Redox chemistry of selenenic acids and the insight it brings on transition state geometry in the reactions of peroxyl radicals

Article abstract of DOI:10.1021/ja411493t

The redox chemistry of selenenic acids has been explored for the first time using a persistent selenenic acid, 9-triptyceneselenenic acid (RSeOH), and the results have been compared with those we recently obtained with its lighter chalcogen analogue, 9-tr

Full text of DOI:10.1021/ja411493t

Article
pubs.acs.org/JACS
Redox Chemistry of Selenenic Acids and the Insight It Brings on
Transition State Geometry in the Reactions of Peroxyl Radicals
Zosia Zielinski,† Nathalie Presseau,† Riccardo Amorati,‡ Luca Valgimigli,*,‡ and Derek A. Pratt*,†
†Department of Chemistry, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada
‡Department of Chemistry “G. Ciamician”, University of Bologna, I-40126 Bologna, Italy
S
* Supporting Information
ABSTRACT: The redox chemistry of selenenic acids has been explored for the first time using a persistent selenenic acid, 9-
triptyceneselenenic acid (RSeOH), and the results have been compared with those we recently obtained with its lighter
chalcogen analogue, 9-triptycenesulfenic acid (RSOH). Specifically, the selenenyl radical was characterized by EPR spectroscopy
and equilibrated with a phenoxyl radical of known stability in order to determine the O−H bond dissociation enthalpy of RSeOH
(80.9 0.8 kcal/mol): ca. 9 kcal/mol stronger than in RSOH. Kinetic measurements of the reactions of RSeOH with peroxyl
radicals demonstrate that it readily undergoes H-atom transfer reactions (e.g., k = 1.7 × 105 M−1 s−1 in PhCl), which are subject
to kinetic solvent effects and kinetic isotope effects similar to RSOH and other good H-atom donors. Interestingly, the rate
constants for these reactions are only 18- and 5-fold smaller than those measured for RSOH in PhCl and CH3CN, respectively,
despite being 9 kcal/mol less exothermic for RSeOH. IR spectroscopic studies demonstrate that RSeOH is less H-bond acidic
than RSOH, accounting for these solvent effects and enabling estimates of the pKas in RSeOH and RSOH of ca. 15 and 10,
respectively. Calculations suggest that the TS structures for these reactions have significant charge transfer between the chalcogen
atom and the internal oxygen atom of the peroxyl radical, which is nominally better for the more polarizable selenenic acid. The
higher than expected reactivity of RSeOH toward peroxyl radicals is the strongest experimental evidence to date for charge
transfer/secondary orbital interactions in the reactions of peroxyl radicals with good H-atom donors.
molecule mimics for therapeutic and/or chemopreventive
purposes against disease wherein oxidative stress has been
implicated.9
INTRODUCTION
■
Selenenic acids (RSeOH) are the selenium analogues of
sulfenic acids (RSOH), both being the heavier (valence)
isoelectronic cousins to the more commonly encountered
hydroperoxides (ROOH). Selenenic acids are believed to be
transient intermediates in a number of redox reactions
involving organoselenium compounds, inferred largely from
the analogous chemistry exhibited by sulfenic acids derived
from organosulfur compounds.1,2 The two most important
reactions that lead to selenenic acids are the oxidation of a
selenol (e.g., with H2O2 as in eq 1), first demonstrated directly
in 2001,3 and syn elimination from a selenoxide (e.g., eq 2), a
common synthetic transformation for late-stage introduction of
an alkene.4−6 The former reaction is believed to be key to the
essential antioxidant enzyme glutathione peroxidase,7,8 whose
active site selenocysteine residue is responsible for the
reduction of hydroperoxides and H2O2 to alcohols and water,
respectively, prompting the widespread pursuit of small-
RSeH + H2O2 → RSeOH + H2O
(1)
RSe(O)CH2CH2R′ → RSeOH + H2CCH−R′
(2)
Despite their prominence in these (and other1) roles, the
properties of selenenic acids have been elusive, owing to the
ease with which they self-condense, disproportionate, and/or
oxidize. While the synthesis of isolable selenenic acids has been
pursued for some time,10 with a handful of examples of
sterically hindered areneselenenic acids being reported,3,11 as
well as a single alkaneselenenic acid,12 essentially no
investigation of their physicochemical properties has accom-
Received: November 18, 2013
Published: January 2, 2014
© 2014 American Chemical Society
1570
dx.doi.org/10.1021/ja411493t | J. Am. Chem. Soc. 2014, 136, 1570−1578

Journal of the American Chemical Society
Article
panied them. Of particular interest are their redox character-
istics, which are expected to figure prominently in their
chemistry.
to minimize the self-condensation of 2 to give the
corresponding selenoseleninate 6. As such, solutions of 2
were prepared immediately prior to use in the experiments
reported below.
A deoxygenated solution of the selenenic acid in benzene was
photolyzed in the presence of 10% (by volume) di-tert-butyl
peroxide in the cavity of an X-band EPR spectrometer,
affording the noisy spectrum shown in Figure 1. The field
We recently reported the results of detailed studies on the
redox chemistry of a persistent sulfenic acid, 9-triptycenesul-
fenic acid (1), in order to provide a thermodynamic13 and
kinetic14 rationale for the involvement of sulfenic acids in the
radical-trapping antioxidant activity of natural product organo-
sulfur compounds, such as the garlic-derived allicin15 and
anamu-derived petivericin,16 as well as sulfenic acid mediated
processes, including cysteine-mediated redox cell signaling.17 In
an effort to shed light on the redox chemistry of selenenic acids
and provide a basis for a direct comparison of their electronic
properties with those of sulfenic acids (and hydroperoxides),
we report here the results of corresponding studies on the
selenium analogue of 1: 9-triptyceneselenenic acid (2).12 The
studies described herein provide important fundamental
insights into the thermodynamics and kinetics of radical
reactions of selenenic acids, in particular, and provide a unique
perspective on the importance of transition state geometry in
formal H-atom transfer reactions to autoxidation chain-carrying
peroxyl radicals, in general.
Figure 1. EPR spectrum of the selenenyl radical generated from
continuous photolysis of a solution of 2 in benzene containing 10% di-
tert-butyl peroxide. The simulated spectrum is shown in red.
RESULTS AND DISCUSSION
■
9-Triptyceneselenenic acid (2)12 was prepared as shown in
Scheme 1. Briefly, 9-bromotriptycene was subjected to
center of the broad singlet (2.4 G line width) is g = 2.0191,
consistent with an oxygen-centered radical with spin delocaliza-
tion onto a heavier atom and slightly larger than that of the
corresponding sulfinyl radical derived from the analogous
sulfenic acid 1, for which we had measured g = 2.0114.13
Continuous photolysis was necessary as the resultant selenenyl
radical decayed on the time scale of the acquisition, yielding
only the weak signal shown which prevented resolution of the
coupling with 77Se (which has a nuclear spin of I = +1/2 and is
7.6% at natural abundance).
Scheme 1. Synthesis of 9-Triptyceneselenenic Acid 2 and Its
Condensation to the Corresponding Selenoseleninate 6
While the selenenyl radical was not particularly persistent, we
hoped that its rapid equilibration with a compound of known
O−H BDE13,18,19 would allow the determination of the O−H
BDE of a selenenic acid for the first time. To provide guidance
on which compound to use in the equilibration experiment, we
carried out computations on a model selenenic acid (t-
BuSeOH), for which CBS-QB320 calculations predicted an
O−H BDE of 81.2 kcal/mol. On the basis of this result we
chose 2,4,6-tri-tert-butylphenol (TTBP) as the reference
compound for the equilibriation experiment, as it has an O−
H BDE of 80.1 kcal/mol19 and yields a persistent phenoxyl
radical upon H-atom abstraction. Gratifyingly, continuous
photolysis of mixtures of selenenic acid and TTBP in benzene
containing 10% di-tert-butyl peroxide (by volume) afforded
spectra showing both the selenenyl and phenoxyl radicals (cf.
lithium−halogen exchange, and the resultant organolithium
was quenched with bis(phenethyl) diselenide (3), which was
prepared from phenethyl bromide as suggested by Thompson
and Boudjouk.36 The resultant selenide 4 was then oxidized
with MCPBA to yield the selenoxide 5, which underwent Cope-
type elimination at room temperature to give 2. The successful
isolation of 2 required the solid-state decomposition of the
precursor selenoxide under high vacuum over a 2 week period
Figure 2). Equilibrium constants of K = 4.0
2.1 were
obtained from simulation of the spectra, which afford ΔH = 0.8
0.4 kcal/mol assuming a negligible entropy change for the
formal H atom transfer process and therefore an O−H BDE of
80.9 0.8 kcal/mol for 2.
1571
dx.doi.org/10.1021/ja411493t | J. Am. Chem. Soc. 2014, 136, 1570−1578

Journal of the American Chemical Society
Article
BuSeOH. These calculations reveal that the O−H bond in t-
BuSeOH (81.2 kcal/mol) is slightly stronger than the Se−H
bond in t-BuSeH (78.9 kcal/mol). This result is particularly
interesting because the trend is completely different from that
observed for the lighter chalcogen: the O−H bond in t-BuSOH
(68.6 kcal/mol) is much weaker than the S−H bond in t-BuSH
(87.6 kcal/mol).23 This fundamentally different trendwhich
is rooted in the very high stability of the sulfinyl radical13
underscores the need for caution when rationalizing the
reactivity of selenium-containing compounds simply on the
basis of their similarity to sulfur-containing compounds.
Electrochemical experiments also suggest that selenenic acids
are more difficult to oxidize than sulfenic acids. Cyclic
voltammograms, which were irreversible at scan rates ranging
from 1 to 1000 mV/s, consistently showed anodic (oxidation)
peaks at potentials that were ca. 200 mV more oxidizing than
for the sulfenic acid under identical conditions (representative
voltammograms are given in the Supporting Information).
Attempts to obtain reversible voltammograms by addition of
acid to suppress deprotonation of the resultant radical cation
(and subsequent rapid reactions of the selenenyl radical; vide
supra) led to no observable signal in the potential window,
presumably due to acid catalysis of the self-condensation of the
selenenic acid to yield the selenoseleninate 6. Likewise, addition
of base to yield the selenenate anion in the hopes of obtaining a
reversible RSeO•/RSeO− couple also led to no observable
signal in the potential window, again due to base-catalyzed
formation of the selenoseleninate 6.
The H-atom transfer reactivity of the selenenic acid was
probed in studies of the kinetics of its reactions with peroxyl
radicals, the results of which could be compared to the results
of analogous experiments we previously carried out on the
corresponding sulfenic acid.14 These measurements were made
using the well-established inhibited autoxidation of styrene
approach, using O2 consumption to monitor reaction progress
(cf. Figure 4), from which rate constants could be determined
for the formal H-atom transfer (kinh) from the initial rates
(Table 1).24 In chlorobenzene, a rate constant of (1.7 0.3) ×
105 M−1 s−1 was obtained, indicating that selenenic acids are
indeed very reactive in H-atom transfer reactions. For
comparison, the analogous sulfenic acid reacts with a rate
Figure 2. Representative EPR spectrum obtained from continuous
photolysis of a 1:1 mixture of 2 and 2,4,6-tri-tert-butylphenol in
benzene containing 10% di-tert-butyl peroxide. The simulated
spectrum for a 4:1 mixture of the 2,4,6-tri-tert-butylphenoxyl and
selenenyl radicals is shown in red.
The increased strength of the O−H bond in the selenenic
acid in comparison to the sulfenic acid (71.9 0.3 kcal/mol,
also measured by the radical equilibration EPR approach; for
comparison CBS-QB3 predicts an O−H BDE of 68.6 kcal/mol
in t-BuSOH)13,42 can be rationalized by the longer Se−O bond
in the selenenyl radical as compared to the S−O bond in the
sulfinyl radical. The calculated minimum energy structures of
the tert-butyl selenenyl and sulfinyl radicals from the CBS-QB3
calculations are shown in Figure 3, wherein the Se−O and S−O
Figure 3. Calculated structures and spin density distributions in t-
BuSeO• (A) and t-BuSO• (B).
bond distances are 1.68 and 1.51 Å, respectively. The associated
spin delocalization in the selenenyl radical is predicted to be ca.
20% on the selenium atom and 80% on the oxygen atom as
compared to ca. 40% on the sulfur atom and 60% on the
oxygen atom in the sulfinyl radical.43
It is of interest to compare the O−H BDE in a selenenic acid
to the Se−H BDE in a selenoloften its direct precursor in
redox reactions, such as in the catalytic cycle of glutathione
peroxidase (as in eq 1). Since, to the best of our knowledge, no
experimental value is available for the Se−H BDE of an
alkaneselenol (the Se−H BDEs for HSeH and PhSeH are
Figure 4. Oxygen consumption during the autoxidation of styrene
(50% by volume) initiated by AIBN (0.05 M) at 303 K without
inhibitor (dashed line) or in the presence of 2 (7.6 μM) in PhCl +
0.5% CH3OH (black line) or PhCl + 0.5% CH3OD (red line). The
resultant kH/kD value is 2.9.
reported to be 79.0
0.221 and 78
4 kcal/mol,22
respectively), we calculated it for t-BuSeH using CBS-QB3 to
allow for direct comparison to the foregoing calculations on t-
1572
dx.doi.org/10.1021/ja411493t | J. Am. Chem. Soc. 2014, 136, 1570−1578

Journal of the American Chemical Society
Article
pKa ≈ 10),10 while 2 has a pKa similar to that of methanol (αH2
= 0.37, pKa ≈ 15).10 We had previously measured a pKa value
of 12.5 for 1 in 4/1 CH3CN/H2O, which would undoubtedly
be slightly lower in water, but attempts to measure the pKa of 2
under similar conditions were unsuccessful owing to
selenoseleninate formation under the conditions of the
potentiometric titration.
Table 1. Rate Constants (kinh) for the Reactions of Selenenic
Acid 2 with Peroxyl Radicals at 303 K
a
kinh (M−1 s−1
)
c
conditions
styrene/PhCl
2
1
(1.7 0.3) × 105
(3.5 0.4) × 104
(3.4 0.3) × 104
(5.1 0.6) × 103
6.7 0.9
(3.0 0.3) × 106
(1.6 0.3) × 105
(1.1 0.2) × 105
(1.8 0.2) × 104
6.1 0.4
styrene/CH3CN
b
styrene/CH3CN/H2O
b
styrene/CH3CN/D2O
kH/kD
a
Literature data for the reactions of sulfenic acid 1 obtained under the
b
same conditions are presented for comparison. H2O or D2O added in
1% v/v. Taken from ref 13.
c
constant of (3.0
0.3) × 106 M−1 s−1 under identical
conditions,14 a reactivity that is identical with that of α-
tocopherol, nature’s premier lipid-soluble radical-trapping
antioxidant, under equivalent conditions.24 A kinetic solvent
effect diminishes the reactivity of 2 by a factor of 4.9 on moving
from chlorobenzene to acetonitrile, in comparison to a factor of
19 for 1. As in our previous studies with sulfenic acid 1,14
a
large primary deuterium kinetic isotope effect was determined
for the reaction of 2 with peroxyls, consistent with a formal H-
atom transfer mechanism.
Ingold has shown that the rates of H-atom transfer reactions
vary in different solvents (S) according to a model that assumes
the H-atom donor (H-A) is unreactive when its labile H-atom
is involved in an H-bond with the solvent.25,26 As a result, the
kinetics can be accurately described only using a model that
involves a predissociation of this H-bonded complex:
•
S‐‐‐H−A ⇆ S + H−A + X• → S + A + H−X
(3)
Ingold went on to show that the solvent effects could be
quantitated using an empirical linear free energy relationship
based on the H-bond donating strength of the H-atom donor
(αH2 ) and the H-bond accepting strength of the solvent (βH2 ):
log kS = −8.3α2HβH + log k0
(4)
Figure 5. Representative FT-IR spectra of the O−H stretching region
of 2 (5 mM, top) and 1 (10 mM, bottom) in CCl4 containing
increasing amounts of acetonitrile as cosolvent.
2
As such, the larger kinetic solvent effect on the reaction of the
sulfenic acid in comparison to the selenenic acid implies that
the former is a better H-bond donor than the latter.
Indeed, FT-IR spectra of 1 and 2 obtained in the non-H-
bond accepting solvent CCl4 in the presence of increasing
amounts of acetonitrile reveal trends that are consistent with
the observed kinetic solvent effects. From these spectra, the
equilibrium constants corresponding to H-bond formation
between 1 or 2 and acetonitrile can be derived from the
integration of the peak corresponding to the free O−H stretch
as a function of acetonitrile concentration (vide infra) to give
At first glance, the slower reaction of the selenenic acid with
peroxyl radicals relative to the sulfenic acid appears fully
consistent with the fact that the selenenic acid has a stronger
O−H bond. However, Evans−Polanyi relationships between
rate constants for the reactions of H-atom donors with peroxyl
radicals (kinh) and the X−H BDEs of the H-atom donors imply
that a much larger reactivity difference should exist between the
selenenic and sulfenic acids given the 9 kcal/mol difference in
their O−H BDEs. For instance, the Evans−Polanyi correlations
in Figure 6 for 4-substituted phenols, 4-substituted 2,6-
dimethylphenols, and 4-substituted 2,6-di-tert-butylphenols
predict differences of 3 orders of magnitude in kinh for an O−
H BDE difference of 9 kcal/mol.28 Interestingly, including the
data for 1 and 2 in these correlations finds them on the lines of
best fit for the 2,6-di-tert-butylated phenols and the
unsubstituted phenols, respectively. The data for substituted
phenols have long been assumed to lie on different lines owing
to the differing steric demands on the reactions, which increases
the entropy of activation for the more hindered substrates,
decreasing the pre-exponential factor for their reactions.28 This
implies that the sulfenic and selenenic acids also have different
steric requirements. Indeed, the longer C−Se and Se−O bonds
2.78
0.45 and 0.92
0.20 M−1, respectively. Using
Abraham’s equation, these equilibrium constants yield values
of αH2 of 0.54 and 0.37 for 1 and 2, respectively. On the basis of
these values of αH2 , we would expect the rate constants for H-
atom transfer from 1 and 2 to peroxyl radicals to drop 22- and
8-fold on moving from chlorobenzene (βH2 = 0.09)27 to
acetonitrile (β2H = 0.39),27,44 from Ingold’s eq 4,25 in excellent
agreement with the experimental results of 19 and 4.9,
respectively.45
It should be pointed out that the values of αH2 derived from
the FT-IR measurements also provide some insight into the
relative acidities of 1 and 2. Since the H-bond donating ability
generally correlates with the pKa of the donor, it can be
expected that 1 has a pKa similar to that of phenol (αH2 = 0.60,
1573
dx.doi.org/10.1021/ja411493t | J. Am. Chem. Soc. 2014, 136, 1570−1578

Journal of the American Chemical Society
Article
CBS-QB3 calculations on less sterically encumbered sulfenic
acids (e.g., t-BuSOH) and peroxyl radicals (MeOO•) indicated
that this conformation is preferred by a significant margin over
a TS structure wherein the substituents on the oxygen atoms
⧧
adopt an anti conformation (ΔGsyn = 10.1 kcal/mol and
ΔGanti = 14.9 kcal/mol, Figure 8B,D),15 we anticipated that
⧧
Figure 6. Evans−Polanyi correlations for the reactions of peroxyl
○
radicals with 4-substituted phenols ( ), 2,6-dimethyl 4-substituted
▲
●
phenols ( ) and 2,6-di-tert-butyl 4-substituted phenols ( ). Also
shown are the corresponding data for 1 and 2.
in the selenenic acid (1.956 and 1.866 Å, respectively)12 relative
to the C−S and S−O bonds in the sulfenic acid (1.833 and
1.622 Å, respectively)29 should minimize steric interactions
between the triptycene moiety and the substituent on the
peroxyl radical to which the H-atom is being transferred.
To provide insight into the transition state (TS) structures
by which these reactions proceed, and the extent to which they
are impeded by steric interactions, we again turned to
computation. Since TS calculations of reactions of 1 and 2
with peroxyl radicals are far too large to be carried out with the
high-accuracy CBS-QB3 approach used for the BDE calcu-
lations above, we turned to density functional theory using
B3LYP and the dispersion-correcting potentials (DCPs) of
DiLabio and Torres.30 The calculated minimum energy TS
structures are shown in Figure 7, and alongside are the
Figure 8. CBS-QB3-calculated syn (A, B) and anti (C, D) transition
state structures for the reactions of t-BuSeOH (A, C) and t-BuSOH
(B, D) with a model alkylperoxyl radical (MeOO•).
the larger substituents would alter this preference.14 In order to
provide a corresponding unhindered comparison for the
selenenic acid, we carried out CBS-QB3 calculations on the
reaction of t-BuSeOH and MeOO•. Indeed, the same geometric
preference is predicted for this reaction (Figure 8A,C), with the
syn TS being preferred by an even greater margin over the anti
⧧
⧧
TS (ΔGsyn = 7.9 kcal/mol and ΔGanti = 19.8 kcal/mol).
However, most interestingly, the order of reactivity is predicted
to be different in the smaller models than in the larger models:
that is, the t-BuSeOH/•OOMe reaction is predicted to be faster
than the corresponding t-BuSOH/•OOMe reaction.47 This is
particularly surprising, given that the strength of the O−H
bond in t-BuSeOH is predicted to be 12.6 kcal/mol stronger
than the O−H bond in t-BuSOH!
The reactions of sulfenic acids with peroxyl radicals have
been described as taking place by proton-coupled electron
transfer,15,31 facilitated by the overlap between orbitals based
on the sulfenic acid sulfur atom and the inner oxygen atom of
the peroxyl radicalseen clearly in the three highest (doubly)
occupied MOs shown in Figure 9accounting for the
preference for the syn TS in the formal H-atom transfer.48
This description would appear to be appropriate for the
reactions of selenenic acids with peroxyl radicals as well, as the
same interactions are evident in the corresponding MOs shown
on the left in Figure 9. The lower barrier for the t-
BuSeOH/•OOMe reaction in comparison to that for t-
BuSOH/•OOMe via the low-energy syn pathway can therefore
be understood on the basis of improved charge transfer
between the larger, less electronegative selenium atom and the
internal oxygen atom of the peroxyl radical in comparison to
the smaller, more electronegative sulfur atom. Indeed, bond
order analysis of the structures in Figure 8 (and Figure 9)
Figure 7. DCP-B3LYP/6-31+G(2d,2p)-calculated transition state
structures for the reactions of selenenic acid 2 (A) and sulfenic acid
1 (B) with a styrylperoxyl radical (representative of the chain-carrying
peroxyl radicals in styrene autoxidations).
corresponding thermokinetic parameters. While the calculated
rate constants (from transition state theory) are overestimated
relative to experiment by roughly 1 order of magnitude,46 they
are consistent with the experimental results in that they indicate
that the selenenic acid 2 reacts more slowly with peroxyl
radicals than the sulfenic acid 1. Indeed, the calculated ratio
k(1)/k(2) = 33 is in very good agreement with the
experimental value of kinh(1)/kinh(2) = 18.
Somewhat surprisingly, the TS structures in Figure 7 feature
a syn relationship of the substituents on the oxygen atoms
between which the H-atom is being transferred. While previous
1574
dx.doi.org/10.1021/ja411493t | J. Am. Chem. Soc. 2014, 136, 1570−1578

Journal of the American Chemical Society
Article
Figure 10. Dependence of the CBS-QB3-calculated ΔG⧧ values for
the reactions of t-BuSeOH (red, open symbols) and t-BuSOH (black,
closed symbols) with MeOO• on the Se−O/O−O and S−O/O−O
torsion angles in the transition state structures, respectively. The
points are labeled with the distance between the chalcogen atom and
the inner oxygen atom of the peroxyl radical in the corresponding
structures.
radicals. Indeed, while the calculated TS structures in Figures 7
(full model) and 8 (smaller model) feature similar overall
geometries, the distances between the chalcogen atom and the
inner oxygen atom of the peroxyl radical are significantly longer
in the former in comparison to the latter (3.12 and 3.14 Å vs
3.01 and 3.00 Å), and the Se−O/O−O and S−O/O−O torsion
angles have opened up to 17 and 20° from 9 and 10°,
respectively,43 diminishing overlap between the orbitals on the
chalcogen atom and the internal oxygen atom of the peroxyl.
The higher than expected reactivity of the persistent
selenenic acid 2 toward peroxyl radicals (relative to the
persistent sulfenic acid 1) is the strongest evidence to date for
charge transfer/secondary orbital interactions in the reactions
of good H-atom donors with peroxyl radicals. While this has
been suggested to be key to the reactions of ubiquitous radical-
trapping antioxidants, such as phenols31,32 and diaryl-
amines,33,34 as well as related compounds,15,35 it is yet to be
firmly supported by experiment. Given that these reactions
underlie the preservation of virtually all hydrocarbon materials
(including us!), we trust the foregoing insights will help clarify
structure−reactivity trends already established or yet to be
uncovered with the design of new compounds.
Figure 9. The four highest energy occupied molecular orbitals
corresponding to the syn transition state structures for the reactions of
t-BuSeOH (A) and t-BuSOH (B) with MeOO• obtained from the
UB3LYP/CBSB7 step of the CBS-QB3 calculation.
reveals a much larger bonding interaction between the selenium
atom and the internal oxygen atom of the peroxyl radical in the
selenenic acid syn TS (0.229) in comparison to the sulfur and
corresponding oxygen atom in the sulfenic acid syn TS
(0.180).49
Since the high reactivity of selenenic acids toward peroxyl
radicals appears to be driven by charge transfer from the
selenium to the internal oxygen atom of the peroxyl radical, the
barrier of the reaction should be highly dependent on the
extent to which orbitals centered on these atoms overlap. To
illustrate this point, we calculated the ΔG⧧ value for the
reactions of t-BuSeOH with MeOO• as a function of the Se−
O/O−O torsion angle in the TS structure. The results are
shown in Figure 10, revealing a substantial dependence of the
barrier height on the torsion angle: rising from ca. 8 kcal/mol
to ca. 20 kcal/mol upon going from the syn structure to the
orthogonal structure. We carried out analogous calculations for
the reactions of t-BuSOH with MeOO•, which revealed a
similar, but less severe, angular dependence; the barrier rises
from ca. 10 kcal/mol to ca. 15 kcal/mol in the orthogonal
structure (see also Figure 10).50 Thus, while selenenic acids
may be more reactive toward peroxyl radicals than sulfenic
acids at the “ideal” TS geometry, they become less reactive as
the geometry is distorted to reduce the Se/O overlap in the TS.
It seems reasonable to suggest that the triptycene moieties
necessary to make the sulfenic and selenenic acids persistent
and amenable to experimental study also make it difficult for
them to achieve “ideal” TS geometries in reactions with peroxyl
CONCLUSIONS
■
A persistent selenenic acid bearing a triptycene moiety has been
used to characterize the redox chemistry of selenenic acids for
the first time. The strength of the O−H bond in the selenenic
acid is ca. 81 kcal/mol, making it weaker than the
corresponding O−H bond in a hydroperoxide (ca. 86 kcal/
mol) but much stronger than in a sulfenic acid (ca. 70 kcal/
mol).13,15 Insights from theoretical calculations indicate that the
lack of any periodic trend in the BDEs is the result of lesser
spin delocalization in the selenenyl radical relative to the
sulfinyl radical, owing to the longer Se−O bond. As such,
selenenic acids are unusual among the three (valence)
isoelectronic species in that the O−H bond is stronger than
the Se−H bond in the selenol (from which it is often derived).
1575
dx.doi.org/10.1021/ja411493t | J. Am. Chem. Soc. 2014, 136, 1570−1578

Journal of the American Chemical Society
Article
2.1 g, 5.4 mmol) was suspended in a minimum amount of dry benzene
and added slowly to the reaction mixture at 0 °C. The reaction mixture
was stirred at room temperature for 12 h, before it was quenched with
water and extracted with ether. The organic phase was washed with
brine and then dried over MgSO4, filtered, and concentrated in vacuo.
The resulting oily yellow solid was loaded on a silica gel flash column
and eluted with 5% dichloromethane in hexanes to yield an off-white
solid which was recrystallized from benzene/ethanol (1.77 g, 75%
yield): 1H NMR (400 MHz, CDCl3) δ 3.16−3.21 (m, 2H), 3.28−3.33
(m, 2H), 5.38 (s, 1H), 7.02−7.05 (m, 6H), 7.27−7.31 (m, 1H), 7.35−
7.41 (m, 7H), 7.56−7.59 (m, 3H); 13C NMR (400 MHz, CDCl3) δ
25.7, 36.8, 54.1, 60.3, 123.4, 124.0, 125.0, 125.6, 126.6, 128.5, 128.6,
140.6, 145.1, 145.6; HRMS (EI+) calculated for C28H22Se 438.0887,
observed 438.0876.
9-Triptycene Phenethyl Selenoxide (5). A solution of 4 (94 mg,
0.21 mmol) in dry dichloromethane (18 mL) was cooled to −78 °C.
mCPBA (77%, 48 mg, 0.21 mmol) was added slowly in 5 mL of dry
dichloromethane. The reaction mixture was worked up immediately by
washing twice with 20 mL of cold 0.7 M aqueous KOH, followed by
20 mL of ice water, then 20 mL of cold brine. The organic phase was
dried over MgSO4, filtered, and concentrated in vacuo at −78 °C. A
white solid was obtained (84 mg, 86% yield). The 9-triptycene
phenethyl selenoxide was generally obtained as a mixture with 9-
triptyceneselenenic acid and styrene due to Cope elimination in situ:
1H NMR (300 MHz, CDCl3) δ 3.47−3.64 (m, 2H), 3.83−3.93 (m,
1H), 4.00−4.09 (m, 1H), 5.37 (s, 1H), 7.02−7.05 (m, 6H), 7.28−7.37
(m, 9H) 7.93−8.02 (m, 1H), 8.31−8.39 (m, 1H); 13C NMR (300
MHz, CDCl3) δ 29.7, 30.5, 48.7, 68.1, 121.6, 123.0, 123.6, 125.3,
125.7, 127.1, 128.8, 129.0, 144.3, 145.7. No molecular ion could be
observed by MS due to its ready fragmentation to the selenenic acid.
9-Triptyceneselenenic Acid (2). Selenoxide 5 was left under high
vacuum for 2 weeks at room temperature: 1H NMR (300 MHz,
CDCl3) δ 5.42 (s, 1H), 7.02−7.05 (m, 6H), 7.41−7.44 (m, 6H); 13C
NMR (300 MHz, CDCl3) δ 54.1, 64.1 123.0, 123.7, 125.3, 125.7,
144.3, 145.7; HRMS (ES−) calculated for C20H14OSe 349.0132,
observed 349.0104. The spectral characteristics are in good agreement
with those presented in the literature.12
EPR Experiments. Spectra were recorded at 298 K by irradiating
deoxygenated benzene solutions of 2 containing di-tert-butyl peroxide
(10% v/v) with a 500W high-pressure Hg lamp in the spectrometer
cavity. The measured g factor was corrected with respect to the known
value of 2,2,6,6-tetramethylpiperidine-N-oxyl radical in benzene (g =
2.0064). Equilibration studies were performed by irradiating mixtures
of 2 and 2,4,6-tri-tert-butylphenol19 in different ratios.18,33,37,38 The
relative amount of the corresponding radicals radicals was determined
by fitting the experimental spectrum with computer simulations using
WinESR software, developed by Prof. Marco Lucarini (Univ.
Bologna), based on the Monte Carlo method.37 Different irradiation
intensities were compared to make sure that equilibrium was
established. The equilibrium constant K was determined according
to eq 5, which yielded ΔH for equilibration by eq 6 under the
assumption that ΔS ≈ 0.33,37,38
This trend is in contrast with those established for sulfenic acids
and hydroperoxides, which are characterized by significantly
weaker bonds (by ca. 20 kcal/mol) to hydrogen than in the
corresponding thiols and alcohols, respectively. These insights
serve as a cautionary note on rationalizing the reactivity of
selenium-containing compounds simply on the basis of their
structural similarity to sulfur-containing compounds.
The kinetics of formal H-atom transfer reactions from the
persistent selenenic acid were evaluated using peroxyl radicals
as a model oxidant, revealing very high reactivity (kinh = 1.7 ×
105 M−1 s−1) relative to the strength of the O−H bond. CBS-
QB3 calculations on a model system wherein the triptycene is
replaced with a tert-butyl group and a methylperoxyl radical is
used as the oxidant reveal that these reactions proceed with a
negligible enthalpic barrier via a transition state stabilized
greatly by interactions between orbitals with significant
contributions from the selenium atom and the internal oxygen
atom of the peroxyl radical. The calculations also indicate that
these interactions are highly sensitive to the transition state
geometry; even small perturbations in the structure that
diminish these interactions significantly increase ΔG⧧ and
account for the lower observed reactivity of the hindered 9-
triptyceneselenenic acid in comparison to that expected on the
basis of the model calculations. These insights reveal the
dramatic impact of small perturbations on the transition state
geometries for formal H-atom transfer reactions facilitated by
secondary orbital interactions (e.g., in proton-coupled electron
transfer reactions). That is, steric arguments for the ration-
alization of reactivity trends must be considered carefully in
light of the impact they have not only on potential approach
trajectories of the reactants but also in their abilities to
maximize secondary orbital overlap in the transition state.
The foregoing results indicate that unhindered selenenic
acids are likely to be among the most reactive H-atom donors
toward radical centers that have an adjacent high-lying electron
pair, such as in peroxyl or phenoxyl radicals, which facilitate the
reaction via interaction with orbitals centered on the selenium
atom. As such, the formation of a selenenic acid either by
oxidation of a selenol or by Cope elimination from a selenoxide
may serve to convert a relatively poor H-atom donor into an
excellent one and may present a complementary mechanism for
the reducing activities of selenium compounds beyond the two-
electron chemistry that is most commonly observed and/or
assumed.
EXPERIMENTAL SECTION
■
Synthesis. Bis(2-phenethyl) Diselenide (3). Dry selenium powder
(3.5 g, 44 mmol), sodium chips (1.06 g, 44 mmol), and naphthalene
(0.57 g, 4.5 mmol) were stirred in dry THF (100 mL) under argon for
12 h. To the dark purple mixture was added 2-bromoethylbenzene
(5.88 mL, 44 mmol) dropwise. The mixture was stirred for 1 h, during
which the color changed to light orange. A tan salt was removed by
filtration, and the filtrate was concentrated in vacuo. The resulting
orange oil was loaded on a silica gel flash column and eluted first with
hexanes and then with benzene to give an orange oil (7.77 g, 48%
yield): 1H NMR (400 MHz, CDCl3) δ 3.03−3.07 (m, 4H), 3.14−3.18
(m, 4H), 7.20−7.24 (m, 6H), 7.29−7.33 (m, 4H); 13C NMR (CDCl3,
300 MHz) δ 30.7, 37.5, 126.3, 128.4, 128.5, 140.7; HRMS (EI+)
K = ([TripSeOH]/[TTBP]) × ([TTBP•]/[TripSeO•])
(5)
(6)
ΔG = ΔH − TΔS = −RT ln K
Autoxidations. Rate constants (kinh) for the reactions of 2 with
peroxyl radicals were dermined by kinetic analysis of inhibited
autoxidations of styrene (50% v/v) in air-saturated chlorobenzene or
acetonitrile solution at 303 K.24 The reaction was thermally initiated at
constant rate Ri (determined experimentally, in the range (2−9) ×
10−9 M s−1) by the decomposition of 2,2′-azodiisobutyronitrile (AIBN,
(1−5) × 10−2 M) and the oxygen consumption was monitored in a
two-channel oxygen-uptake apparatus already described elsewhere.39,40
2,2,5,7,8-Pentamethyl-6-chromanol (PMHC) was used as reference
antioxidant.39,40 From the slope of the oxygen consumption during the
inhibited period (Rinh), kinh values were obtained by using eq 7, where
R0 is the rate of oxygen consumption in the absence of antioxidants,
2kt is the bimolecular termination rate constant of styrene (4.2 × 107
M−1 s−1), and n is the stoichiometric coefficient of the antioxidant,
1
calculated for C16H18Se2 369.6739, observed 369.6735. The H NMR
spectrum is in good agreement with that presented in the literature;36
the 13C NMR spectrum has not been reported to date.
9-Triptycene Phenethyl Selenide (4). 9-Bromotriptycene13 (1.8 g,
5.4 mmol) was dissolved in dry benzene (100 mL), and dry methyl
tert-butyl ether (70 mL), cooled to −18 °C, and n-butyllithium (3.8
mL, 5.4 mmol) were added dropwise. Bis(phenethyl) diselenide (3;
1576
dx.doi.org/10.1021/ja411493t | J. Am. Chem. Soc. 2014, 136, 1570−1578

Journal of the American Chemical Society
Article
which was determined experimentally from the length of the inhibited
period (τ) by eq 8.39,40 When the inhibited period was not clearly
visible, kinetic data were confirmed by fitting the experimental traces
with numerical simulations using Gepasi 3.0 software, as previously
described.41
(3) Goto, K.; Nagahama, M.; Mizushima, T.; Shimada, K.;
Kawashima, T.; Okazaki, R. Org. Lett. 2001, 3, 3569−3572.
(4) Jones, D. N.; Mundy, D.; Whitehouse, R. D. J. Chem. Soc. D 1970,
86.
(5) Sharpless, K. B.; Lauer, R. F. J. Am. Chem. Soc. 1973, 95, 2697−
2699.
(6) Reich, H. J.; Renga, J. M.; Reich, I. L. J. Am. Chem. Soc. 1975, 97,
5434−5447.
(7) Flohe, L.; Gunzler, W. A.; Schock, H. H. FEBS Lett. 1973, 32,
132−134.
(R0/Rinh) − (Rinh/R0) = nkinh[AH]/√(2ktRi)
(7)
n = τRi/[AH]
(8)
Deuterium kinetic isotope effects were determined by comparing
inhibited autoxidations recorded upon addition of 1% v/v H2O or
D2O.39,40
(8) Rotruck, J. T.; Pope, A. L.; Ganther, H. E.; Swanson, A. B.;
Hafeman, D. G.; Hoekstra, W. G. Science 1973, 179, 588−590.
(9) Bhabak, K. P.; Mugesh, G. Acc. Chem. Res. 2010, 43, 1408−1419.
(10) Reich, H. J.; Jasperse, C. P. J. Org. Chem. 1988, 53, 2389−2390.
(11) Saiki, T.; Goto, K.; Okazaki, R. Angew. Chem., Int. Ed. Engl.
1997, 36, 2223−2224.
IR Measurements. Spectra of 1 and 2 were recorded at 298 K in a
FT-IR spectrometer under a nitrogen atmosphere using a sealed KBr
cell with an optical path of 0.5 mm. Solutions of the test compound
(1−10 mM) in CCl4 and in CCl4/CH3CN mixtures were analyzed in
absorbance mode referenced to the blank spectrum of the
corresponding solvent mixture. The integrated signal of the “free”
O−H stretching mode at ca. 3530 cm−1, obtained after manual
baseline correction, was plotted vs the concentration of acetonitrile
and fit to eq 9.
(12) Ishii, A.; Matsubayashi, S.; Takahashi, T.; Nakayama, J. J. Org.
Chem. 1999, 64, 1084−1085.
(13) McGrath, A. J.; Garrett, G. E.; Valgimigli, L.; Pratt, D. A. J. Am.
Chem. Soc. 2010, 132, 16759−16761.
(14) Amorati, R.; Lynett, P. T.; Valgimigli, L.; Pratt, D. A. Chem. Eur.
J. 2012, 18, 6370−6379.
(15) Vaidya, V.; Ingold, K. U.; Pratt, D. A. Angew. Chem., Int. Ed.
2009, 48, 157−160.
[1 or 2]free = [1 or 2]tot /(1 + Ksolv × [Solv])
(9)
From the measured equilibrium constant Ksolv, the corresponding α2H
values were obtained by Abraham’s eq 10,27 using the revised value of
βH2 = 0.39 for acetonitrile.44
(16) Lynett, P. T.; Butts, K.; Vaidya, V.; Garrett, G. E.; Pratt, D. A.
Org. Biomol. Chem. 2011, 9, 3320−3330.
(17) Paulsen, C. E.; Carroll, K. S. Chem. Rev. 2013, 113, 4633−4679.
(18) Lucarini, M.; Pedulli, G. F.; Cipollone, M. J. Org. Chem. 1994,
59, 5063−5070.
log(Ksolv/M−1) = 7.354α2HβH − 1.094
(10)
2
(19) Mulder, P.; Korth, H.-G.; Pratt, D. A.; DiLabio, G. A.;
Valgimigli, L.; Pedulli, G. F.; Ingold, K. U. J. Phys. Chem. A 2005, 109,
2647−2655.
Computations. All calculations were carried out using the
Gaussian 09 quantum chemistry package51 using the CBS-QB3
complete basis set approach,20 except for the calculated transition state
structures shown in Figure 10, which were obtained using B3LYP and
a 6-31+G(2d,2p) basis set including the dispersion-correcting
potentials of DiLabio and Torres.30 Wiberg bond order analysis52
was carried out using the AOMix software developed by Gorelsky.53
(20) Montgomery, J. A.; Frisch, M. J.; Ochterski, J. W.; Petersson, G.
A. J. Chem. Phys. 1999, 110, 2822.
(21) Gibson, S. T. J. Chem. Phys. 1986, 85, 4815−4824.
(22) Leeck, D. T.; Li, R.; Chyall, L. J.; Kenttamaa, H. I. J. Phys. Chem.
̈
1996, 100, 6608−6611.
(23) Janousek, B. K.; Reed, K. J.; Brauman, J. I. J. Am. Chem. Soc.
1980, 102, 3125−3129.
ASSOCIATED CONTENT
* Supporting Information
■
S
(24) Burton, G. W.; Doba, T.; Gabe, E.; Hughes, L.; Lee, F. L.;
Prasad, L.; Ingold, K. U. J. Am. Chem. Soc. 1985, 107, 7053−7065.
(25) Snelgrove, D. W.; Lusztyk, J.; Banks, J. T.; Mulder, P.; Ingold, K.
U. J. Am. Chem. Soc. 2001, 123, 469−477.
NMR spectra of new compounds, representative voltammo-
grams, Cartesian coordinates, and energies of stationary points
determined by computation. This material is available free of
charge via the Internet at http://pubs.acs.org.
(26) Litwinienko, G.; Ingold, K. U. Acc. Chem. Res. 2007, 40, 222−
230.
(27) Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Morris, J. J.;
Taylor, P. J. J. Chem. Soc., Perkin Trans. 2 1990, 521.
(28) Valgimigli, L.; Pratt, D. A. In Encyclopedia of Radicals in
Chemistry, Biology and Materials; Wiley: Chichester, U.K., 2012, Vol. 3,
pp 1623−1677.
AUTHOR INFORMATION
Corresponding Author
luca.valgimigli@unibo.it; dpratt@uottawa.ca
■
Notes
(29) Ishii, A.; Komiya, K.; Nakayama, J. J. Am. Chem. Soc. 1996, 118,
12836−12837.
The authors declare no competing financial interest.
(30) Torres, E.; DiLabio, G. A. J. Phys. Chem. Lett. 2012, 3, 1738−
1744.
(31) DiLabio, G. A.; Johnson, E. R. J. Am. Chem. Soc. 2007, 129,
ACKNOWLEDGMENTS
■
We are grateful to the Natural Sciences and Engineering
Research Council of Canada and the Canada Foundation for
Innovation for grants to D.A.P. andthe Italian MIUR, Rome
(PRIN2011 2010PFLRJR (PROxi project)) for a grant to L.V.
and R.A. The computations described herein were made
possible by the High Performance Computing Virtual
Laboratory. D.A.P. also acknowledges the support of the
Canada Research Chairs program.
6199−6203.
(32) Valgimigli, L.; Amorati, R.; Petrucci, S.; Pedulli, G. F.; Hu, D.;
Hanthorn, J. J.; Pratt, D. A. Angew. Chem., Int. Ed. 2009, 48, 8348−
8351.
(33) Hanthorn, J. J.; Valgimigli, L.; Pratt, D. A. J. Am. Chem. Soc.
2012, 134, 8306−8309.
(34) Hanthorn, J. J.; Amorati, R.; Valgimigli, L.; Pratt, D. A. J. Org.
Chem. 2012, 77, 6895−6907.
(35) Amorati, R.; Pedulli, G. F.; Pratt, D. A.; Valgimigli, L. Chem.
Commun. 2010, 46, 5139.
REFERENCES
■
(36) Thompson, D. P.; Boudjouk, P. J. Org. Chem. 1988, 53, 2109−
2112.
(1) Rappoport, Z. The Chemistry of Organic Selenium and Tellurium
Compounds; Wiley: Hoboken, NJ, 2012.
(37) Lucarini, M.; Pedrielli, P.; Pedulli, G. F.; Cabiddu, S.; Fattuoni,
C. J. Org. Chem. 1996, 61, 9259−9263.
(2) Mugesh, G.; du Mont, W.-W.; Sies, H. Chem. Rev. 2001, 101,
2125−2180.
1577
dx.doi.org/10.1021/ja411493t | J. Am. Chem. Soc. 2014, 136, 1570−1578

Journal of the American Chemical Society
Article
(38) Shanks, D.; Amorati, R.; Fumo, M. G.; Pedulli, G. F.; Valgimigli,
L.; Engman, L. J. Org. Chem. 2006, 71, 1033−1038.
(39) Lucarini, M.; Pedulli, G. F.; Valgimigli, L.; Amorati, R.; Minisci,
F. J. Org. Chem. 2001, 66, 5456−5462.
(40) Amorati, R.; Pedulli, G. F.; Valgimigli, L. Org. Biomol. Chem.
2011, 9, 3792−3800.
(41) Valgimigli, L.; Amorati, R.; Fumo, M. G.; DiLabio, G. A.;
Pedulli, G. F.; Ingold, K. U.; Pratt, D. A. J. Org. Chem. 2008, 73, 1830−
1841.
(42) The increased O−H BDE in 1 in comparison to t-BuSOH is
likely the result of the three sp2 carbons bound to the quaternary
carbon in the former being more inductively withdrawing than the sp3
carbons in the latter, stabilizing the starting material and destabilizing
the radical from 1 relative to t-BuSOH.
(43) The CCSD(T), MP2, and MP4 steps of the CBS-QB3
calculation each predict a spin distribution of ca. 0.2/0.8 (Se/O) in the
selenenyl radical and ca. 0.4/0.6 (S/O) in the sulfinyl radical, whereas
B3LYP predicts roughly 0.5/0.5 in both. The inclusion of effective core
potentials in these calculations did not change the spin distributions to
any significant extent.
(44) Valgimigli, L.; Bartolomei, D.; Amorati, R.; Haidasz, E.;
Hanthorn, J. J.; Nara, S. J.; Brinkhorst, J.; Pratt, D. A. Beilstein J.
Org. Chem. 2013, 9, 2781−2792.
(45) Ingold’s equation refers to the KSE in pure solvents, while
kinetics were measured in 50% v/v styrene/solvent mixtures.
(46) This is unsurprising, given that the calculations are of isolated
molecules and the experiments were carried out in chlorobenzene,
which would slow the reaction slightly due to H bonding; see refs 25
and 26.
(47) It must be pointed out that the DCP-B3LYP approach used to
calculate the TS structures for the reactions of 1 and 2 with
styrylperoxyl radicals also predicts a lower free energy barrier (by 1.2
kcal/mol) for the t-BuSeOH/•OOMe reaction in comparison to the t-
BuSOH/•OOMe reaction, in good agreement with the CBS-QB3
calculations.
(48) This can also be described simply as an H-atom transfer reaction
facilitated by a secondary orbital interaction, which lowers the energy
of the syn TS relative to the anti TS. For a related example, see: Hu,
D.; Pratt, D. A. Chem. Commun. 2010, 46, 3711−3713.
(49) In contrast, the anti TS structures have essentially the same
bond order between the chalcogen atom and the internal oxygen atom
of the peroxyl radical (S/O, 0.096; Se/O, 0.101).
(50) The optimal TS structure is calculated for a Se−O/O−O or S−
O/O−O torsion angle of roughly 20° due to the orbital overlap
between the chalcogen atom and inner peroxyl atom in the HOMO,
which is impossible at a torsion angle of 0°.
(51) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Keith, T.; Kobayashi, R.; Normand, J.;
Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.;
Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.;
Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.;
Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.;
Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador,
P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
revision B.01; Gaussian, Inc., Wallingford, CT, 2009.
(52) Mayer, I. Chem. Phys. Lett. 1983, 97, 270−274. Wiberg, K. B.
Tetrahedron 1968, 24, 1083−1096.
(53) Gorelsky, S. I. AOMix: Program for Molecular Orbital Analysis,
version 6.82, 2013; http://www.sg-chem.net/. Gorelsky, S. I.; Lever, A.
B. P. J. Organomet. Chem. 2001, 635, 187−196.
1578
dx.doi.org/10.1021/ja411493t | J. Am. Chem. Soc. 2014, 136, 1570−1578