Oxyfunctionalization of non-natural targets by dioxiranes. 6. On the selective hydroxylation of cubane

Article abstract of DOI:10.1021/ol901302d

Image Persented By using methyl(trifluoromethyl)dioxirane (TFDO), the direct mono- and bishydroxylation of cubane could be achieved in high yield under remarkably mild conditions. Comparison of the rates of dioxirane O-insertion with those of standard reference compounds, such as adamantane and cyclopropane, as well as ab initio computations provide useful hints concerning the mechanism of these transformations.

Full text of DOI:10.1021/ol901302d

ORGANIC
LETTERS
2009
Vol. 11, No. 16
3574-3577
Oxyfunctionalization of Non-Natural
Targets by Dioxiranes. 6. On the
Selective Hydroxylation of Cubane
Cosimo Annese,† Lucia D’Accolti,† Caterina Fusco,† Remo Gandolfi,‡ Philip
E. Eaton,*,§ and Ruggero Curci*,†,|
Dipartimento Chimica, CNR.-ICCOM, UniVersita` di Bari, V. Amendola 173, I-70126
Bari, Italy, Dipartimento di Chimica Organica, UniVersita` di PaVia, V.le Taramelli 10,
27100 PaVia, Italy, Department of Chemistry, The UniVersity of Chicago, Chicago,
Illinois 60637, and Department of Chemistry, Brown UniVersity,
ProVidence, Rhode Island 02912
curci@ba.iccom.cnr.it; eaton@uchicago.edu
Received June 10, 2009
ABSTRACT
By using methyl(trifluoromethyl)dioxirane (TFDO), the direct mono- and bishydroxylation of cubane could be achieved in high yield under
remarkably mild conditions. Comparison of the rates of dioxirane O-insertion with those of standard reference compounds, such as adamantane
and cyclopropane, as well as ab initio computations provide useful hints concerning the mechanism of these transformations.
In the last decades, the direct oxyfunctionalization of
“unactivated” C-H bonds using dioxiranes (1)1,2 has opened
an important new area of oxidation chemistry. The efficient
hydroxylation of simple and/or structurally complex alkanes
under mild conditions is the highlight of the chemistry of
these extraordinarily effective oxidants. These transforma-
tions are earmarked by high selectivity, and tertiary C-H
bonds are considerably more reactive toward dioxirane
O-insertion than secondary or primary units.2,3
As a result, high tertiary vs secondary selectivities (Rst
from 15 to over 250) can be routinely achieved.4
The selective hydroxylation of adamantane (2) by diox-
iranes 1a (DDO) and 1b (TFDO) to yield adamantan-1-ol is
illustrative. With the powerful methyl(trifluoromethyl)diox-
irane (1b) in excess, the all-bridgehead hydroxylated tetraol
3 is obtained in 73% yield4 (Figure 1).
In these oxidations, kinetic data have shown that TFDO
is more reactive than DDO by a factor of almost 103, with
no loss of selectivity. This and many other facets of alkane
oxyfunctionalization using dioxiranes have been discussed
in a recent account.4
† Universita` di Bari.
‡ Universita` di Pavia.
§ University of Chicago.
| Brown University.
We have investigated the reactivity of a few strained
polycyclic hydrocarbons having a peculiar architecture, i.e.,
2,4-didehydroadamantane1 and BinorS,5 along with bicyclics
that function as fast radical-clock probes.4
(1) Part 5: D’Accolti, L.; Dinoi, A.; Fusco, C.; Russo, A.; Curci, R. J.
Org. Chem. 2003, 68, 7806–7810.
(2) For examples, see:(a) Curci, R. In AdVances in Oxygenated Pro-
cesses; Baumstark, A. L., Ed.; JAI: Greenwich, CT, 1990; Vol 2, Chapter
I, pp 1-59. (b) Adam, W.; Zhao, C.-G. In The Chemistry of Peroxides;
Patai, S., Ed.; Wiley: New York, 2006; Vol. 2, Chapter 14.
(3) (a) Curci, R.; Dinoi, A.; Rubino, M. F. Pure Appl. Chem. 1995, 67,
811–22. (b) Mello, R.; Fiorentino, M.; Fusco, C.; Curci, R. J. Am. Chem.
Soc. 1989, 111, 6749–6757. (c) Murray, R. W.; Jeyaraman, R. J. Org. Chem.
(4) Curci, R.; D’Accolti, L.; Fusco, C. Acc. Chem. Res. 2006, 39, 1–9.
See references therein.
(5) D’Accolti, L.; Fusco, C.; Lucchini, V.; Carpenter, G. B.; Curci, R.
J. Org. Chem. 2001, 66, 9063–9066.
1985, 50, 2847–2853
.
10.1021/ol901302d CCC: $40.75
Published on Web 07/22/2009
 2009 American Chemical Society

in the high-field region (at 40.9 and 40.0 ppm, respectively).
The EI mass spectrum gave M+(21) m/z 120.06, and the
formula C8H8O was confirmed by combustion analysis.
Further oxidation of cubanol 6 with TFDO in acetone (or
CH2Cl2) solution (eq 2) affords selectiVely the symmetrical
Figure 1. Dioxiranes isolated and representative substrates for
selective hydroxylations.
We now add the direct dioxirane oxyfunctionalization of
cubane (4). This is relevant because, after the pioneering
work of Eaton and co-workers in the 1960s,6a during the
past decade cubane chemistry has had a great revival.6b
In addition to the theoretical and synthetic challenges
inherent to this strained compound, much current interest
originates from the unusual stability and compact framework
exhibited by cubane functionalized derivatives, making them
attractive hyperenergetic materials.6b To our knowledge, no
effective methods for the direct hydroxylation of this
substrate have been reported. Classical “electrophilic” oxi-
dants (e.g., MCPBA) are ineffective to this end; also, cubane
is found as a poor substrate for methane monooxygenase
systems (MMO) or for P-450 enzyme oxidations, requiring
extended reaction times to effect hydroxylation in pitifully
low yields (ca. 0.05%).6c
cubane-1,4-diol (7) as the only isolable product, and this was
identified by physical and spectroscopic data in full agree-
ment with the literature.7 Structure verification of diol 7 was
straightforward given the analytical details7 and the symmetry
evident from the number of carbon signals (two) in the
{1H}13C NMR spectrum of the compound.
The ease of the transformations above is remarkable. In
view of the known facile access to cubyl cations,6a it seems
likely that these intermediates are involved as incipient
species in the oxidation at hand. Actually, experimental
evidence, as well as high-level computations, pointed out
thatsrather than a caged-biradical “rebound” mechanism
resembling P-450 oxygenationssa rather concerted “ox-
enoid” mechanism applies (Figure 2).4,8 On the whole, the
In the course of investigations herein, we verified that low
yield hydroxylation also holds when dimethyldioxirane (1a)
is applied to cubane (4). This can be ascribed to the fact
that the exocyclic orbitals in this strained cage compound
are highly pyramidalized (C-C-H angle 123-127 ( 2°)
and s rich.6a
We now report on the reactivity of the powerful TFDO
(1b) toward cubane (4) in comparison with a few reference
compounds of choice, namely, adamantane (2) and cyclo-
propane (5).
We find that treatment of cubane (4) in acetone or CH2Cl2
with isolated TFDO (1b) at subambient temperature gives
cubanol (6) in excellent yield (98%) (eq 1). In a typical
laboratory-scale reaction, cubane could be functionalized in
milligram quantities, and 5-10 mg of cubanol (6) could be
obtained readily within short reaction times.
Figure 2. Concerted mechanism and general FMO model for
dioxirane O-insertion into alkane C-H bonds.
The structure of isolated 6 could be established spectro-
scopically. Its IR spectrum showed a strong O-H stretching
absorption at 3421 cm-1. In its {1H}13C NMR spectrum, four
resonances are apparent: one C-OH at 56.0 ppm and three
well-resolved methine carbon signals attributed to unfunc-
tionalized cubane carbons, one at 55.7 ppm (R-CH) and two
process sketched in Figure 2 represents a practically con-
certed (albeit nonsynchronous) dioxirane O-insertion into
C-H bonds.4
The transition state thereof involves a shallow triangular
C-H-O assembly as verified through high-level calcula-
(6) (a) Eaton, P. E. Angew. Chem., Int. Ed. Engl. 1992, 21, 1421–1436.
(b) Eaton, P. E.; Gilardi, R. L.; Zhang, M.-X. AdV. Mater. 2000, 12, 1143–
1148. (c) Cohi, S.-Y.; Eaton, P. E.; Hollenberg, P. F.; Liu, K. E.; Lippard,
S. J.; Newcomb, M.; Putt, D. A.; Upadhyaya, S. P.; Xiong, Y. J. Am. Chem.
Soc. 1996, 118, 6547–6555. See references therein.
(7) Smitt, R. J.; Botaro, J. C.; Penwell, P. E. Report no. AD-A217147,
1989; SRI Int., Menlo Park, CA; Chem. Abstr. 1991, 114, 172.
(8) Glukhovtsev, M. N.; Canepa, C.; Bach, R. D. J. Am. Chem. Soc.
1998, 120, 10528–10533
.
Org. Lett., Vol. 11, No. 16, 2009
3575

tions,8-10 including those reported below. These support the
view that negative charge is being transferred to the dioxirane
from cubane, thus leaving behind partial positive charge
distributed over the proximal hydrocarbon C-H centers.
As expected, due to the electron-withdrawing OH group
of cubanol (6), further TFDO “electrophilic” hydroxylation
to the diol 7 is slower under comparable conditions (eq 2),
as verified by the kinetic data reported in Table 1 (entries 2
and 5).
Actually, our initial B3LPY calculations, performed using
the parent dioxirane H2CO2 in lieu of TFDO as the oxidant,
failed to predict exclusive functionalization at C-4 of cubanol
(6). Further work is in order to clarify the origin of this
peculiar lack of congruence between computations and the
experimental result.
In an attempt to find if s-character in the tertiary C-H
bonds undergoing O-insertion is important, we determined
the dioxirane oxidation rate of cubane (4) in comparison with
that of two reference substrates, namely, adamantane (2)
(purely sp3 C-H) and cyclopropane (5) (C-H quasi sp2)
(Table 1).
Because of the high reactivity of TFDO and of the
volatility of most substrates (especially 5), kinetic runs were
performed most conveniently at subambient temperatures.
The data in Table 1 establish that dioxirane 1b reactivity
toward the given substrates is adamantane > cubane .
cyclopropane; even correcting for statistical factors, this is
in line with the increasing s character of the C-H bonds,
i.e., respectively, ca. 25, 30, and 33%. Nonetheless, as cubane
is more reactive than cyclopropane by a factor of over 105,
despite similar formal C-H hybridization, something more
must be important. Tentatively, one might envisage that this
small-ring compoundsin reaching a distorted dsp3 transition
state such as that sketched in Figure 2swould face a
considerable increase in endocyclic angle strain.
On closer analysis, we found that in fact no appreciable
amount of cyclopropanol, nor of ring-open oxidation prod-
ucts, could be detected by GC-MS or NMR after exposure
of cyclopropane to TFDO. Indeed, the first-order rate
constants measured for TFDO decay in the absence of
cyclopropane are k1 ) 2.8 × 10-6 s-1 at -19 °C and 10.7
× 10-6 s-1 at 0 °C, i.e., practically identical (within (15%)
with the pseudo-first-order rate constants estimated for TFDO
decay in the presence of excess cyclopropane (Table 1, note
c).
Table 1. Rates of Cubane and of Cubanol Monohydroxylation
by TFDO (1b) in Acetone, as Compared to Rates of
Adamantane and Cyclopropane
a
b
no.
substrate
t (°C)
k2 (M-1·s-1
)
1
2
3
4
5
6
7
adamantane (2)
cubane (4)
”
”
cubanol (6)
cyclopropane (5)
”
- 19
- 19
- 10
0
- 19
- 19
0
1.80
0.084
0.150
0.300
0.035
c
< 0.8·10-6
c
< 2.8·10-6
a ( 0.2 °C. b Unless noted otherwise, kinetic data were obtained by
monitoring the decay with time of substrate concentration (calibrated GC).
Initial concentrations of the alkane and of TFDO were kept in the range (8
÷ 4) × 10-3 M and rate constants calculated from integrated second-order
rate-law plots linear to over 50-80% reaction. The data shown are average
values ((8%) from two or more independent runs. c Second-order constants
were estimated as k1/[cyclopropane]o, with k1 values ((10%) from experi-
ments run under pseudo-first-order conditions with [cyclopropane]o in 60-
fold excess over [1b]o (0.04 M); the decrease of TFDO concentration with
time was followed spectrophotometrically by monitoring the absorbance at
333 nm; with [5]o ) 2.43 M and [TFDO]o ) 0.04 M, the averaged first-
order rate constants were k1 ) 1.9 × 10-6 s-1 and 6.8 × 10-6 s-1
respectively, at -19 and 0 °C.
,
It seems unlikely, however, that distance from the electron-
withdrawing OH group would be sufficient to account for
the fact that the further hydroxylation of cubanol (6) to the
diol 7 is exclusiVely at C-4. It would be easier to understand
had this second hydroxylation occurred instead at C-3; the
ab intio 6-31G* calculations (including electron correlation)
by Borden and Hrovat11a on cubyl cation point to efficient
charge delocalization to the C-2 and C-4 positions transmitted
via the strained, p-rich cubane CC bonds,11 in agreement
with the results of Eaton et al. on the relative rates of
solvolysis of C-4 substituted cubyl tosylates.11b
As shown in Figure 3 and Table 2, the markedly higher
reactivity of cubane over cyclopropane with dioxiranes
TFDO (1b) and DDO (1a) is in agreement with ab initio
calculations for the gas phase and in solution at 25 °C (cf:
entry 1 and 2 with 5 and 6, Table 2).12
On the basis of the ∆∆G298+ values from data in Table 2,
one could estimate a relative rate of cubane/cyclopropane
of over 1011 at 25 °C in acetone!
In the cyclopropane case, it should be mentioned that the
possibility of dioxirane attack at the C-C bond (TS 3, Figure
3) can not be discounted a priori. However, inspection of
data collected in Table 2 points out that this avenue also
continues to be largely disfavored with respect to cubane
oxidation (entries 1 and 2).12
On the basis of these computational and experimental
results, formation of a positive charge at C-4 of 1-cubanol
(6) would be expected to be disfavored. Therefore, the
exclusive functionalization of cubanol at C-4 argues against
a transition structure in which much carbocationic character
is developed at this carbon.
In line with these results, one should recall that oxidative
scission of the cyclopropane ring was never observed when
a series of alkyl cyclopropanes1 or of some polycyclic
hydrocarbons encompassing a cyclopropane ring5 were made
to react with the powerful TFDO. It seems that the only
(9) (a) Freccero, M.; Gandolfi, R.; Sarzi-Amade´, M.; Rastelli, J. Org.
Chem. 2003, 68, 811–823. (b) Du, X. H.; Houk, K. N. J. Org. Chem. 1998,
63, 6480–6483. (c) Shustov, G. V.; Rauk, A. J. Org. Chem. 1998, 63, 5413–
5422
.
(10) For a comprehensive review of dioxirane computations, see:Bach,
R. D. In The Chemistry of Peroxides; Patai, S., Ed.; Wiley: New York,
2006; Vol 2, Chapter 1. See references therein.
(11) (a) Hrovat, D. A.; Borden, W. T. J. Am. Chem. Soc. 1990, 112,
3227–3228. (b) Eaton, P. E.; Zhou, J. P. J. Am. Chem. Soc. 1992, 114,
3118–3120.
(12) For both TFDO and DDO oxidation, B3LYP/6-311+G(2d,p) data
are compared with those at the B3LYP/6-31G(d) level in the Supporting
Information.
3576
Org. Lett., Vol. 11, No. 16, 2009

Table 2. Calculated [(B3LYP/6-311+G(2d,p)] Energy Barriers
and Activation Parameters (kcal·mol-1) for the
Oxyfunctionalization of Cubane and of Cyclopropane with
TFDO and with DDO12
∆E+ ∆H298
∆G298
+
+
no.
1
reaction
phase
cubane (4) + TFDO
(TS1)
gas
18.91 16.59
22.47
14.96
33.18
28.24
2
3
4
”
acetone 11.40
gas 29.59 27.55
acetone 24.65 22.61
9.08
cubane (4) + DDO
”
cycloprop.(5) + TFDO
5
6
7
8
(TS2)
gas
acetone 26.16 23.79
37.23 34.95
acetone 34.61 32.33
29.82 27.45
33.99
30.33
41.22
38.60
”
cycloprop. (5) + DDO gas
”
cycloprop (5) + TFDO
9
10
(TS3)
gas
30.35 30.67
38.25
32.85
”
acetone 24.95 25.27
The direct TFDO hydroxylation of cubane in high yield
and, if desired, a position-specific second hydroxylation (also
high yield as reported here) are especially valuable since the
synthesis of cubanol (6) and of cubane-1,4-diol (7) otherwise
requires far more laborious procedures.7,16
Figure 3. Optimized [B3LYP/6-311+G(2d,p)] transition structures
for the oxyfunctionalization of cubane and of cyclopropane with
(TFDO). Bond distances in Å; CT ) charge transfer.
exception reported so far is the DDO oxidation of the
unstable 1,3-didehydroadamantane, which presents a highly
strained cyclopropane C-C bond having a distinct biradical
character.13
The calculations do corroborate the experimental finding
that DDO is unsuitable for the oxidation of cubane (Table
2, entries 3 and 4); indeed, the energy barrier opposing the
DDO reaction is estimated to be some 13 kcal·mol-1 higher
than that for TFDO in acetone. The far more powerful TFDO
(1b) allows one to carry out the oxygenations of cubane
efficiently.4
Acknowledgment. This work was sponsored by the
University of Bari and by the University of Pavia (Italy).
Thanks are also due to the National Research Council of
Italy (CNR, Rome) and the Ministry of Education of Italy
(MIUR) for partial financial support. We wish to dedicate
this article to the memory of Prof. A. M. Tamburro
(University of Basilicata, Italy), who prematurely passed
away on June 23, 2009.
Supporting Information Available: General experimental
procedures and compound characterization data. Sample
kinetics runs. Energy, imaginary frequency, Cartesian co-
ordinates, and activation parameters for all TSs at the B3LYP/
6-311+G(2d,p) level for TFDO and DDO reactions. This
material is available free of charge via the Internet at
http://pubs.acs.org.
As observed in several cases, the hydroxylation of
hydrocarbon C-H bonds using DDO can be inconveniently
slow, so that the radical decomposition of the oxidant might
be triggered.14,15
(13) Fokin, A. A.; Tkachenko, B. A.; Korshunov, O. I.; Gunchenko,
P. A.; Schreiner, P. R. J. Am. Chem. Soc. 2001, 123, 11248–11252.
(14) Curci, R.; Dinoi, A.; Fusco, C.; Lillo, M. A. Tetrahedron Lett. 1996,
OL901302D
37, 249–252
(15) Simakov, P. A.; Choi, S.-Y.; Newcomb, M. Tetrahedron Lett. 1998,
39, 8187–8190
.
(16) Choi, S.-Y.; Eaton, P. E.; Hollenberg, P. F.; Liu, K. E.; Lippard,
S. J.; Newcomb, M.; Putt, D. A.; Upadhyaya, S. P.; Xiong, Y. J. Am. Chem.
Soc. 1996, 118, 6547–6555.
.
Org. Lett., Vol. 11, No. 16, 2009
3577