003 12193

A R T I C L E S
Wagner et al.
1406, 1033, 912, 846, 635, 565 cm-1. HRMS (CI): m/z calcd for C12H24-
OBS (M + H)+ 227.1643, found 227.1647.
was added, and the reaction was heated to 114 °C with stirring (12 h).
Solvents were removed in vacuo, and the residue was purified by
column chromatography (SiO2, CHCl3:EtOH 17:3). To the purified
residue was added acetic anhydride (12 mL, 126 mmol), and the reaction
was heated to 100 °C with stirring for 1 h. The excess acetic anhydride
was removed in vacuo, and the crude product was purified by column
chromatography (SiO2, hexanes:EtOAc 23:2) to give a mixture of
triacetates 13b, 14b, and 15b as a clear, pale yellow oil (0.74 g, 93%
yield). Triacetate 13b accounted for 43% of the mixture by GC (0.31
g). A sample of triacetate 13b was obtained by column chromatography
cis-1-Hydroxyethyl-3,5-bis(hydroxymethyl)cyclohexane 10. To a
solution of 1-borahomoadamantane‚DMSO 9 (0.65 g, 2.9 mmol) in
toluene (7.0 mL) was added trimethylamine-N-oxide dihydrate (0.97
g, 8.6 mmol), and the reaction was refluxed overnight. The toluene,
trimethylamine, and H2O were removed in vacuo, and the residue was
purified by flash chromatography (9:1 CHCl3:MeOH) to give the triol
10 (0.47 g, 87%) as an oil. 1H NMR (500 MHz, D2O): δ 3.54 (t, J )
6.5, 2H), 3.30 (d, J ) 11.8, 4H), 1.64 (d, J ) 12.0, 3H), 1.45 (m, 2H),
1.38 (m, 3H), 0.46 (q, J ) 12.2, 3H). 13C NMR (125.8 MHz, D2O): δ
1
(SiO2, hexanes:EtOAc 23:2) as a clear, pale yellow oil (0.015 g). H
67.6, 59.9, 39.3, 35.9, 33.1, 32.3. IR: 3332, 2914, 1062, 1015 cm-1
.
NMR (500 MHz, CDCl3): δ 4.12 (t, J ) 6.8, 4H), 3.89 (d, J ) 6.4,
2H), 2.07 (s, 3H), 2.06 (s, 6H), 1.79 (m, 3H), 1.62 (m, 2H), 1.55 (m,
3H), 1.47 (m, 2H), 0.64 (m, 3H). 13C NMR (125.8 MHz, CDCl3): δ
171.2, 69.2, 62.5, 39.4, 36.7, 35.8, 33.7, 21.0, 20.9. IR (neat): 2922,
1738, 1462, 1367, 1237, 1033, 805 cm-1. HRMS (CI): m/z calcd for
C17H29O6 (M + H)+ 329.1964, found 329.1964.
HRMS (CI): m/z calcd for C10H21O3 (M + H)+ 189.1490, found
189.1493.
cis-1,3,5-Tris(hydroxymethyl)cyclohexane Triacetate 11. To a
finely crushed powder of cis-1,3,5-trihydroxymethylcyclohexane (7)
(0.20 g, 1.1 mmol) was added acetic anhydride (5.0 mL, 53.0 mmol)
with stirring at room temperature. The reaction solution was heated to
100 °C and stirred (1 h). The reaction was observed to be complete by
TLC. The excess acetic anhydride was removed in vacuo, and the crude
product was purified by flash chromatography (3:1 hexanes:EtOAc)
to give the triacetate 11 (0.29 g, 83%) as a colorless oil. 1H NMR (500
MHz, CDCl3): δ 3.86 (d, J ) 6.2, 6H), 1.98 (s, 9H), 1.76 (d, J )
14.4, 3H), 1.69 (m, 3H), 0.65 (q, J ) 12.4, 3H). 13C NMR (125.8 MHz,
CDCl3): δ 170.5, 68.9, 36.0, 32.4, 21.0. IR (neat): 2942, 1739, 1463,
1369, 1242, 1035 cm-1. HRMS (CI): m/z calcd for C15H28NO6 (M +
NH4)+ 318.1916, found 318.1915.
cis-1-Hydroxymethyl-3-hydroxyethyl-5-hydroxypropylcyclohex-
ane 14a. To 1-borahomoadamantane‚DMSO (9) (0.85 g, 3.8 mmol)
dissolved in toluene (9 mL) was added a solution of ylide 1 in toluene
(5.9 mL, 0.64 M, 3.8 mmol) at -50 °C. After being warmed to 0 °C
with stirring (2 h), the reaction was warmed to room temperature with
stirring (2 h). Trimethylamine-N-oxide dihydrate (1.29 g, 11.6 mmol)
was added, and the reaction was heated to 114 °C with stirring (12 h).
Solvents were removed in vacuo, and the residue was purified by
column chromatography (SiO2, CHCl3:EtOH 17:3) to give a mixture
of triols 10, 13a, and 14a as a clear, colorless oil (0.73 g, 95%). A
sample of the crude product was purified by preparative TLC (SiO2,
CHCl3:EtOH 17:3) to allow the isolation of triol 14a as a clear, colorless
oil (0.030 g). 1H NMR (500 MHz, D2O): δ 3.59 (t, J ) 6.6, 2H), 3.51
(t, J ) 6.7, 2H), 3.35 (d, J ) 6.3, 2H), 1.69 (m, 3H), 1.51 (m, 3H),
1.41 (m, 3H), 1.29 (m, 1H), 1.19 (m, 2H), 0.51 (m, 3H). 13C NMR
(125.8 MHz, CDCl3): δ 67.5, 62.2, 59.7, 39.4, 39.3, 39.0, 36.1, 35.6,
35.5, 33.1, 32.9, 28.8. IR (neat): 3359 (br), 2912, 1658, 1455, 1360,
1053, 1012 cm-1. HRMS (CI): m/z calcd for C12H25O3 (M + H)+
217.1804, found 217.1806.
cis-1-Hydroxyethyl-3,5-bis(hydroxymethyl)cyclohexane Triac-
etate 12. To cis-1-hydroxyethyl-3,5-dihydroxymethylcyclohexane (10)
(0.30 g, 1.6 mmol) was added acetic anhydride (7.1 mL, 75.2 mmol)
with stirring. The reaction solution was heated to 100 °C and maintained
at this temperature (3 h) with stirring. The reaction was observed to be
complete by TLC. The excess acetic anhydride was removed in vacuo,
and the crude product was purified by flash chromatography (3:1
hexanes:EtOAc) to give the triacetate 12 (0.44 g, 86%) as a colorless
1
oil. H NMR (500 MHz, CDCl3): δ 4.09 (t, J ) 6.7, 2H), 3.88 (d, J
) 6.3, 4H), 2.04 (s, 6H), 2.02 (s, 3H), 1.77 (d, J ) 14.0, 3H), 1.71 (m,
2H), 1.55 (q, J ) 6.7, 2H), 1.47 (m, 1H), 0.66 (p, J ) 12.7, 3H). 13C
NMR (125.8 MHz, CDCl3): δ 171.1 (two peaks), 69.0, 62.3, 36.3,
35.7, 35.69, 33.3, 32.3, 21.0, 20.9. IR (neat): 2922, 1739, 1461, 1366,
1240, 1034 cm-1. Anal. Calcd for C16H26O6: C, 61.13; H, 8.34.
Found: C, 60.91; H, 8.34. HRMS (CI): m/z calcd for C16H27O6 (M +
H)+ 315.8070, found 315.8070.
cis-1-Hydroxymethyl-3-hydroxyethyl-5-hydroxypropylcyclohex-
ane Triacetate 14b. To 1-borahomoadamantane‚DMSO (9) (0.55 g,
2.4 mmol) dissolved in toluene (7 mL) was added a solution of ylide
1 in toluene (3.8 mL, 0.65 M, 2.5 mmol) at -50 °C. After being warmed
to 0 °C with stirring (2 h), the reaction was warmed to room temperature
with stirring (2 h). Trimethylamine-N-oxide dihydrate (0.84 g, 7.5
mmol) was added, and the reaction was heated to 114 °C with stirring
(12 h). Solvents were removed in vacuo, and the residue was purified
by column chromatography (SiO2, CHCl3:EtOH 17:3). To the purified
residue was added acetic anhydride (12 mL, 126 mmol), and the reaction
was heated to 100 °C with stirring for 1 h. The excess acetic anhydride
was removed in vacuo, and the crude product was purified by column
chromatography (SiO2, hexanes:EtOAc 23:2) to give a mixture of
triacetates 13b, 14b, and 15b as a clear, pale yellow oil (0.74 g, 93%
yield). Triacetate 14b accounted for 31% of the mixture by GC (0.23
g). A sample of triacetate 14b was obtained by column chromatography
cis-1-Hydroxymethyl-3,5-bis(hydroxyethyl)cyclohexane 13a. To
1-borahomoadamantane‚DMSO (9) (0.85 g, 3.8 mmol) dissolved in
toluene (9 mL) was added a solution of ylide 1 in toluene (5.9 mL,
0.64 M, 3.8 mmol) at -50 °C. After being warmed to 0 °C with stirring
(2 h), the reaction was warmed to room temperature with stirring (2
h). Trimethylamine-N-oxide dihydrate (1.29 g, 11.6 mmol) was added,
and the reaction was heated to 114 °C with stirring (12 h). Solvents
were removed in vacuo, and the residue was purified by column
chromatography (SiO2, CHCl3:EtOH 17:3) to give a mixture of triols
10, 13a, and 14a as a clear, colorless oil (0.73 g, 95%). A sample of
the crude product was purified by preparative TLC (SiO2, CHCl3:EtOH
17:3) to allow the isolation of triol 13a as a clear, colorless oil (0.010
1
(SiO2, hexanes:EtOAc 23:2) as a clear, pale yellow oil (0.03 g). H
NMR (500 MHz, CDCl3): δ 4.12 (t, J ) 6.8, 2H), 4.05 (t, J ) 6.8,
2H), 3.89 (d, J ) 6.4, 2H), 2.07 (s, 3H), 2.06 (s, 6H), 1.78 (m, 3H),
1.70 (m, 3H), 1.67 (m, 2H), 1.54 (m, 2H), 1.45 (m, 1H), 1.35 (m, 1H),
1.27 (m, 2H), 0.61 (m, 3H). 13C NMR (125.8 MHz, CDCl3): δ 171.9,
69.4, 64.7, 62.5, 39.5, 36.8, 36.4, 36.0, 35.9, 35.8, 33.8, 33.3, 25.9,
21.1, 21.0, 20.9. IR (neat): 2920, 1739, 1366, 1240, 1034 cm-1. HRMS
(CI): m/z calcd for C18H31O6 (M + H)+ 343.2120, found 343.2121.
cis-1-Hydroxymethyl-3,5-bis(hydroxybutyl)cyclohexane Triac-
etate 20. To 1-boraadamantane‚THF (2) (0.20 g, 1 mmol) dissolved
in toluene (10 mL) was added a solution of ylide 1 in toluene (36.0
mL, 0.55 M, 20 mmol) at 25 °C. After the mixture was stirred (2 h),
trimethylamine-N-oxide dihydrate (0.35 g, 3.1 mmol) was added, and
the reaction was heated to 114 °C with stirring (12 h). Solvents were
1
g). H NMR (500 MHz, D2O): δ 3.49 (t, J ) 6.5, 4H), 3.26 (d, J )
6.4, 2H), 1.58 (m, 3H), 1.39 (m, 2H), 1.32 (m, 5H), 0.42 (m, 3H). 13
C
NMR (125.8 MHz, CDCl3): δ 67.4, 59.7, 39.3, 39.2, 38.9, 35.5, 33.0.
IR (neat): 3334 (br), 2913, 1648, 1438, 1377, 1049, 1012 cm-1. HRMS
(CI): m/z calcd for C11H23O3 (M + H)+ 203.1647, found 203.1645.
cis-1-Hydroxymethyl-3,5-bis(hydroxyethyl)cyclohexane Triac-
etate 13b. To 1-borahomoadamantane‚DMSO (9) (0.55 g, 2.4 mmol)
dissolved in toluene (7 mL) was added a solution of ylide 1 in toluene
(3.8 mL, 0.65 M, 2.5 mmol) at -50 °C. After being warmed to 0 °C
with stirring (2 h), the reaction was warmed to room temperature with
stirring (2 h). Trimethylamine-N-oxide dihydrate (0.84 g, 7.5 mmol)
9
12194 J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003

The Polyhomologation of 1-Boraadamantane
A R T I C L E S
removed in vacuo, and both the polymeric product (4) and the residual
triols were purified by column chromatography (SiO2, CHCl3:EtOH
17:3). To the purified triols was added acetic anhydride (12 mL, 126
mmol), and the reaction was heated to 100 °C with stirring for 1 h.
The excess acetic anhydride was removed in vacuo, and the crude
product was purified by column chromatography (SiO2, hexanes:EtOAc
23:2) to give a mixture of triacetates as a clear, pale yellow oil (0.24
g). Triacetate 20 accounted for 35% of the mixture by GC. A sample
of triacetate 20 was obtained by column chromatography (SiO2,
1739, 1461, 1366, 1238, 1036 cm-1. HRMS (CI): m/z calcd for
C17H29O6 (M + H)+ 384.2590, found 384.2600.
Acknowledgment. We thank the Chemistry Division of the
National Science Foundation for financial support of this work
(Grant CHE-9617475). C.E.W. wishes to thank Allergan, Inc.,
for a graduate fellowship. Gracious thanks is also given to Dr.
A. W. deGroot of the Dow Chemical Co. for analysis of the
high molecular weight star polymers.
1
hexanes:EtOAc 23:2) as a clear, pale yellow oil (0.005 g). H NMR
Supporting Information Available: X-ray data for com-
pounds 2 and 9 (PDF and CIF). This material is available free
of charge via the Internet at http://pubs.acs.org.
(500 MHz, CDCl3): δ 4.07 (t, J ) 6.8, 4H), 3.89 (d, J ) 6.4, 2H),
2.09 (s, 3H), 2.07 (s, 6H), 1.75 (m, 3H), 1.65 (m, 3H), 1.36 (m, 12H),
0.57 (m, 3H). 13C NMR (125.8 MHz, CDCl3): δ 171.3, 171.2, 69.6,
64.6, 39.8, 36.9, 36.8, 36.2, 29.7, 28.9, 23.2, 21.0, 20.9. IR (neat): 2924,
JA0361291
9
J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003 12195