Experimental and computational bridgehead C-H bond dissociation enthalpies

Article abstract of DOI:10.1021/jo202519w

Bridgehead C-H bond dissociation enthalpies of 105.7 ± 2.0, 102.9 ± 1.7, and 102.4 ± 1.9 kcal mol^-1 for bicyclo[2.2.1]heptane, bicyclo[2.2.2]octane, and adamantane, respectively, were determined in the gas phase by making use of a thermodynamic cycle (i.e., BDE(R-H) = ΔH°_acid(H-X) - IE(H^?) + EA(X^?)). These results are in good accord with high-level G3 theory calculations, and the experimental values along with G3 predictions for bicyclo[1.1.1]pentane, bicyclo[2.1.1]hexane, bicyclo[3.1.1]heptane, and bicyclo[4.2.1]nonane were found to correlate with the flexibility of the ring system. Rare examples of alkyl anions in the gas phase are also provided.

Full text of DOI:10.1021/jo202519w

Article
pubs.acs.org/joc
Experimental and Computational Bridgehead C−H Bond Dissociation
Enthalpies
Alireza Fattahi,*,† Lev Lis,‡ Zahra A. Tehrani,† Sudha S. Marimanikkuppam,‡ and Steven R. Kass*,‡
†Department of Chemistry, Sharif University of Technology, Tehran, Iran
‡Department of Chemistry, University of Minnesota, Minneapolis, Minnesota 55455, United States
S
* Supporting Information
ABSTRACT: Bridgehead C−H bond dissociation enthalpies of 105.7 2.0, 102.9 1.7, and
102.4
1.9 kcal mol−1 for bicyclo[2.2.1]heptane, bicyclo[2.2.2]octane, and adamantane,
respectively, were determined in the gas phase by making use of a thermodynamic cycle (i.e.,
BDE(R−H) = ΔH°acid(H−X) − IE(H·) + EA(X·)). These results are in good accord with
high-level G3 theory calculations, and the experimental values along with G3 predictions for
bicyclo[1.1.1]pentane, bicyclo[2.1.1]hexane, bicyclo[3.1.1]heptane, and bicyclo[4.2.1]nonane
were found to correlate with the flexibility of the ring system. Rare examples of alkyl anions in
the gas phase are also provided.
of these values comes from the most reliable procedure (i.e., a
INTRODUCTION
■
B3LYP/6-311++G(2df,p)//B3LYP/6-31(d) extrapolation to
give a G3B3 prediction).7b,c For 3H, the experimental C−H
BDE is also not well established, and values spanning from 92.2
to 99.6 kcal mol−1 have been reported.1g,5b,8 The most recent
experimental determination of the bond energy was carried out
in 1998 and gave a value of 96.3 kcal mol−1, whereas a G3MP2
study three years later gave a prediction of 100.4 kcal mol−1.
This computation suggests that the actual bond energy maybe
at the high end of the experimentally reported range.9
Reactive intermediates formally derived from rigid hydro-
carbons have been extensively studied because their lack of
structural flexibility provides an ideal platform for probing
geometry-dependent phenomena (e.g., orbital interactions,
substituent effects, and reactivity), and their derivatives, while
posing many synthetic challenges, are valuable in many practical
contexts.1 For example, adamantane-containing compounds
have been used in the synthesis of polymers for electronic
displays and in the treatment of influenza, HIV-1, leukemia, and
Parkinson’s and Alzheimer’s diseases.2 Likewise, bicyclo[2.2.1]-
heptanes and bicyclo[2.2.2]octanes are found in naturally
occurring compounds and have been used to treat depression,
cocaine abuse, and malaria.3
The bridgehead radicals of these rigid hydrocarbons have
been generated by the thermolysis of peresters and azoalkanes
as well as from tin hydride reductions of the corresponding
halides.4 On the basis of these studies, it is generally accepted
that the 1-bicyclo[2.2.1]heptyl (1r), 1-bicyclo[2.2.2]octyl (2r)
and 1-adamantyl (3r) radicals are harder to form than the tert-
butyl radical because they are thermodynamically less stable
(i.e., BDE (CH3)3C-H < 3-H < 2-H < 1-H).5 The C−H bond
dissociation enthalpies of bicyclo[2.2.1]heptane (1H),
bicyclo[2.2.2]octane (2H), and adamantane (3H), however,
are uncertain despite considerable effort to determine these
quantities.
Two decades ago, Walton noted in his excellent review
article on bridgehead radicals that bridgehead C−H bond
energies are extremely valuable but not well established or
generally available.1g This situation has not changed signifi-
cantly in the intervening years. Consequently, in this study we
report the bridgehead C−H BDEs of 1H−3H. These values
were determined from the gas-phase acidities (ΔH°acid) of these
three hydrocarbons and the electron affinities (EA) of their
bridgehead radicals by combining them in a thermodynamic
cycle to afford the corresponding bond dissociation enthalpies
·
(eq 1); IE(H) stands for the ionization energy of hydrogen
atom and is well-known to be 313.6 kcal mol−1.10 The
experimental results were supplemented with G3 computations
on 1H−3H and a number of related bicyclic compounds.
·
·
BDE(HX) = ΔH° (HX) − IE(H ) + EA(X )
(1)
acid
Experimental measurements of the bond dissociation
enthalpies for 1H range from 96.7 to 101.8 kcal mol−1,1g,5b,6
but recent computations give predictions spanning from 102 to
112 kcal mol−1.7 The most reliable theoretical methods (CBS-
QB3, G3B3, and G3) indicate that the BDE is 107−108 kcal
mol−1,7c,d but these values, along with other computational
predictions, are in poor accord with the experimental data. A
similar situation applies for 2H in that BDEs of 93.2 and 97.7
kcal mol−1 have been reported,1g,5b whereas computational
predictions range from 98.3 to 101.9 kcal mol−1, and the largest
RESULTS AND DISCUSSION
Bicyclo[2.2.1]heptane (1H). Fluoride ion was used to
deprotonate bicyclo[2.2.1]heptane-1-carboxylic acid (1CO2H)
in the gas phase, and the resulting conjugate base was
transferred to the second cell of a dual cell Fourier transform
mass spectrometer (FTMS) where it was fragmented by
■
Received: December 9, 2011
Published: January 16, 2012
© 2012 American Chemical Society
1909
dx.doi.org/10.1021/jo202519w | J. Org. Chem. 2012, 77, 1909−1914

The Journal of Organic Chemistry
Article
energetic collisions with argon gas (i.e., collision-induced
dissociation (CID)). It was anticipated that the bridgehead
anion of bicyclo[2.2.1]heptane (i.e., 1a) would result from this
process since other alkyl anions have been produced in this
way.11 To test this expectation, the decarboxylated (M −
CO2)− ion was isolated and allowed to react with a variety of
neutral reagents. In accord with the known extreme basicity of
alkyl anions,12 the (M − CO2)− ion readily deprotonated water
−1
ΔH° (RH) = (−0.613 0.022 kcal mol )
acid
·
× EA(R) + (414.7 0.7)
−1
(2)
kcal mol
Combining ΔH°acid(1H) and EA(1r) along with the well-
known ionization energy of hydrogen into the thermodynamic
cycle given in eq 1 leads to 105.7
2.0 kcal mol−1 for the
(ΔH°acid = 390.27 0.02 kcal mol−1) and ammonia (ΔH°acid
=
bridgehead C−H bond dissociation enthalpy of bicyclo[2.2.1]-
heptane. This value is well reproduced by B3LYP/6-311+
+G(2d,p), CBS-QB3, G3B3, and G3 predictions of 104.0,
107.3, 107.6, and 107.9 kcal mol−1, respectively, and is also
403.4 0.1 kcal mol−1).13 This places its conjugate acid at the
high end of the gas-phase acidity scale (i.e., ΔH°acid > 403.4
0.1 kcal mol−1), which is as expected for 1H. As a result, an
alternative approach to equilibrium and bracketing measure-
ments is needed for determining its acidity. This was
accomplished by measuring the electron affinity (EA) of the
bridgehead radical (1r) and taking advantage of a recently
consistent with Allinger’s estimate of 102.1 kcal mol−1.7
A
MP2/6-31+G(d) BDE of 112.0 kcal mol−1, however, is clearly
too large.7b
·
Experimental determinations for the bridgehead C−H bond
energy of 1H of 96.7 2.5 and 99.4 kcal mol−1 were originally
reported in 1970 and 1971.5b,6 The first of these energies was
determined by reacting 1-iodobicyclo[2.2.1]heptane (1I) with
HI to afford bicyclo[2.2.1]heptane and I2, but as noted in
Walton’s 1992 review,1g measurements of this type sub-
sequently were found to underestimate bond dissociation
enthalpies by ∼3 kcal mol−1 for tertiary C−H bonds. If one
arbitrarily employs a 3.0 kcal mol−1 correction for 1H as done
by Walton, then a revised estimate of 99.7 kcal mol−1 is
obtained. The second bond energy was determined by
comparing the polar-effect corrected relative iodine abstraction
rates of phenyl radical with a series of alkyl iodides (RI)
including 1I in which the reference compounds have known
R−H bond dissociation enthalpies. Since updated C−H BDEs
were available for the standards, a revised bond energy to 101.8
kcal mol−1 was given by Walton. Both of these changes moved
the original BDE determinations upward, but the resulting
values are still ∼5 kcal mol−1 too small. In the latter instance
(phenyl radical abstraction of iodine), this appears to be a
systematic error since the BDEs for several other compounds
using this method are also ∼5 kcal mol−1 too weak; for further
details, see below.
Bicyclo[2.2.2]octane (2H). We attempted to generate the
bridgehead anion of 2H (i.e., 2a) in a similar manner as was
done for 1a, but in this instance, as happens from time to time,
a rearrangement to an unknown ion took place upon
fragmentation of the bridgehead carboxylate anion. This was
readily apparent in that the (M − CO2)− anion did not
deprotonate NH3 or H2O nor did it undergo hydrogen−
deuterium exchange with ND3. These observations are
inconsistent with all known alkyl anions and the G3 predicted
acidity for 2H of 413.4 kcal mol−1. A less basic species such as
the deprotonated allylic anion of vinylcyclohexane is a plausible
structure, but this ion was not explored further. An alternative
approach for the preparation of 2a was carried out instead by
reacting fluoride anion with 1-trimethylsilylbicyclo[2.2.2]octane
(2SiMe3). This methodology has been successfully employed in
a similar case (bicyclo[1.1.1]pentane) where the loss of carbon
dioxide from a carboxylate anion led to a rearrangement.17 In
this instance, however, even though 2SiMe3 did give the
desired (M − TMS)− anion, it was produced in very small
amounts. The resulting ion appears to be extremely basic which
is consistent with the formation of 2a, but an insufficient signal
kept us from collisionally cooling and reliably probing the
reactivity of this ion. Our data does suggest, nevertheless, that
2a is bound with respect to electron loss as originally predicted
reported correlation between ΔH°acid(RH) and EA(R) for
hydrocarbons whose conjugate bases are localized anions.14
Sulfur dioxide, chloropentafluorobenzene, and carbon
disulfide were found to react with 1a via electron transfer,
whereas carbonyl sulfide, oxygen, and nitrous oxide, all of which
have smaller electron affinities, did not (Table 1).15 These
Table 1. Experimental Determination of the Electron
Affinity of 1-Bicyclo[2.2.1]heptyl Radical (1r)
a
ref compd
EA (eV)
electron transfer
SO2
1.107 0.008
0.82 0.11
0.58 0.05
0.46 0.20
0.448 0.006
0.22 0.10
yes
yes
C6F5Cl
CS2
b
yes
c
COS
O2
no
no
no
N2O
a
b
Experimental values come from ref 10. A small amount of sulfur-
c
atom transfer was also observed. Sulfur-atom transfer and adduct
formation were observed.
results indicate that the electron affinity of 1r is between that
for O2 and CS2 or 0.51 0.07 eV (i.e., 11.8 1.6 kcal mol−1).
This value is in accord with other alkyl anions which are known
to be weakly bound.12 It is also in between the measured
electron affinities for dodecahedryl (0.17 0.10 eV)16 and 3-
tert-butyl-1-bicyclo[1.1.1]pentyl (0.64 0.14 eV)17 radicals as
one might expect. A high level G3 theory18 calculation gives a
298 K value of 0.453 eV (10.44 kcal mol−1), and this too is in
excellent agreement with the experimental determination.
Substitution of the experimentally determined electron
affinity of 1r into eq 2 affords ΔH°acid(1H) = 407.5
1.2
kcal mol−1; the given uncertainty is the statistical value, but a
more conservative and arbitrary estimate is 2−3 kcal mol−1.
This result is in reasonable accord with a G3 theory prediction
of 411.9 kcal mol−1 but is smaller than a preliminary estimate
for this quantity of 414 kcal mol−1 using the DePuy kinetic
method (see below for a detailed description of this
procedure).19 The latter difference is not necessarily surprising,
however, since 1-trimethylsilylbicyclo[2.2.1]heptane was used
for the kinetic acidity measurement, and it was subsequently
discovered that when the two substituents attached to silicon
(1-bicyclo[2.2.1]heptyl and CH3 in this case) differ sterically
then acidities that are numerically too large can result. For
example, cubane and dodecahedrane gave acidities that were
too weak by 9 and 10 kcal mol−1, respectively.16,20
1910
dx.doi.org/10.1021/jo202519w | J. Org. Chem. 2012, 77, 1909−1914

The Journal of Organic Chemistry
Article
by Sauer and subsequently indicated by the G3 computations
reported herein.7b
To determine the bridgehead acidity of 2H, 1-
triphenylsilylbicyclo[2.2.2]octane was prepared from its corre-
sponding bromide (eq 3), and the triphenylsilyl derivative,
Table 2. DePuy Kinetic Acidity Measurement Data for 1-
Triphenylsilybicyclo[2.2.2]octane and 1-
Triphenylsilyladamantane
Ph3SiO−/ Ph Si(R)O−
ln (3 ×
2a
b
RSiPh3
R = CH3
(ratio)
ratio)
ΔH°acid(RH)
0.010 0.002
0.045 0.004
0.042 0.004
1.00
−3.49
−2.01
−2.08
0.00
416.7 0.7
411.5 2.0
409.4 0.6
401.7 0.5
394.2 1.2
411.9 0.4
c-C3H5
C2H3
C6H5
1-naphthyl
3.32 0.22
0.030 0.003
2.30
rather than the corresponding trimethylsilane, was used in a
DePuy kinetic method determination.12a,b In this approach, a
series of reference triphenylsilyl derivatives RSiPh3 were used in
which RH has a known gas-phase acidity. By allowing these
compounds to react with OH−, two product ions (Ph3SiO− and
RPh2SiO−) were produced (eq 4). A plot of the natural
1-bicyclo[2.2.2]
octyl
−2.40
1-adamantyl
0.021 0.003
−2.79
413.4 0.6
a
b
Observed ratio was isotopically corrected. Values are in kcal mol−1.
G3 value of 0.148 eV (3.42 kcal mol−1). By combining
ΔH°acid(2H) and EA(2r) via eq 1, one obtains 102.9 1.7 kcal
mol−1 for the bridgehead C−H bond of 2H. This value is also
in excellent agreement with the G3 estimate for this quantity of
102.4 kcal mol−1 but is larger than the B3LYP/6-311++G(2d,p)
and MP2/6-31+G(d) predictions of 98.3 and 99.9 kcal mol−1,
respectively.7b
ΔH°acid(RH) = −3.90 ln(3Ph3SiO−
logarithm of the statistically and isotopically corrected product
ion ratio [ln(3Ph3SiO−/Ph2Si(R)O−)] versus ΔH°acid(RH)
afforded a linear calibration line that can be used to obtain
the acidity of an unknown acid. This methodology has been
used most commonly with trimethylsilyl derivatives, but several
large substrates (i.e., cubane and dodecahedrane) gave
anamolously low acidities and this was attributed to the large
steric difference between the two silicon substituents.16,20 That
is, pentacoordinate silconate 4a is expected and was computed
to be more stable than 4b when R is large (Figure 1), but the
/R(Ph)2SiO−) + 402.6, r2 = 0.987
(5)
An experimental value for 2H of 93.2 kcal mol−1 for the
bridgehead C−H bond dissociation enthalpy was originally
obtained by Danen using the phenyl radical abstraction method
noted above.5b This bond energy was subsequently revised by
Walton to 97.7 kcal mol−1 using newer reference data,1g but it is
still ∼5 kcal mol−1 too small as is the case for bicyclo[2.2.1]-
heptane.
Adamantane (3H). Both 1-trimethylsilyl- and 1-triphenyl-
silyladamantane were synthesized from the commercially
available bridgehead bromide to determine the acidity of
adamantane and explore the effect of methyl vs phenyl
substitution in the trialkylsilanes. To accomplish this, a five-
point reference line was obtained using (CH3)3SiR, where R =
CH2CH3, CH3, c-C3H5, C2H3, and C6H5 (Table 3, eq 6).
Figure 1. Trimethylsilyl pentacoordinate anion intermediates formed
in the DePuy kinetic method.
Table 3. DePuy Kinetic Acidity Measurement Data for 1-
Trimethylsilyladamantane
latter species is the one that loses RH more readily. This leads
to a reduced amount of (CH3)3SiO− and a weaker acidity than
would be obtained otherwise. If one replaces all three methyl
groups in the tetraalkylsilane with larger phenyl rings, then the
difference in stability between the two pentacoordinate anions
should diminish and the resulting product ion ratio is more
likely to reflect the thermodynamic acidity of RH. This was
found to be the case for dodecahedrane, and it is for this reason
that 1-triphenylsilylbicyclo[2.2.2]octane (2SiPh3) was em-
ployed in this work.
Me3SiO−/ Me2Si(R)O−
ln (3 ×
ratio)
a
b
RSiMe3
(ratio)
ΔH°acid(RH)
R = CH3CH2
CH3
0.13 0.01
1.00
−0.95
0.00
1.08
1.99
3.29
1.04
420.1 2.0
416.7 0.7
411.5 2.0
409.4 0.6
401.7 0.5
411.9 0.3
c-C3H5
0.98 0.06
2.43 0.08
8.93 0.62
0.95 0.06
C2H3
C6H5
1-adamantyl
a
b
Observed ratio was isotopically corrected. Values are in kcal mol−1.
A five point calibration line was constructed using Ph3SiR
compounds, where R = methyl, cyclopropyl, vinyl, phenyl, and
1-naphthyl (Table 2 and eq 5). The isotopically corrected ratio
Subsequent measurements on both adamantane derivatives
afforded ΔH°acid(3H) = 413.4 0.6 (SiPh3) and 411.9 0.3
(SiMe3) kcal mol−1 which are similar enough that both results
were averaged to give our recommended value of 412.7 0.8
kcal mol−1. This quantity is well reproduced by the G3 estimate
of 411.4 kcal mol−1. It leads also to EA(3r) = 0.14 0.08 eV
for 2SiPh3 was found to be 0.030
ΔH°acid(2H) = 411.9
excellent accord with the G3 theory prediction of 413.4 kcal
mol−1, and a previous estimate for this quantity of 414 kcal
mol−1 using the trimethylsilyl derivative.19 It also leads to
0.003, which leads to
0.4 kcal mol−1. This finding is in
(3.3
1.8 kcal mol−1) via eq 2, and this too is accurately
EA(2r) = 0.20
0.07 eV (4.6
1.6 kcal mol−1) via the
predicted by G3 theory which gives a value of 2.1 kcal mol−1 for
this quantity. Both the experimental and the G3 electron
correlation given in eq 2, and this too is well reproduced by the
1911
dx.doi.org/10.1021/jo202519w | J. Org. Chem. 2012, 77, 1909−1914

The Journal of Organic Chemistry
Article
affinities for 3r are numerically small values which is expected
for an alkyl radical, but they are also larger than zero and this
indicates that the anion is stable with respect to electron loss.
This contrasts with a conclusion based upon B3LYP and other
DFT functionals that indicated that 1-adamantyl anion is an
unbound species with respect to its radical.21 The sign of the
electron affinity (positive or negative), consequently, can be
viewed as an open question.
hybridization,22,24 and the difference in the out-of-plane
distance d between the bridgehead carbon and the plane
defined by the 3 carbon atoms attached to it in the hydrocarbon
and its radical was used as an indicator of the rigidity of the ring
systems; more commonly employed out-of-plane angles were
not used because most of the compounds in Table 4 do not
have C3-rotation axes and consequently more than one such
angle can be defined. A plot of the BDEs versus the 13C−1H
coupling constants (Figure S1 in the Supporting Information)
shows considerable scatter of the data points and only a loose
correlation. In contrast, there is a good relationship between
the bond dissociation enthalpies and the ability of the radicals
to flatten out as given by the d values for the different ring
systems (Figure 2).25 This correlation is consistent with the
−
ΔH° (RH) = −4.19 ln(3Me SiO
acid
3
2
−
/R(Me) SiO ) + 416.3, r = 0.985
(6)
2
Our data leads to BDE(3H) = 102.4 1.9 kcal mol−1 for the
tertiary C−H bond, and this determination is in excellent
accord with the G3 prediction of 100.6 kcal mol−1. It is larger,
however, than a recent appearance energy measurement of 96.3
kcal mol−1,8b and it is ∼5 kcal mol−1 bigger than Danen’s value
of 97.0 kcal mol−1.5b The latter discrepancy is consistent with
the experimental and computational findings for bicyclo[2.2.1]-
heptane and bicyclo[2.2.2]octane, however, and suggests that
all of the bond energies determined by phenyl abstraction of
iodine are 5 kcal mol−1 too small. Our bond dissociation
enthalpy, on the other hand, is in excellent accord with
Beauchamp’s measurement of 100.2
1.3 kcal mol−1 by
photoelectron spectroscopy.8a
To put our bridgehead C−H bond dissociation enthalpies in
a broader perspective, values for bicyclo[1.1.1]pentane (5H),
bicyclo[2.1.1]hexane (6H), bicyclo[3.1.1]heptane (7H), and
bicyclo[4.2.1]nonane (8H) were computed using G3 theory.
All of these results are summarized in Table 4, and the resulting
Figure 2. Plot of bridgehead bond dissociation enthalpies for 1H−3H
and 5H−8H versus d, a measure of the flexibility of the ring system as
described in the text.
commonly held assumption that the more rigid the ring system
and the more pyramidal the radical the larger the bridgehead
C−H bond dissociation enthalpy.1g The radical stabilization
energies (RSEs, i.e., BDE(RH) − BDE((CH3)3C−H)26 also
track with the flexibility of the ring system (i.e., d), but the
RSEs and BDEs only cover a 12 kcal mol−1 range.
Table 4. Experimental and Computed (G3) Bridgehead
Bond Dissociation Enthalpies and Radical Stabilization
Energies for Bicyclic Hydrocarbons 1H−3H and 5H−8H
along with Heteronuclear 13C−1H Coupling Constants and a
a
Measure of the Ring System's Flexibility
M06-2X/aug-cc-pVDZ computations27,28 were carried out
on the tert-butyl radical. When it is distorted from its
equilibrium structure by carrying out partial geometry
optimizations, the energy was found to change by less than
1.5 kcal mol−1 when the same range in d spanned by the bicyclic
hydrocarbons is used (i.e., 0.181 − 0.034 or 0.147 Å). This
indicates that distorting (bending) the tert-butyl radical requires
less energy than the bicyclic bridgehead radicals studied herein.
BDE (kcal mol−1)
J
(13C−1H, d (RH
b
c
d
̇
compd
Hz)
− R)
expt
G3
RSE
1H
2H
3H
5H
6H
7H
8H
141.0
134.5
133.4
167.8
151.8
144.9
−
0.072
0.097
0.105
0.044
0.034
0.083
0.181
105.7 2.0
102.9 1.7
102.4 1.9
109.7 3.3
107.9
102.4
100.6
106.5
109.1
103.8
97.4
10.5 (9.2 2.0)
5.0 (6.4 1.7)
3.2 (5.9 1.9)
9.1
e
11.7
6.4
CONCLUSIONS
■
0.0
It appears that bridgehead anions from small strained ring
systems such as bicyclo[1.1.1]pentane to larger and more
flexible hydrocarbons such as bicyclo[2.2.2]octane and possibly
even adamantane are stable with respect to electron detach-
ment. This is in accord with weakly bound alkyl anions such as
CH3− and (CH3)3CCH2− but contrasts with CH3CH2−,
(CH3)2CH−, and (CH3)3C− which are unstable with respect
to spontaneous electron detachment (i.e., the corresponding
radicals are more stable than their anions). The electron affinity
of 1-bicyclo[2.2.1]heptyl radical was large enough to be
a
b
Bond and radical stabilization energies are given in kcal mol−1. All
values come from ref 22. The differences in the distances (in
c
angstroms) between the bridgehead carbon and the plane defined by
the three carbon atoms attached to it for the hydrocarbons and their
d
corresponding radicals are given by d. RSE = BDE(RH) −
BDE((CH3)3C−H). The computed G3 energies are given first and
the experimental values follow in parentheses. BDE ((CH3)3C−H) =
96.5 0.4 kcal mol−1 (expt, see ref 23) and 96.6 kcal mol−1 (G3, ref
e
7c). This value is for the 1-tert-butyl derivative (ref 17).
heats of formation and strain energies are given in Table S1 in
the Supporting Information. One might expect these bond
energies to correlate with the hybridization of the bridgehead
C−H bonds or some measure of the flexibility of the ring
systems. Heteronuclear 13C−1H coupling constants for the
bridgehead C−H bonds were used as a measure of their
measured by a titration (bracketing) method (11.8
1.6
[expt] and 10.4 [G3] kcal mol−1), whereas the 1-bicyclo[2.2.2]-
octyl and 1-adamantyl anions were too fragile for us to
characterize them in the same way (EA(2r) = 4.6 1.6 [expt]
and 3.4 [G3] kcal mol−1, and EA(3r) = 3.3 1.8 [expt] and 2.1
[G3] kcal mol−1). We therefore measured the acidities of their
1912
dx.doi.org/10.1021/jo202519w | J. Org. Chem. 2012, 77, 1909−1914

The Journal of Organic Chemistry
Article
acidified with 1.0 mL of 10% HCl and diluted at the same time with 7
mL of CH2Cl2. After separation of the organic layer, the aqueous
solution was extracted twice with 2 mL portions of CH2Cl2. The
combined organic material was washed three times with 1 mL of brine,
dried over MgSO4, and concentrated under aspirator pressure with a
rotary evaporator. Purification of the residue by flash chromatography
on a silica gel column with hexane as the eluent afforded 38 mg (35%)
of 1-triphenylsilylbicyclo[2.2.2]octane with a melting point of 149−
conjugate acids using a well-established kinetic method that
correlates ΔH°acid(R′H) to the product ratio of the reaction
between hydroxide ion and a series of substituted silanes (i.e.,
HO− + R3SiR′ → R3SiO− + R′H and R2Si(R′)O− + RH). Since it
was important to use a large R group for dodecahedrane and
cubane, a phenyl group was employed in this work. This may
play a role for bicyclo[2.2.1]heptane (407.5 (R = SiPh3) vs 414
(R = SiMe3) kcal mol−1), but appears to be a small issue, if one
at all, for bicyclo[2.2.2]octane (411.9 (R = SiPh3) vs 414 (R =
SiMe3)) and is unimportant for adamantane. The resulting
acidities and the one electron affinity were substituted into a
1
151 °C: H NMR (500 MHz, CDCl3) δ 1.53 (m, 6H), 1.57 (sextet,
1H), 1.82 (m, 6H), 7.34−7.38 (m, 6H), 7.39−7.43 (m, 3H), 7.54−
7.57 (m, 6H); 13C NMR (75 MHz, CDCl3) δ 18.8, 24.1, 26.1, 27.8,
127.8, 129.3, 134.9, 137.0; HRMS-CI (NH3) calcd for C26H32SiN (M
+ NH4)+ 386.2304, found 386.2296.
1-Triphenylsilyladamantane (3SiPh3). The procedure above
was followed using 215 mg (1.0 mmol) of 1-bromoadamantane to
afford 188 mg (48%) of 3SiPh3 (mp = 167−168 °C): 1H NMR (300
MHz, CDCl3) δ 1.78 (m, 6H), 1.88 (m, 3H), 2.03 (m, 6H), 7.34−7.47
(m, 9H), 7.59−7.63 (m, 6H); 13C NMR (75 MHz, CDCl3) δ 24.1,
28.0, 37.6, 39.0, 127.8, 129.3, 134.5, 136.9; HRMS (EI) calcd for
C28H30Si (M+) 394.2111, found 394.2100.
·
recently reported correlation between ΔH°acid(RH) and EA(R)
for localized carbanions to obtain the other quantity, and this
enabled us to determine the homolytic bond dissociation
enthalpies for 1H−3H. These bond energies are in accord with
the previously proposed order (CH3)3C−H < 3-H < 2-H < 1-
H and are consistent with high-level G3 computations.
Additional bridgehead bond dissociation enthalpies for
bicyclo[1.1.1]pentane, bicyclo[2.1.1]hexane, bicyclo[3.1.1]-
heptane, and bicyclo[4.2.1]nonane were also calculated using
G3 theory and all 7 BDEs examined in this work were found to
correlate with the ability of the ring system to flatten out upon
formation of the radical. This intuitively reasonable finding
suggests that bridgehead C−H bond energies are now finally
well established to an accuracy of 2−3 kcal mol−1.
1-Trimethylsilyladamantane (3SiMe3). This substrate was
prepared as previously described.34 It had a melting point of 48−50
35
1
°C and a H NMR spectrum that was the same as in the literature.
Gas-Phase Experiments. A dual cell Fourier transform mass
spectrometer (FTMS) equipped with a 3 T superconducting magnet
and controlled by a Sun workstation running the Odyssey Software 4.2
package or a data system from Ion Spec running IonSpec99 Ver. 7.0
was used in these studies. A solid probe inlet kept at room temperature
was employed to introduce solid samples into the FTMS. For the
DePuy kinetic acidity method, hydroxide was prepared by electron
ionization (6 eV) of water in the analyzer cell. The OH− ions were
transferred into the source cell and after an argon pulse up to a
nominal pressure of ∼10−5 Torr and a subsequent 700 ms delay, all of
the product ions were ejected using a chirp broad band excitation.36
The reaction of thermalized hydroxide with a static pressure of the
neutral species (∼2.0 × 10−8 Torr) was then monitored over time
(0.2−4 s) and the product ion ratio was found to be constant.
Fluoride ion was prepared by electron ionization of CF4 at 7 eV.
The resulting F− was used to deprotonate bicyclo[2.2.1]heptane-1-
carboxylic acid and bicyclo[2.2.2]octane-1-carboxylic acid in the source
cell. The resulting carboxylate anions were transferred to the other cell
where they were thermalized with an argon gas pulse up to a nominal
pressure of ∼10−5 Torr. On-resonance collisionally induced dissoci-
ation (CID) of these ions resulted in the formation of (M − CO2)−
ions. The ion derived from bicyclo[2.2.1]heptane-1-carboxylate was
isolated using a series of chirp and SWIFT excitations37 and its
reactions subsequently were examined as a function of time.
EXPERIMENTAL SECTION
■
1-Bromoadamantane and bicyclo[2.2.2]octane-1-carboxylic acid were
purchased from a commercial supplier and the Florida Center for
Heterocyclic Compounds, respectively, while bicyclo[2.2.1]heptane-1-
carboxylic acid and 1-bromobicyclo[2.2.2]octane were kindly provided
by Prof. William Adcock.29,30
1-Trmethylsilylbicyclo[2.2.2]octane (2SiMe3). A solution of
trimethylsilylsodium was prepared under a stream of Ar by adding 0.36
mL (1.76 mmol) of hexamethyldisilane dropwise to a suspension of
freshly prepared sodium methoxide (from 25 mg of Na and 1.0 mL of
methanol) in 0.85 mL of anhydrous hexamethylphosphortriamide
(HMPA). After the resulting red solution was stirred for 4 h at
ambient temperature, a solution of 0.050 g (0.265 mmol) of 1-
bromobicyclo[2.2.2]octane in 0.25 mL of dry HMPA was added
dropwise over ∼10 min and then was allowed to stand overnight with
stirring. The resulting reaction mixture was quenched with 3 mL of
water and extracted with pentane (3 × 10 mL), and the combined
organic layers were washed with water (3 × 5 mL) and dried over
MgSO4. Removal of the solvent was carried out by distillation at
atmospheric pressure since the product is volatile and vacuum
operations are best avoided.31 Purification of the residue was carried
out with a gas chromatograph equipped with a 16′ long SE-30/
Chromosorb P column held at 140 °C to afford 10 mg (26%) of 1-
trimethylsilylbicyclo[2.2.2]octane (mp 66−67 °C (hexanes), lit.31,32
mp 61−63 °C and 67−68 °C): 1H NMR (300 MHz, CDCl3) δ −0.12
(s, 9H), 1.41−1.52 (m, 13H); 13C NMR (75 MHz, CDCl3) δ −4.6,
15.6, 24.0, 26.0, 26.2.
1-Triphenylsilylbicyclo[2.2.2]octane (2SiPh3). A 30% disper-
sion of sodium in toluene (230 mg, 3.0 mmol) was placed in a round-
bottomed flask, and the solvent was removed under vacuum. The
pressure was restored to an atmosphere with argon and then 2 mL of
dry ether was syringed into the flask. A solution of 221 mg (0.75
mmol) of chlorotriphenylsilane and 19.2 mg (0.15 mmol) of
naphthalene in 1.5 mL of anhydrous ether was rapidly added to the
reaction vessel.33 The resulting mixture was stirred for 1 h before
adding 0.31 mL (1.8 mmol) of HMPA. This solution turned deep
green−black and was stirred for an additional 0.5 h before 56.7 mg
(0.30 mmol) of 1-bromobicyclo[2.2.2]octane in 0.5 mL of ether was
added by syringe over the course of 2 min. After 0.5 h, the reaction was
quenched under argon by the careful dropwise addition of 1.0 mL of
water to react away the excess sodium. The resulting solution was
Computations. All of the calculations were carried out at the
Minnesota Supercomputer Institute for Advanced Computational
Research using the Gaussian suite of programs.38 G3 energies were
computed as previously described in the literature,18 and M06-2X-aug-
cc-pVDZ27,28 partial optimizations on (CH3)3C were carried out in
·
C3v symmetry at fixed out-of-plane angles at the carbon radical center.
ASSOCIATED CONTENT
■
S
* Supporting Information
Computed MP2(full)/6-31G(d) geometries and G3 energies,
heats of formation, and strain energies are provided along with
experimental ΔH°f,298 K values, a plot of bridgehead C−H BDEs
vs 13C−1H coupling constants, the 1H and 13C NMR spectra of
2SiPh3 and 3SiPh3, and the complete citation to ref 38. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: fattahi@sharif.edu, kass@umn.edu.
1913
dx.doi.org/10.1021/jo202519w | J. Org. Chem. 2012, 77, 1909−1914

The Journal of Organic Chemistry
Article
(10) Bartmess, J. E. NIST Chemistry WebBook, NIST Standard
Reference Database Number 6; Mallard, W. G., Linstrom, P. J., Eds.;
National Institute of Standards and Technology: Gaithersburg, MD
(http://webbook.nist.gov).
(11) Graul, S. T.; Squires, R. R. J. Am. Chem. Soc. 1990, 112, 2506−
2516.
(12) (a) DePuy, C. H.; Bierbaum, V. M.; Damrauer, R. J. Am. Chem.
Soc. 1984, 106, 4051−4053. (b) DePuy, C. H.; Gronert, S.; Barlow, S.
E.; Bierbaum, V. M.; Damrauer, R. J. Am. Chem. Soc. 1989, 111, 1968−
1973. (c) Graul, S. T.; Squires, R. R. J. Am. Chem. Soc. 1990, 112,
2517−2529.
(13) Ervin, K. M.; DeTuri, V. F. J. Phys. Chem. A 2002, 106, 9947−
9956.
(14) Fattahi, A.; Kass., S. R. J. Org. Chem. 2004, 69, 9176−9183.
(15) All thermochemical values come from ref 10 unless otherwise
noted.
(16) Broadus, K. M.; Kass, S. R.; Osswald, T.; Prinzbach, H. J. Am.
Chem. Soc. 2000, 122, 10964−10968.
(17) Reed, D. R.; Kass, S. R.; Mondanaro, K. R.; Dailey, W. P. J. Am.
Chem. Soc. 2002, 124, 2790−2795.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Professor William Adcock for kindly providing us
with samples of bicyclo[2.2.1]heptane-1-carboxylic acid and 1-
bromobicyclo[2.2.2]octane. Generous support from the Na-
tional Science Foundation, the Petroleum Research Fund, and
the Minnesota Supercomputer Institute for Advanced Compu-
tational Research are also gratefully acknowledged.
REFERENCES
■
(1) (a) Harman, D. G.; Blanksby, S. J. Org. Biomol. Chem. 2007, 5,
3495−3503. (b) Eguchi, S. Recent Res. Dev. Pure Appl. Chem. 1998, 2
(Part 2), 311−320. (c) Ohga, Y.; Takeuchi, K. Pure Appl. Chem. 1997,
1, 223−240. (d) Takeuchi, K.; Ohga, Y. Bull. Chem. Soc. Jpn. 1996, 69,
833−851. (e) Della, E. W.; Schiesser, C. H. Adv. Carbocation Chem.
1995, 2, 91−122. (f) Mueller, P.; Mareda, J.; Milin, D. J. Phys. Org.
Chem. 1995, 8, 507−528. (g) Walton, J. C. J. Chem. Soc. Rev. 1992, 21,
105−112. (h) Applequist, D. E.; Roberts, J. D. Chem. Rev. 1954, 54,
1065−1089.
(18) Curtiss, L. A.; Raghavachari, K.; Redfern, P. C.; Rassolov, V.;
Pople, J. A. J. Chem. Phys. 1998, 109, 7764−7776.
(19) Damrauer, R. In Selective Hydrocarbon Activation: Principles and
Progress; Davies, J. A., Watson, P. L., Greenberg, A., Liebman, J. F.,
Eds.; VCH Publishers: New York, 1990; pp 535−556.
(20) Hare, M.; Emrick, T.; Eaton, P. E.; Kass, S. R. J. Am. Chem. Soc.
1997, 119, 237−238.
(2) (a) Jeong, H. Y.; Lee, Y. K.; Talaie, A.; Kim, K. M.; Kwon, Y. D.;
Jang, Y. R.; Yoo, K. H.; Choo, D. J.; Jang, J. Thin Solid Films 2002, 417,
171−174. (b) Ruppersberg, J. P.; Fakler, B. P. PCT Int. Appl. 2002, 31.
(c) Spasov, A. A.; Khamidova, T. V.; Bugaeva, L. I.; Morozov, I. S.
Pharm. Chem. J. 2000, 34, 1−7. (d) Mathias, L. J.; Jensen, J. J.;
Reichert, V. R.; Lewis, C. M.; Tullos, G. L. Polym. Prepr. 1995, 36,
741−742. (e) Stotskaya, L. L.; Serbin, A. V.; Munshi, K.; Kozeletskaya,
K. N.; Sominina, A. A.; Kiselev, O. I.; Zaitseva, K. V.; Natochin, Y. V.
Khim.- Farm. Zh. 1995, 29, 19−22. (f) Boukrinskaia, A. G.; Serbin, A.
V.; Bogdan, O. P.; Stotskaya, L. L.; Alymova, I. V.; Klimochkin, Y. N.
PCT Int. Appl. 1995, 32.
(3) (a) Seebacher, W.; Berger, H.; Kaiser, M.; Brun, R.; Saf, R.; Weis,
R. Monatsh. Chem. 2006, 137, 471−482. (b) Hindmarch, I. H.
Psychopharmacol. Clin. Exp. 2001, 16, 203−218. (c) Nelson, J. C. J.
Clin. Psychiatry 1998, 59, 65−68. (d) Owens, M. J.; Morgan, W. N.;
Plott, S. J.; Nemeroff, C. B. J. Pharmacol. Exp. Ther. 1997, 283, 1305−
1322. (e) Martin, S. F.; Assercq, J. M.; Austin, R. E.; Dantanarayana, A.
P.; Fishpaugh, J. R.; Gluchowski, C.; Guinn, D. E.; Hartmann, M.;
Tanaka, T.; Wagner, R.; White, J. B. Tetrahedron 1995, 51, 3455−
3482. (f) Yates, P.; Burnell, D. J.; Freer, V. J.; Sawyer, J. F. Can. J.
Chem. 1987, 65, 69−77. (g) Baldessarini, R. J. Drugs and the
Treatment of Psychiatic Disorders. In The Pharmacological Basis of
Therapeutics, 7th ed.; Goodman, A. G., Gilman, L. S., Rall, T. W.,
Murad, F., Eds.; Macmillan: New York, 1985; p 416.
(4) (a) Engel, P. S. Chem. Rev. 1980, 80, 99−150. (b) Golzke, V.;
Groeger, F.; Oberlinner, A.; Ruechardt, C. Nouv. J. Chim. 1978, 2,
169−178. (c) Ruechardt, C. Angew. Chem., Int. Ed. Engl. 1970, 9, 830−
843. (d) Fort, R. C.; Schleyer, P. v. R. Adv. Alicycl. Chem. 1966, 1,
283−370.
(5) (a) Fort, R. C. Jr.; Hiti, J. J. Org. Chem. 1977, 42, 3968−3970.
(b) Danen, W. C.; Tipton, T. J.; Saunders, D. G. J. Am. Chem. Soc.
1971, 93, 5186−5189. (c) Oberlinner, A.; Ruchardt, C. Tetrahedron
Lett. 1969, 4685−4688. (d) Applequist, D. E.; Kaplan, L. J. Am. Chem.
Soc. 1965, 87, 2194−2200.
(6) O’Neal, H. E.; Bagg, J. W.; Richardson, W. H. Int. J. Chem. Kinet.
1970, 2, 493−496.
(21) Yan, G.; Brinkmann, N. R.; Schaefer, H. F. III. J. Phys. Chem. A
2003, 107, 9479−9485.
(22) Craciun, L.; Jackson, J. E. J. Phys. Chem. A 1998, 102, 3738−
3745.
(23) Blanksby, S. J.; Ellison, G. B. Acc. Chem. Res. 2003, 36, 255−263.
(24) Muller, N.; Pritchard, D. E. J. Chem. Phys. 1959, 31, 1471−1476.
(25) The experimental bond energies were used for 1H−3H and the
G3 energies for the remaining compounds.
(26) (a) Coote, M. L.; Lin, C. Y.; Beckwith, A. L. J.; Zavitsas, A. A.
Phys. Chem. Chem. Phys. 2010, 12, 9597−9610. (b) Zavitsas, A. A.;
Rogers, D. W.; Matsunaga, N. J. Org. Chem. 2010, 75, 5697−5700.
(c) Poutsma, M. L. J. Org. Chem. 2011, 76, 270−276. (d) Wodrich, M.
D.; McKee, W. C.; Schleyer, P. v. R. J. Org. Chem. 2011, 76, 2439−
2447.
(27) (a) Zhao, Y.; Truhlar, D. G. J. Phys. Chem. A 2008, 112, 1095−
1099. (b) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215−
241. (c) Zhao, Y.; Truhlar, D. G. Acc. Chem. Res. 2008, 41, 157−167.
(28) Dunning, T. H. Jr. J. Chem. Phys. 1989, 90, 1007−1023.
(29) Adcock, W.; Clark, C. I. J. Org. Chem. 1995, 60, 723−724.
(30) Adcock, W.; Kok, G. B. J. Org. Chem. 1985, 50, 1079−1087.
(31) Adcock, W.; Krstic, A. R.; Duggan, P. J.; Shiner, V. J. Jr.; Coope,
J.; Ensinger, M. W. J. Am. Chem. Soc. 1990, 112, 3140−3145.
(32) (a) Della, E. W.; Tsanaktsidis, J. Organometallics 1988, 7, 1178−
1182. (b) Adcock, W.; Gangodawila, H.; Kok, G. B.; Iyer, V. S.;
Kitching, W.; Drew, G. M.; Young, D. Organometallics 1987, 6, 156−
166.
(33) Fearon, F. W. G.; Gilman, H. J. Organomet. Chem. 1967, 9, 403−
416.
(34) Grob, C. A.; Sawlewicz, P. Helv. Chim. Acta 1988, 71, 1508−
1510.
(35) Sasaki, T.; Shimizu, K.; Ohno, M. Synth. Commun. 1984, 14,
853−856.
(7) (a) Liu, R.; Allinger, N. L. J. Phys. Org. Chem. 1993, 6, 551−554.
(b) Sauers, R. R. Tetrahedron 1999, 55, 10013−10026. (c) Feng, Y.;
Liu, L.; Wang, J. T.; Zhao, S. W.; Guo, Q. X. J. Org. Chem. 2004, 69,
3129−3138. (d) Nunes, P. M.; Estcio, S. G.; Lopes, G. T.; Cabral, B. J.
C.; Borges dos Santos, R. M.; Simes, J. A. M. Org. Lett. 2008, 10,
1613−1616.
(8) (a) Kruppa, G. H.; Beauchamp, J. L. J. Am. Chem. Soc. 1986, 108,
2162−2169. (b) Aubry, C.; Holmes, J. L.; Walton, J. C. J. Phys. Chem.
A 1998, 102, 1389−1393.
(9) Abboud, J. -L. M.; Castano, O.; Dav
Chem. Phys. Lett. 2001, 337, 327−330.
(36) Marshall, A. G.; Roe, D. C. J. Chem. Phys. 1980, 73, 1581−1590.
(37) Wang, T. C. L.; Ricca, T. L.; Marshall, A. G. Anal. Chem. 1986,
58, 2935−2938.
(38) Frisch, M. J., et al. Gaussian 09, Gaussian, Inc., Wallingford, CT,
2009.
́
alos, J. Z.; Gromperts, R.
̃
1914
dx.doi.org/10.1021/jo202519w | J. Org. Chem. 2012, 77, 1909−1914