On the synergy of Coulombic and chelate effects in bidentate diboranes: Synthesis and anion binding properties of a cationic 1,8-diborylnaphthalene

Article abstract of DOI:10.1021/om2012216

As part of our investigation in the chemistry of bidentate Lewis acids for anion complexation, we have carried out the reaction of 1-(dimesitylboryl)-8- (1′-bora-9′-thia-anthryl)naphthalene (1) with methyltriflate. This reaction proceeds via alkylation of the sulfur atom to afford a bidentate diborane ([2]⁺) decorated by a peripheral sulfonium unit. This new diborane, which has been isolated as the triflate salt, reacts with both fluoride and azide anions to form the corresponding anion chelate complexes 2-μ₂-F and 2-μ₂-N₃, respectively. Titration experiments carried out in chloroform indicate that the fluoride binding constant of [2]⁺ exceeds that of 1 by at least 4 orders of magnitude. These results, which are supported by spectroscopic, structural, and computational data, show that chelate and Coulombic effects are additive and can be combined to boost the anion affinity of bidentate Lewis acids.

Full text of DOI:10.1021/om2012216

Article
pubs.acs.org/Organometallics
On the Synergy of Coulombic and Chelate Effects in Bidentate
Diboranes: Synthesis and Anion Binding Properties of a Cationic
1,8-Diborylnaphthalene
̈
Haiyan Zhao and Franco̧ is P. Gabbaı*
Department of Chemistry, Texas A&M University, College Station, Texas 77843-3255, United States
S
* Supporting Information
ABSTRACT: As part of our investigation in the chemistry of
bidentate Lewis acids for anion complexation, we have carried
out the reaction of 1-(dimesitylboryl)-8-(1′-bora-9′-thia-
anthryl)naphthalene (1) with methyltriflate. This reaction
proceeds via alkylation of the sulfur atom to afford a bidentate
diborane ([2]+) decorated by a peripheral sulfonium unit. This
new diborane, which has been isolated as the triflate salt, reacts
with both fluoride and azide anions to form the corresponding
anion chelate complexes 2-μ2-F and 2-μ2-N3, respectively.
Titration experiments carried out in chloroform indicate that
the fluoride binding constant of [2]+ exceeds that of 1 by at least 4 orders of magnitude. These results, which are supported by
spectroscopic, structural, and computational data, show that chelate and Coulombic effects are additive and can be combined to
boost the anion affinity of bidentate Lewis acids.
or B−F→Te (for [D]+) motif, whose formation illustrates the
INTRODUCTION
■
The chemistry of boron-based bidentate Lewis acids1 is an area
Lewis acidic behavior of the heavy onium moiety. The stability
of the resulting complexes also benefits from strong Coulombic
effects, which prevent dissociation of the anion from the
cationic host. The favorable influence of these Coulombic
effects has also been demonstrated in the case of triarylboranes
decorated by peripheral cationic moieties such as the
phosphonium borane [E]+, which, unlike neutral boranes,
captures F− in aqueous solution.5j
Building on these earlier achievements, we have now decided
to target bidentate diboranes that incorporate a peripheral
cationic functionality. From a simple conceptual viewpoint, we
anticipate that the anion affinity of such system would benefit
from (i) the chelate effect provided by the chelating diborane
moiety; (ii) the Coulombic effect imparted by the presence of a
peripheral cationic group.
of active investigation with application in the domains of anion
complexation,2 organometallic catalysis,3 and small-molecule
activation.4 Since the properties of these bidentate Lewis acids
can be influenced by the respective positions and electronic
features of the binding sites, a great deal of attention has been
dedicated to the synthesis of compounds with juxtaposed
electron-poor boron functionalities.3a,4a Prototypical examples
of such compounds include the fluorinated diboranes A and B,
which have both been shown to chelate small anions.3a,4a
RESULTS AND DISCUSSION
■
Synthesis and Characterization of the Cationic
Bidentate Lewis Acid. Inspired by Katz’s seminal contribu-
tion on the anion affinity of 1,8-dimethylborylnaphthalene,1b,6
our group has investigated the synthesis of numerous
naphthalene-based diboranes7 including 1 (Scheme 1),8
whose fluoride binding constants exceed that of monofunc-
tional analogues by at least 3 or 4 orders of magnitude. With
the synthesis of a cationic bidentate borane as an objective, it
occurred to us that the sulfur atom of the thiaborin9 moiety
In a variant of this approach, several groups, including ours,
have become interested in boron-based bidentate Lewis acids
that bear a cationic group in proximity to the boron atom.3d,5
Examples of such compounds include [C]+ and [D]+, which
have been investigated for the complexation of fluoride
anions.5h,i For these compounds, the coordination of the
anion is supported by the formation of a B−F→Sb (for [C]+)
Received: December 8, 2011
Published: March 9, 2012
© 2012 American Chemical Society
2327
dx.doi.org/10.1021/om2012216 | Organometallics 2012, 31, 2327−2335

Organometallics
Article
Scheme 1. Synthesis of the Sulfonium Diborane of [2]+
determined experimentally, was subjected to a time-dependent
DFT calculation using the PCM solvation model with
chloroform as a solvent. The LUMO of [2]+ is localized on
the cationic and thus inherently electron-deficient sulfonium
boryl moiety, with a dominating contribution from the pz
orbital of the boron atom. The LUMO+1 is centered on the
other boryl moiety, again with a large contribution from the pz
orbital of the boron atom. These orbital characteristics contrast
with those of the neutral precursor 1, for which the LUMO
shows an almost equal contribution from both boron pz orbitals
(Figure 2).8,12 These differences illustrate the asymmetry
induced by the cationic sulfonium moiety, making one side of
the molecule distinctly more electron-deficient than the other
one. TD-DFT calculations indicate that both the LUMO and
LUMO+1 of [2]+ are the main accepting orbitals of the
electronic transitions that contribute to the low-energy band
observed at λmax 349 nm. We have also compared the
electrochemical properties of 1 and [2]+. Unlike the cyclic
voltammogram of 1, which shows two reversible reduction
waves at E1/2 −2.20 V and −2.57 V (vs Fc/Fc+),8b [2]+ displays
an irreversible wave only at Ered −1.66 V (vs Fc/Fc+) in THF.
The potential of this wave, which is much more positive than
that of 1, indicates that [2]+ is substantially more electrophilic
than 1. Its irreversibility also suggests that the resulting neutral
radical is unstable.
could possibly be alkylated. With this in mind, 1 was allowed to
react with MeOTf in refluxing dichloromethane to afford
[2]OTf as a moisture-sensitive compound (Scheme 1). This
salt has been characterized by NMR spectroscopy, UV−vis
spectroscopy, and single-crystal X-ray diffraction. The 1H NMR
spectrum of this compound shows a singlet at 3.60 ppm
corresponding to the methyl group of the sulfonium ion. In
addition, six distinct methyl groups are observed for the mesityl
substituents, indicating that the structure of [2]+ is sterically
congested. Although the S-alkylation of thiaborins has, to our
knowledge, never been reported, we note that related reactions
are known for phosphaborins, which can be easily converted
into the corresponding phosphonium species.10
The crystal structure of [2]OTf shows that (1) both boron
centers B(1) and B(2) adopt a trigonal-planar coordination
geometry, as indicated by the sum of the Caryl−B−Caryl angles
(∑∠C−B(1)−C = 356.02°, ∑∠C−B(2)−C = 359.55°); (2)
the distance separating the two boron atoms (3.276(3) Å) is
almost identical to that of the neutral precursor 1 (3.279(4) Å);
and (3) the sulfur-bound methyl group points outward from
the diboron pocket (Figure 1). The UV−vis spectrum of
Fluoride Anion Complexation. The cationic borane [2]+
quickly reacts with [nBu4N][Ph3SiF2] in CDCl3 to afford the
corresponding fluoride complex 2-μ2-F, as confirmed by NMR
spectroscopy. Synthesis of this fluoride adduct can also be
carried out in CH2Cl2 by reaction of [2]OTf with [(Me2N)3S]-
[Me3SiF2]. Adduct 2-μ2-F, which is air and moisture stable, has
been fully characterized. The 11B NMR resonances at −0.4 and
4.4 ppm are consistent with the presence of two four-
coordinate boron centers.8a The H NMR resonance of the
1
sulfur-bound methyl group appears at 3.28 ppm, upfield from
that of [2]+ at 3.60 ppm. The 19F NMR signal at −174.4 ppm is
comparable to the chemical shift observed in other fluoride
chelate complexes3a including [1-μ2-F]− (−188.7 ppm).8a The
structure of 2-μ2-F has also been determined by single-crystal
X-ray diffraction (Figure 3). In contrast with the relatively
symmetrical B−F−B bridge of [1-μ2-F]− (B−F bond lengths =
1.585(5) and 1.633(5) Å),8a the structure of 2-μ2-F shows that
the fluorine atom forms a short bond with B(1) (1.539(4) Å)
and a long one with B(2) (1.822(4) Å). This unsymmetrical
coordination is logically reflected by a notable difference in the
extent of pyramidalization observed for each boron center
(∑(∠C−B(1)−C) = 334.6° and ∑(∠C−B(2)−C) = 350.6°).
The asymmetry of the B−F−B bridge in 2-μ2-F can be
correlated to the presence of a sulfonium moiety, which
enhances the Lewis acidity of the adjacent B(1) atom. This
conclusion is in agreement with the computational studies of
[2]+ that show that the pz orbital of the B(1) boron atom
contributes to the LUMO, while that of the B(2) atom
contributes to the LUMO+1. It is also interesting to note that
the sulfur-bound methyl group is oriented inward. This
surprising orientation places the C(41) atom at 3.452 Å from
the fluorine atom, which is above the range that one would
expect if a CMe−H---F interaction was present.
Figure 1. ORTEP drawing of [2]+ with thermal ellipsoid plots (50%
probability). For clarity, the hydrogen atoms and triflate anion are
omitted, and the mesityl subsitutents are represented by a thin line.
Selected bond lengths (Å) and angles (deg): S(1)−C(23) 1.796(2),
C(1)−B(1) 1.568(3), C(17)−B(1) 1.569(3), C(11)−B(1) 1.564(3),
C(24)−B(2) 1.576(3), C(33)−B(2) 1.578(3), C(10)−B(2) 1.567(3);
C(11)−B(1)−C(1) 117.23(18), C(11)−B(1)−C(17) 117.97(18),
C(1)−B(1)−C(17) 120.82(18), C(10)−B(2)−C(24) 124.41(18),
C(10)−B(2)−C(33) 116.85(17), C(24)−B(2)−C(33) 118.29(17),
C(6)−C(10)−B(2) 128.27(17), C(6)−C(1)−B(1) 131.89(17).
[2]OTf in CHCl3 exhibits a band centered at λmax = 349 nm.11
This band, whose position is close to that observed for 1
(λmax = 363 nm in THF),8a originates from the trigonal-planar
boron-centered chromophores, thus implying that the triflate
counteranion does not associate with [2]+ in solution. To
confirm this photophysical assignment, the structure of [2]+ has
been computationally optimized using DFT methods (func-
tional: B3LYP; mixed basis set: B: 6-31+G(d′); S: 6-31+G(d);
C, H: 6-31G). The optimized geometry, which is close to that
The structure of 2-μ2-F has been computationally optimized
using DFT methods (functional: B3LYP; mixed basis set: B: 6-
31+G(d′); S: 6-31+G(d); C, H: 6-31G). Interestingly, the
optimized structure shows a large deviation from that observed
experimentally (Figure 4). The largest deviation is observed in
2328
dx.doi.org/10.1021/om2012216 | Organometallics 2012, 31, 2327−2335

Organometallics
Article
Figure 2. Top: Experimental (chloroform) and calculated UV−vis spectra for [2]+. Bottom: Views of LUMO and LUMO+1 of [2]+ and LUMO of 1
(isovalue = 0.035).
a
Scheme 2. Reactions of [2]+ with Different Anions in Chloroform
a
All charges appearing in the scheme are formal charges.
Figure 3. ORTEP drawings of 2-μ2-F and 2-μ2-N3 with thermal ellipsoid plots (50% probability). For clarity, the hydrogen atoms are omitted and
the mesityl subsitutents are represented by a thin line. Selected bond lengths (Å) and angles (deg) for 2-μ2-F: F(1)−B(1) 1.540(5), F(1)−B(2)
1.822(4), C(2)−B(2) 1.608(6), C(3)−B(2) 1.626(5), C(4)−B(2) 1.599(6), C(10)−B(1) 1.622(6), C(11)−B(1) 1.631(6), C(29)−B(1) 1.604(6),
S(1)−C(41) 1.760(4); B1−F1−B2 124.1(2), C(29)−B(1)−C(10) 112.8(3), C(29)−B(1)−C(11) 110.3(3), C(10)−B(1)−C(11) 111.5(3), C(9)−
C(29)−B(1) 125.5(3), C(4)−B(2)−C(2) 112.3(3), C(4)−B(2)−C(3) 120.3(3), C(2)−B(2)−C(3) 118.0(3), C(9)−C(4)−B(2) 125.2(3).
Selected bond lengths (Å) and angles (deg) for 2-μ2-N3: N(1)−N(2) 1.243(5), N(1)−B(1) 1.643(6), N(1)−B(2) 1.704(6), N(2)−N(3) 1.136(5),
C(1)−B(1) 1.621(7), C(11)−B(1) 1.637(6), C(22)−B(1) 1.626(7), C(24)−B(2) 1.653(6), C(8)−B(2) 1.622(6), C(33)−B(2) 1.652(6), S(1)−
C(23) 1.714(7); B(1)−N(1)−B(2) 124.3(3), C(9)−C(1)−B(1) 125.8(4), C(9)−C(8)−B(2) 124.4(4), C(1)−B(1)−C(11) 111.0(4), C(1)−
B(1)−C(22) 111.8(4), C(11)−B(1)−C(22) 112.1(4), C(8)−B(2)−C(33) 115.9(4), C(8)−B(2)−C(24) 108.7(3), C(33)−B(2)−C(24) 119.5(4).
2329
dx.doi.org/10.1021/om2012216 | Organometallics 2012, 31, 2327−2335

Organometallics
Article
Figure 4. (a) Optimized structure of 2-μ2-F. Calculated bond lengths (Å) and angles (deg) for 2-μ2-F: F(1)−B(1) 1.486, F(1)−B(2) 2.393, C(2)−
B(2) 1.604, C(3)−B(2) 1.601, C(4)−B(2) 1.589, C(10)−B(1) 1.648, C(11)−B(1) 1.658, C(29)−B(1) 1.625, S(1)−C(41) 1.836; B1−F1−B2
116.9, C(29)−B(1)−C(10) 113.0, C(29)−B(1)−C(11) 109.6, C(10)−B(1)−C(11) 110.2, C(9)−C(29)−B(1) 127.2, C(4)−B(2)−C(2) 116.2,
C(4)−B(2)−C(3) 122.0, C(2)−B(2)−C(3) 119.3, C(9)−C(4)−B(2) 129.0. (b) Enegy changes based on different B(2)−F(1) lengths in the
optimization calculations of 2-μ2-F. (c) AIM electron density maps with relevant bond paths and bond critical points of the B(1)−F(1)−B(2) bond
in 2-μ2-F (left) and the B(1)−N(1)−B(2) bond in 2-μ2-N3 (right).
the B(2)−F(1) bond, which converges to a value of 2.393 Å as
opposed to 1.822(4) Å in the crystal structure. To elucidate the
origin of this difference, we decided to study how variations in
the B(2)−F(1) bond length affect the total energy of the
structure. To this end, the structure of 2-μ2-F has been
optimized with the B(2)−F(1) constrained at 1.600, 1.822,
2.000, 2.200, 2.400, 2.600, 2.800, 3.000, and 3.200 Å. As it can
been seen from Figure 4, the total energy of the molecule varies
by only 1.95 kcal/mol in the 1.822 < B(2)−F(1) < 2.800 Å
range. This shallow energy well indicates that the B(2)−F(1)
bond length has a limited effect on the total energy of the
molecule, the structure of which can therefore be influenced by
subtle solvation or crystal packing effects.
A geometry optimization carried out by constraining the
B(2)−F(1) bond length to its experimental value (B(2)−F(1) =
1.822 Å) afforded a structure that lies only 1.95 kcal/mol above
the computed minimum with a B(1)−F(1) bong length of
1.555 Å (Figure 4). The latter is in good agreement with the
B(1)−F(1) bond length of 1.540(5) Å determined by X-ray
diffraction (Figure 3). An atom in molecules (AIM)
calculation13 carried out at this constrained geometry identifies
a bond path for both the B(1)−F(1) and the B(2)−F(2)
linkages (Figure 4). The electron density ρ(r) of 0.100 e bohr−3
at the bond critical point (BCP) of the B(1)−F(1) bond is
significantly larger than that of the B(2)−F(1) bond (0.053
e bohr−3), in agreement with the observed asymmetry of the
B−F−B bridge. By contrast, AIM calculations carried out at the
optimized geometry of the [1-μ2-F]− indicate a much more
symmetrical B−F−B bridge, with similar electron density at the
BCP of the B(1)−F(1) bond (ρ(r) = 0.074 e bohr−3) and
B(2)−F(1) bond (ρ(r) = 0.077 e bohr−3).
Another point that we decided to address computationally is
the change observed in the orientation of the sulfur-bound
methyl group, which switches from outward in [2]+ to inward
in 2-F. A priori, this change could occur by inversion of the
sulfonium atom after fluoride complexation (scenario 1,
Scheme 3). Alternatively, a fluoride anion could first bind
externally to the boron atom of the thiaborinium fragment,
followed by rotation of the resulting fluoroborate moiety about
the B−C1 bond connecting the thiaborinium to the
naphthalene backbone (scenario 2, Scheme 3). To test the
first scenario, we carried out a transition-state calculation
(functional: B3LYP; mixed basis set: B: 6-31+G(d′); S: 6-
31+G(d); C, H: 6-31G), which indicates that the energy barrier
to inversion of the sulfonium unit is equal to 27.11 kcal/mol.
For the second scenario we scanned the energy of the complex
as a function of the C0−C1−B−C2 torsional angle (denoted a τ,
Scheme 3) between the values of 101.3° found by optimization
of the exo/nonchelate complex 2-F and −103.6° found by the
unconstrained optimization of the chelate complex 2-μ2-F. To
this end, the structure of the molecule was optimized with the
angle τ constrained to 40°, 20°, 10°, 0°, −10°, −20°, and −40°.
The energies of the resulting structures, which are plotted in
Figure 5, show that this rotation is a relatively low-energy
pathway, the barrier of which does not exceed 14 kcal/mol.
This barrier to rotation is almost 50% lower than the barrier
determined for the inversion of the sulfonium unit.
Accordingly, we propose that the change observed in the
2330
dx.doi.org/10.1021/om2012216 | Organometallics 2012, 31, 2327−2335

Organometallics
Article
where K(borane) is the absolute fluoride binding constant of
the borane under study and K(Ph3SiF) is the absolute fluoride
binding constant of Ph3SiF (see Experimental Section for details
of the derivation). Krel(1) could be derived from the integrated
1H NMR and 19F NMR spectra of CDCl3 solutions containing 1
and [nBu4N][Ph3SiF2] in different molar ratios, affording Krel(1)
= 5.4 ( 1) (see SI for details of the experiments). Analogous
NMR experiments carried out with [2]+ indicated the quantitative
formation of 2-μ2-F. For this reason, we decided to resort to UV−
vis spectroscopy, which is more appropriate to study large binding
constants. However, addition of [nBu4N][Ph3SiF2] to a dilute
chloroform solution of [2]+ (6.7 × 10−5 M in CHCl3) resulted in
the stoichiometric quenching of the absorbance of the boron-
centered chromophore at 349 nm, in agreement with the
quantitative formation of the fluoride complex (Figure 6). These
observations indicate that the relative fluoride binding constant
of [2]+ (Krel([2]+) exceeds the value of 105 and is thus at least 4
orders of magnitude greater than that of 1. The drastic difference
observed in the fluoride binding properties of 1 and [2]+
underscores the higher fluoride affinity of [2]+, which can be
assigned to its cationic nature. Although 2-μ2-F is very stable, it
slowly reacts with B(C6F5)3 to afford [2][FB(C6F5)3], thus
indicating that fluoride binding by [2]+ is reversible. A similar
reaction was observed for [1-μ2-F]−, which also releases fluoride
to B(C6F5)3.8a
Scheme 3. The Two Different Scenarios Envisaged for the
Switching of the Methyl Group from Outward in [2]+ to
Inward in 2-μ2-F
a
a
All charges appearing in the scheme are formal charges.
orientation of the sulfur-bound methyl group upon conversion
of [2]+ into 2-μ2-F occurs via scenario 2. These computa-
tional results also indicate that rotations of the groups at the
peri-positions of these naphthalene systems are not energeti-
cally prohibited despite the large steric crowding affecting these
compounds.
To further understand the influence of the cationic sulfonium
moiety in [2]+, we decided to measure the fluoride binding
constant of [2]+ and compare it to that of 1. Since we have
previously shown that the binding constant of neutral 1 exceeds
the range measurable by a direct titration with tetrabutylam-
monium fluoride (nBu4NF),8a we decided to employ the
commercially available tetrabutylammonium difluorotriphenyl-
silicate [nBu4N][Ph3SiF2] as fluoride source and determine the
relative binding constant Krel(borane) = K(borane)/K(Ph3SiF),
We also decided to compare the fluoride anion affinity of
[2]+ to that of another cationic borane. For the purpose of this
study, we selected the phosphonium borane [3]+, which was
previously shown to react with fluoride to afford 3-F.5b
Interestingly, mixing equimolar amounts of [2]+ and 3-F in
CDCl3 at ambient temperature leads to the quantitative
formation of [3]+ and 2-μ2-F (Scheme 4). Fluoride ion transfer
1
from 3-F to [2]+ is confirmed by H, 19F, and 31P NMR
spectroscopy. Specifically, the 31P NMR spectrum shows full
conversion of 3-F (28.3 ppm) into [3]+ (23.9 ppm).
Accordingly, the 19F NMR spectrum indicates formation of
2-μ2-F, the resonance of which appears at −174.4 ppm. Related
changes are also observed in the 1H NMR spectrum, where the
Figure 5. Relative energy of 2-F as a function of the torsional angle τ as defined in Scheme 3. The first point of the graph (left) corresponds to the
exo/nonchelate complex 2-F, the energy of which is taken as the reference (τ = 101.3°). The last point of the graph (right) corresponds to 2-μ2-F
(τ = 103.6°). The structure shown in the middle is that obtained for τ = −10°. The fitted line is arbitrary and does not accurately map the energy
profile of the process.
2331
dx.doi.org/10.1021/om2012216 | Organometallics 2012, 31, 2327−2335

Organometallics
Article
Figure 6. Left: Absorbance changes upon addition of [nBu4N][Ph3SiF2] to a CHCl3 solution of [2]+ (0.067 mM). Right: Binding isotherm
monitored at λ = 349 nm (Krel([2]+ ≥ 105, ε = 10 040 for [2]OTf, ε = 3270 for 2-μ2-F).
a
Scheme 4. Competition Reaction of [2]+ with 3-F in CDCl3
group. Presumably the cumulative Coulombic and chelate
effects occurring with this compound are responsible for this
enhanced fluoride affinity.
Reaction with Other Anions. To conclude this study, we
decided to probe the reactivity of [2]+ toward other anions.
Interestingly, addition of nBu4Cl, nBu4Br, or nBu4I to a CDCl3
solution of [2]+ resulted in the formation of the neutral
diborane 1, indicating demethylation of the diarylmethylsulfo-
nium (Scheme 2). Such demethylation reactions are not
unprecedented and have been observed in related sulfonium
borane species.3d,14 The reactivity of [2]+ toward cyanide was
also tested using nBu4CN as a cyanide source. In this case,
however, the reaction afforded a mixture of compounds that
could not be unambiguously identified. A cleaner anion com-
plexation reaction was observed with nBu4N3 in CDCl3, leading
to the corresponding azide complex 2-μ2-N3. Formation of the
latter, which can be reverted by addition of B(C6F5)3, could be
monitored by UV−vis spectroscopy to afford a binding
constant exceeding 107 M−1 (Figure 7).
a
All charges appearing in the scheme are formal charges.
P- or S-bound methyl signals of [2]+ (3.60 ppm) and 3-F
(3.12 ppm) disappear to give rise to those of 2-μ2-F (3.28 ppm)
and [3]+ (2.67 ppm). This reaction indicates that [2]+ is one of
the most Lewis acidic cationic boranes ever investigated in our
The azide complex 2-μ2-N3, which could be isolated from the
reaction of [2]OTf with nBu4N3 in dichloromethane, has been
fully characterized. It features a broad 11B NMR resonance
Figure 7. Left: Absorbance changes upon addition of nBu4NN3 to a CHCl3 solution of [2]+ (0.095 mM). Right: Binding isotherm monitored at λ =
349 nm (K ≥ 107 M−1, ε = 10 040 for [2]OTf, ε = 3100 for 2-μ2-N3).
2332
dx.doi.org/10.1021/om2012216 | Organometallics 2012, 31, 2327−2335

Organometallics
Article
(Mo Kα radiation, λ = 0.71069 Å) for [2]OTf, 2-μ2-F, and 2-μ2-N3. In
each case, a specimen of suitable size and quality was selected and
mounted onto a nylon loop. The structures were solved by direct
methods, which successfully located most of the non-hydrogen atoms.
Subsequent refinement on F2 using the SHELXTL/PC package
(version 5.1) allowed location of the remaining non-hydrogen atoms.
Synthesis of [2]OTf. Methyl triflate (0.1 mL, 0.89 mmol) was added
to a solution of diborane 2 (0.2 g, 0.35 mmol) in dichloromethane
(10 mL) at room temperature. The mixture was refluxed overnight and
then cooled to room temperature. The solvent was removed under
vacuum to yield a solid, which was washed with diethyl ether to afford
[2]OTf as a pale yellow product (0.21 g, yield 81%). Single crystals of
[2]OTf-CH2Cl2 were obtained by slow diffusion of diethyl ether into a
at −4.6 ppm, which, we speculate, corresponds to the
overlapping signals of the two four-coordinate boron centers.
In line with the coordination of an anion to the diborane, the
resonance of the sulfur-bound methyl group at 3.26 ppm is
close to that measured for 2-μ2-F (3.28 ppm). The crystal
structure of 2-μ2-N3 confirms formation of a chelate complex,
with the azide anion bridging the two boron centers in a μ−η1
fashion (Figure 3). While the structure of 2-μ2-F showed a
rather unsymmetrical B−F−B bridge, the N(1)−B(1) (1.635
Å) and N(1)−B(2) (1.706 Å) bond lengths in 2-μ2-N3 indicate
a rather symmetrical B−N−B bridge, which is consistent with
the propensity of the azide anion to adopt a μ−η1 coordination
mode. The DFT-optimized structure 2-μ2-N3 (functional:
B3LYP; mixed basis set: B, N: 6-31+G(d′); S: 6-31+G(d); C,
H: 6-31G) features a B(1)−N(1) distance of 1.644 Å, which is
very close to that experimentally observed. The calculated
N(1)−B(2) distance of 1.832 Å is slightly elongated when
compared to that experimentally observed, a phenomenon
reminiscent of that observed in the computed structure of 2-μ2-
F. Presumably, small variation of this bond length has only a
limited effect on the total energy of the molecule, making subtle
solvation or crystal packing effects the main culprit for this
small discrepancy. An AIM analysis carried out at the optimized
geometry affords consistent results with BCP electron densities
ρ(r) = 0.108 e bohr−3 for N(1)−B(1) and ρ(r) = 0.070
e bohr−3 for N(1)−B(2), again illustrating the increased acidity
of the sulfonium-decorated boryl moiety (Figure 4).
1
dichloromethane solution of [2]OTf at −25 °C. H NMR (CDCl3): δ
0.95 (s, 3H, Mes-CH3), 1.39 (s, 3H, Mes-CH3), 1.47 (s, 3H, Mes-CH3),
1.82 (s, 3H, Mes-CH3), 1.84 (s, 3H, Mes-CH3), 2.23 (s, 3H, Mes-CH3),
3.60 (s, 3H, S-CH3), 5.71 (s, 1H, Mes-CH), 6.59 (s, 1H, Mes-CH), 6.60
(s, 1H, Mes-CH), 6.69 (s, 1H, Mes-CH), 7.39 (d, 1H, 3J(HH) = 7.2 Hz,
nap-CH), 7.46 (t, 1H, 3J(HH) = 7.2 Hz, nap-CH), 7.52 (d, 1H, 3J(HH) =
3
7.6 Hz, nap-CH), 7.53−7.63 (m, 3H, nap-CH), 7.77 (t, 1H, J(HH) =
3
8.0 Hz, CH), 7.82 (t, 1H, J(HH) = 7.6 Hz, CH), 7.88−7.94 (m, 2H,
CH), 8.04 (d, 1H, 3J(HH) = 8.4 Hz, CH), 8.13 (d, 1H, 3J(HH) = 8.0 Hz,
CH), 8.18 (d, 1H, 3J(HH) = 8.0 Hz, CH), 8.23 (d, 1H, 3J(HH) = 8.0 Hz,
CH). 13C NMR (CDCl3): δ 21.00 (Mes-CH3), 21.11 (Mes-CH3), 21.19
(Mes-CH3), 22.11 (Mes-CH3), 23.74 (Mes-CH3), 23.89 (Mes-CH3),
25.26 (S-CH3), 124.39, 126.51, 126.87, 127.40, 128.20, 128.29, 128.96,
130.12, 130.85, 130.88, 131.70, 132.28, 132.48, 132.51, 132.99, 133.13,
133.40, 133.98, 134.22, 135.93, 137.71, 138.64, 139.18, 139.76, 140.44,
141.47, 142.80, 143.49, 145.42, 147.93. 11B NMR (CDCl3): not detected.
Anal. Calcd for C42H39B2F3O3S2·CH2Cl2: C 63.03; H 5.04. Found: C
65.74; H 5.28. (The sample used for EA was obtained by recrystallization
from CH2Cl2; the EA results indicate ∼50% loss of interstitial CH2Cl2
molecule. The calculated values for C42H39B2F3O3S2·1/2(CH2Cl2) are C
65.70; H 5.19.)
CONCLUSION
■
As demonstrated by our earlier work on naphthalene-based
diboranes,2a the fluoride binding constants of neutral bidentate
derivatives exceed those of their monofunctional analogues by 3
or 4 orders of magnitude. The results presented in this paper
show that the fluoride anion affinity of such diboranes can be
further enhanced by the simple introduction of a cationic
moiety in the proximity of one of the boron atoms. This
conclusion is substantiated by the observation that the fluoride
binding constant of [2]+ exceeds that of its neutral precursor by
at least 4 orders of magnitude. A more general lesson that can
be derived from this work is that chelate and Coulombic effects
are additive and can be combined to boost the anion affinity of
bidentate Lewis acids.
Synthesis of 2-μ2-F. To a solution of [2]OTf (50 mg, 0.068 mmol)
in CH2Cl2 (5 mL) at 25 °C was added a CH2Cl2 solution
(5 mL) of [(Me2N)3S][Me3SiF2] (21 mg, 0.076 mmol). After 15 min,
the solvent was evaporated and the residue was dissolved in a mixture
of CH2Cl2 (2 mL) and diethyl ether (18 mL). The resulting solution
was filtered and evaporated to dryness to afford 2-μ2-F as a colorless
solid (35 mg, 85% yield). Large colorless monocrystals of 2-μ2-F could
be obtained by slow evaporation of an acetone solution of 2-μ2-F at
room temperature. 1H NMR (CDCl3): δ 0.86 (s, 3H, Mes-CH3), 1.63
(s, 3H, Mes-CH3), 1.79 (s, 3H, Mes-CH3), 1.92 (d, 3H, J(H−F) =
5.2 Hz, Mes-CH3), 2.07 (s, 3H, Mes-CH3), 2.23 (s, 3H, Mes-CH3),
3.28 (s, 3H, S-CH3), 5.99 (s, 1H, Mes-CH), 6.37 (s, 1H, Mes-CH),
6.39 (s, 1H, Mes-CH), 6.62−6.66 (m, 2H, CH), 6.67 (s, 1H, Mes-
3
3
CH), 6.78 (d, 1H, J(HH) = 7.2 Hz, CH), 7.00 (t, 1H, J(HH) =
7.6 Hz, CH), 7.12−7.25 (m, 5H, CH), 7.35 (t, 1H, 3J(HH) = 7.6 Hz,
CH), 7.49 (dd, 2H, 3J(HH) = 6.8 Hz, CH), 7.75 (d, 1H, 3J(HH) = 8.0
Hz, CH), 7.87 (d, 1H, 3J(HH) = 8.0 Hz, CH). 13C NMR (CDCl3): δ
20.74 (Mes-CH3), 21.01 (Mes-CH3), 22.89 (Mes-CH3), 24.58 (Mes-
CH3), 25.16 (Mes-CH3), 25.24 (Mes-CH3), 39.23 (S-CH3), 124.20,
125.12, 125.92, 126.47, 126.68, 127.17, 127.47, 127.70, 127.92, 127.95,
128.01, 128.22, 129.16, 123.00 (d, J(CF) = 7.3 Hz), 130.50 (d, J(CF) =
7.1 Hz), 132.24 (d, J(CF) = 6.8 Hz), 132.61 (d, J(CF) = 2.7 Hz),
134.97 (d, J(CF) = 1.9 Hz), 135.10, 140.37, 140.76 (d, J(CF) = 6.2
Hz), 140.98, 141.57, 141.73 (d, J(CF) = 5.9 Hz), 141.84 (d, J(CF) =
6.1 Hz), 142.90. 11B NMR (CDCl3): +0.4 (s), +4.4 (bs). 19F NMR
(CDCl3): −174.4 (s). Anal. Calcd for C41H39B2FS: C 80.47; H 6.50.
Found: C 80.91; H 6.49.
EXPERIMENTAL SECTION
■
General Considerations. Commercially available chemicals were
purchased and used as provided (commercial sources: Aldrich for
Mes2BF, TMEDA, [(Me2N)3S][Me3SiF2], [nBu4N][Ph3SiF2], TMSCl,
nBu4NF, nBu4NCl, nBu4NBr, nBu4NI, and nBu4NN3; TCI America for
Ph2S; Alfa Aesar for BBr3 and n-butyllithium (2.8 M in hexanes)).
Diborane 1 was prepared by reaction of tetrakis(THF)lithium
dimesityl-1,8-naphthalenediylborate7b,15 with 10-bromo-9-thia-10-
boranthracene as previously described.8a Solvents were dried by reflux
under N2 over drying agents and freshly distilled prior to use. The
drying agents employed were CaH2 for dichloromethane and Na/K for
diethyl ether and THF. Air-sensitive compounds were handled under
N2 atmosphere using standard Schlenk and glovebox techniques. UV−
vis spectra were recorded on an Ocean Optics USB4000 spectrometer
with an Ocean Optics ISS light source. Elemental analyses were
performed at Atlantic Microlab (Norcross, GA, USA). NMR spectra
were recorded on Varian Unity Inova 400 FT NMR (399.59 MHz for
1H, 376.03 MHz for 19F, 128.19 MHz for 11B, 100.45 MHz for 13C)
spectrometers at ambient temperature. Chemical shifts δ are given in
ppm and are referenced against external BF3·Et2O (11B and 19F).
Crystallography. The crystallographic measurements were
performed using a Bruker APEX-II CCD area detector diffractometer
Synthesis of 2-μ2-N3. To a solution of [2]OTf (84 mg,
0.11 mmol) in CH2Cl2 (5 mL) at 25 °C was added a CH2Cl2 solution
(5 mL) of nBu4NN3 (35 mg, 0.12 mmol). After 15 min, the solvent
was evaporated and the residue was dissolved in a solvent mixture of
CH2Cl2 (2 mL) and diethyl ether (18 mL). The resulting solution was
filtered and evaporated to dryness to afford 2-μ2-N3 as a colorless solid
(41 mg, 59% yield). Large colorless monocrystals of 2-μ2-N3 could be
obtained by slow evaporation of an acetone/hexane mixture of 2-μ2-N3
1
at room temperature. H NMR (CDCl3): δ 1.25 (s, 3H, Mes-CH3),
2333
dx.doi.org/10.1021/om2012216 | Organometallics 2012, 31, 2327−2335

Organometallics
Article
1.73 (s, 3H, Mes-CH3), 1.81 (s, 3H, Mes-CH3), 2.10 (s, 6H, Mes-
CH3), 2.23 (s, 3H, Mes-CH3), 3.26 (s, 3H, S-CH3), 6.15 (bs, 1H, Mes-
CH), 6.34 (bs, 1H, Mes-CH), 6.38 (d, 2H, J(HH) = 8.0 Hz, CH),
(STQN) method and further confirmed by the intrinsic reaction
coordinate method (see Table S1 and Figure S1 in the Supporting
Information).
3
6.73 (bs, 1H, Mes-CH), 6.86 (bs, 1H, Mes-CH), 6.78 (m, 2H, CH),
3
7.06 (t, 1H, J(HH) = 6.0 Hz, CH), 7.18−7.33 (m, 4H, CH), 7.40−
ASSOCIATED CONTENT
3
■
7.48 (m, 3H, CH), 7.66 (d, 1H, J(HH) = 8.0 Hz, CH), 7.77 (d, 1H,
3J(HH) = 8.0 Hz, CH). 13C NMR (CDCl3): 20.73 (Mes-CH3), 23.87
(Mes-CH3), 25.41 (Mes-CH3), 26.08 (Mes-CH3), 30.95 (Mes-CH3),
41.16 (S-CH3), 123.73, 123.83, 125.03, 125.18, 126.08, 126.15,
126.31, 126.37, 126.86, 127.83, 128.15, 129.07, 129.57, 130.06, 130.68,
131.50, 132.52, 132.87, 133.07, 133.57, 133.95, 139.74, 140.53, 142.75.
11B NMR (CDCl3): −4.6. Anal. Calcd for C41H39B2N3S: C 78.48; H
6.26. Found: C 78.36; H 6.37.
* Supporting Information
S
Cartesian coordinates of the optimized structures and X-ray
data in CIF format. These materials are available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Anion (X−) Complexation in CHCl3. A CHCl3 solution of
Corresponding Author
*E-mail: francois@tamu.edu.
[2]OTf (3 mL, c = 6.73 × 10−5 M for X− = F− and c = 9.5 × 10−5
M
−
for X− = N3 ) was placed in a cuvette and titrated with incremental
Notes
amounts of the anion by addition of a CHCl3 solution of
[nBu4N][Ph3SiF2] for X− = F− (7.2 mM) or nBu4NN3 for X−
=
The authors declare no competing financial interest.
N3− (33.5 mM). For both experiments, the absorbance of the diborane
was monitored at λ = 349 nm, showing stoichiometric complexation of
the anion (ε = 10 040 for [2]OTf, ε = 3270 for 2-μ2-F, ε = 3100 for 2-
μ2-N3).
ACKNOWLEDGMENTS
■
This work was supported by the National Science Foundation
(CHE-0952912), the Welch Foundation (A-1423), and Texas
A&M. We thank Ching-Wen Chiu for her original contribution
to this project.
NMR Study of the Reaction of Diborane 1 with [nBu4N]-
[Ph3SiF2] in CDCl3. In a typical assay, a mixture of diborane 1 and
[nBu4N][Ph3SiF2] in CDCl3 was placed in an NMR tube and analyzed
1
by H NMR and 19F NMR spectroscopy. Integration of the spectra
was used to determine the respective concentration of the relevant
species. Two data sets (labeled as experiments 1 and 2) are provided
below, along with the derivation of the equation used to calculate
Krel(1) (F0 is the initial concentration of Ph3SiF2 , and B0 is the initial
concentration of the diborane 1).
REFERENCES
■
(1) (a) Biallas, M. J.; Shriver, D. F. J. Am. Chem. Soc. 1966, 88, 375−
376. (b) Katz, H. E. Inclusion Compd. 1991, 4, 391−405.
(c) Schmidtchen, F. P.; Berger, M. Chem. Rev. 1997, 97, 1609−
−
̈
1646. (d) Gabbaı, F. P. Angew. Chem., Int. Ed. 2003, 42, 2218−2221.
−
1
+
Ph3SiF
+
2 ×
⇔
1F−
+
Ph3SiF2
F−
̈
̈
(e) Melaımi, M.; Gabbaı, F. P. Adv. Organomet. Chem. 2005, 53, 61−
̈
99. (f) Wang, H.; Sole, S.; Gabbaı, F. P. ACS Symp. Ser. 2006, 917,
B0 −
F0 − y
F0 − x
− y
x
y
208−220.
x
̈
(2) (a) Hudnall, T. W.; Chiu, C.-W.; Gabbaı, F. P. Acc. Chem. Res.
experiment 1 5.1
experiment 2 70
51
51
100
25
−
2009, 42, 388−397. (b) Wade, C. R.; Broomsgrove, A. E. J.; Aldridge,
100
100
̈
S.; Gabbaı, F. P. Chem. Rev. 2010, 110, 3958−3984.
K1 = [1F]/([1]*[F−]) = x/[(B0 − x)(F0 − x − y)]. KPh3SiF2
=
(3) (a) Williams, V. C.; Piers, W. E.; Clegg, W.; Elsegood, M. R. J.;
Collins, S.; Marder, T. B. J. Am. Chem. Soc. 1999, 121, 3244−3245.
(b) Piers, W. E.; Irvine, G. J.; Williams, V. C. Eur. J. Inorg. Chem. 2000,
2131−2142. (c) Lewis, S. P.; Taylor, N. J.; Piers, W. E.; Collins, S.
−
[Ph3SiF2 ]/([Ph3SiF] [F−]) = y/[(F0 − y)(F0 − x − y)] ⇒ Krel(1) =
−
K1/KPh3SiF2 = [x(F0 − y)]/[y(B0 − x)] = 5.4 ( 1).
Electrochemistry. Electrochemical experiments were performed
with an electrochemical analyzer from CH Instruments (model 610A)
with a glassy carbon working electrode, a platinum auxiliary electrode,
and a reference silver electrode. The reference electrode solution was
built by immersing a silver wire in a Vycor-capped glass tube
containing a THF solution of [nBu4N][PF6] (0.1 M) and AgNO3
(0.005 M). All three electrodes were placed in a THF solution (3 mL)
containing [nBu4N][PF6] (0.1 M) as supporting electrolyte and [2]OTf
(0.010 M). Ferrocene was used as an internal standard, and the potentials
are reported relative to the E1/2 of the Fc+/Fc redox couple.
Computational Details. DFT calculations (full geometry
optimization) were carried out with the Gaussian 03 program using
the gradient-corrected Becke exchange functional (B3LYP) and the
Lee−Yang−Parr correlation functional. Geometry optimization was
carried out with the following mixed basis set: 6-31+G(d′) for the
boron, nitrogen, and fluorine atoms, 6-31+G(d) for the sulfur atom,
and the 6-31G basis set was used for the other remaining carbon and
hydrogen atoms. Frequency calculations for [2]+, 2-μ2-F, and 2-μ2-N3
confirmed the absence of any imaginary frequencies. In the case of 2-F
(τ = 101.3°), a weakly negative frequency associated with a methyl
group rotation at the para-position of one of the mesityl groups was
identified. All other structures with the τ angle constrained to specific
values are not minima and were therefore not subjected to frequency
calculations. DT-DFT calculation was carried out with the Gaussian 03
program using the B3LYP functional with the same basis sets as the
DFT calculation. The transition state of the sulfonium unit inversion
reaction was located with the synchronous transit-guided quasi-Newton
̈
J. Am. Chem. Soc. 2003, 125, 14686−14687. (d) Zhao, H.; Gabbaı, F.
P. Org. Lett. 2011, 1444−1446.
(4) (a) Jiang, C.; Blacque, O.; Berke, H. Chem. Commun. 2009,
5518−5520. (b) Zhao, X.; Stephan, D. W. Chem. Commun. 2011, 47,
1833−1835.
̈
(5) (a) Chiu, C.-W.; Gabbaı, F. P. J. Am. Chem. Soc. 2006, 128,
14248−14249. (b) Hudnall, T. W.; Kim, Y.-M.; Bebbington, M. W. P.;
̈
Bourissou, D.; Gabbaı, F. P. J. Am. Chem. Soc. 2008, 130, 10890−
10891. (c) Kim, Y.; Hudnall, T. W.; Bouhadir, G.; Bourissou, D.;
̈
Gabbaı, F. P. Chem. Commun. 2009, 3729−3731. (d) Kim, Y.; Zhao,
̈
H.; Gabbaı, F. P. Angew. Chem., Int. Ed. 2009, 48, 4957−4960.
(e) Agou, T.; Sekine, M.; Kobayashi, J.; Kawashima, T. Chem.Eur. J.
̈
2009, 15, 5056−5062. (f) Wade, C. R.; Zhao, H.; Gabbaı, F. P. Chem.
Commun. 2010, 46, 6380−6381. (g) Broomsgrove, A. E. J.; A. Addy,
D.; Di Paolo, A.; Morgan, I. R.; Bresner, C.; Chislett, V.; Fallis, I. A.;
Thompson, A. L.; Vidovic, D.; Aldridge, S. Inorg. Chem. 2010, 49,
̈
157−173. (h) Wade, C. R.; Gabbaı, F. P. Organometallics 2011, 30,
̈
4479−4481. (i) Zhao, H.; Gabbaı, F. P. Nat. Chem. 2010, 2, 984−990.
̈
(j) Kim, Y.; Gabbaı, F. P. J. Am. Chem. Soc. 2009, 131, 3363−3369.
(6) (a) Katz, H. E. J. Org. Chem. 1985, 50, 5027−5032. (b) Katz, H.
E. J. Am. Chem. Soc. 1985, 107, 1420−1421. (c) Katz, H. E.
Organometallics 1987, 6, 1134−1136.
̈
(7) (a) Hoefelmeyer, J. D.; Gabbaı, F. P. J. Am. Chem. Soc. 2000, 122,
̈
9054−9055. (b) Hoefelmeyer, J. D.; Gabbaı,
F. P. Organometallics
2002, 21, 982−985. (c) Hoefelmeyer, J. D.; Schulte, M.; Tschinkl, M.;
̈
Gabbaı,
F. P. Coord. Chem. Rev. 2002, 235, 93−103. (d) Hoefelmeyer,
̈
J. D.; Sole,
́
S.; Gabbaı, F. P. Dalton Trans. 2004, 1254−1258.
2334
dx.doi.org/10.1021/om2012216 | Organometallics 2012, 31, 2327−2335

Organometallics
Article
̈
́
(8) (a) Sole, S.; Gabbaı, F. P. Chem. Commun. 2004, 1284−1285.
̈
̈
(b) Melaımi, M.; Sole, S.; Chiu, C.-W.; Wang, H.; Gabbaı, F. P. Inorg.
Chem. 2006, 45, 8136−8143.
(9) (a) Agou, T.; Kobayashi, J.; Kawashima, T. Chem.Eur. J. 2007,
̈
13, 8051−8060. (b) Matsumoto, T.; Wade, C. R.; Gabbaı, F. P.
Organometallics 2010, 29, 5490−5495. (c) Agou, T.; Kobayashi, J.;
Kawashima, T. Phosphorus, Sulfur Silicon Relat. Elem. 2010, 185, 947−
951. (d) Rohr, A. D.; Banaszak, H. M. M.; Kampf, J. W.; Ashe, A. J.
Organometallics 2011, 30, 3698−3700.
(10) (a) Agou, T.; Kobayashi, J.; Kawashima, T. Inorg. Chem. 2006,
45, 9137−9144. (b) Lee, M. H.; Agou, T.; Kobayashi, J.; Kawashima,
̈
T.; Gabbaı, F. P. Chem. Commun. 2007, 1133−1135. (c) Agou, T.;
̈
Kobayashi, J.; Kim, Y.; Gabbaı, F. P.; Kawashima, T. Chem. Lett. 2007,
36, 976−977.
(11) Yamaguchi, S.; Akiyama, S.; Tamao, K. J. Am. Chem. Soc. 2001,
123, 11372−11375.
(12) Dorsey, C. L.; Jewula, P.; Hudnall, T. W.; Hoefelmeyer, J. D.;
̈
Taylor, T. J.; Honesty, N. R.; Chiu, C.-W.; Schulte, M.; Gabbaı, F. P.
Dalton Trans. 2008, 4442−4450.
(13) Bader, R. F. W. Atoms in Molecules: A Quantum Theory; Oxford
University Press: UK, 1994.
̈
(14) Kim, Y.; Kim, M.; Gabbaı, F. P. Org. Lett. 2010, 12, 600−602.
̈
(15) Schulte, M.; Gabbaı, F. P. Chem.Eur. J. 2002, 8, 3802−3807.
2335
dx.doi.org/10.1021/om2012216 | Organometallics 2012, 31, 2327−2335