Effect of para-substituents and solvent polarity on the formation of triphenylboroxine·amine adducts

Article abstract of DOI:10.1021/jp062055e

Density functional theory (B3LYP//6-311+G*) calculations including Poisson-Boltzmann implicit solvent and NMR were used to study the formation of a series of para-substituted triphenylboroxine·amine adducts with respect to their phenylboronic acid monomers and free amine in solution. Our calculations suggest that the intermediate prior to forming trimer·amine is a dimeramine adduct. Formation of dimeramine can proceed via two pathways. Electron-donating substituents favor dimerization of two monomers before addition of the amine, and electron-withdrawing substituents favor formation of a monomer·amine adduct before addition of the second monomer. We also find that π-electron acceptors destabilize formation of the dimer and trimer with respect to its monomers. Electron-withdrawing substituents favor adduct formation. Adduct formation is enthalpically stabilized by increasing the polarity of the solvent but differential solubility of the monomer compared to trimer·amine also has an effect on the equilibrium constant.

Full text of DOI:10.1021/jp062055e

8158
J. Phys. Chem. A 2006, 110, 8158-8166
Effect of Para-Substituents and Solvent Polarity on the Formation of
Triphenylboroxine‚Amine Adducts
Jeremy Kua,* Matthew N. Fletcher, and Peter M. Iovine
Department of Chemistry, UniVersity of San Diego, 5998 Alcala Park, San Diego, California 92110
ReceiVed: April 3, 2006; In Final Form: May 15, 2006
Density functional theory (B3LYP//6-311+G*) calculations including Poisson-Boltzmann implicit solvent
and NMR were used to study the formation of a series of para-substituted triphenylboroxine‚amine adducts
with respect to their phenylboronic acid monomers and free amine in solution. Our calculations suggest that
the intermediate prior to forming trimer‚amine is a dimer‚amine adduct. Formation of dimer‚amine can proceed
via two pathways. Electron-donating substituents favor dimerization of two monomers before addition of the
amine, and electron-withdrawing substituents favor formation of a monomer‚amine adduct before addition of
the second monomer. We also find that π-electron acceptors destabilize formation of the dimer and trimer
with respect to its monomers. Electron-withdrawing substituents favor adduct formation. Adduct formation
is enthalpically stabilized by increasing the polarity of the solvent but differential solubility of the monomer
compared to trimer‚amine also has an effect on the equilibrium constant.
Introduction
construction from monomeric boronic acids (step 1, Figure 1)
and subsequent boroxine complexation by amines to form a 1:1
adduct (step 2, Figure 1). In that preliminary study we chose to
use water as an implicit (but nonreactive) solvent because the
implicit solvent parameters were well tested in the Jaguar23
program. We found that the trimerization of phenylboronic acids
to form arylboroxine rings (step 1) is enthalpically unfavorable.
In contrast, the formation of stable 1:1 adducts (step 2) was
highly favorable; in fact, sufficiently favorable to drive the two-
step reaction forward toward formation of the products.
Substitution of π-electron-withdrawing groups in the para-
position of the phenyl ring further disfavored step 1, whereas
the opposite was observed for π-electron-donors. On the other
hand, substituents that were oVerall electron-withdrawing
favored step 2, whereas electron donors disfavored it. We also
found that formation of 1:2 adducts of arylboroxine‚amine were
less favorable enthalpically compared to 1:1 adducts.
We have extended the previous study to examine the ther-
modynamics of intermediate steps and identify potential stable
intermediates using both computational methods and NMR. One
possible intermediate in step 1 could be formation of a dimer
from two boronic acid monomers. Furthermore, in the presence
of ligand, it may be possible to form monomer‚amine and dimer‚
amine, prior to the formation of trimer‚amine. We seek to
answer the following questions: (1) Can stable adducts be
formed between the base and the monomer or dimer? (2) What
is the lowest energy pathway to form the stable trimer‚amine?
(3) How do electron-acceptor and electron-donor substituents
influence which reaction pathway will be most favorable
enthalpically? (4) How does solvent polarity affect the thermo-
dynamics of this system?
Boroxines, the dehydration product of organoboronic acids,1
have found commercial use in such diverse areas as flame
retardant materials,2 dopants for lithium ion transference in
polymer electrolytes,3-5 acid alternatives in Suzuki-Miyaura
coupling reactions,6 and nonlinear optical materials.7 Boroxines
are known to form stable adducts with many nitrogen donor
compounds including amines,8-12 pyridines,13 hydrazines,14
azaindoles,15 and even salen type ligands.16 Low-temperature
NMR studies on trimethylboroxine‚ligand adducts indicate that
the ligands are in fast exchange with activation barriers of 9-13
kcal/mol.15,17
There are few studies investigating the thermodynamics of
boroxine formation. Tokunaga et al. measured the equilibrium
constant between phenylboronic acid and triphenylboroxine for
a series of p-phenyl substituents.18 This equilibrium corresponds
to step 1 in Figure 1. There have been two computational studies,
one by us (summarized below), and the other by Beckmann et
al. in which a ring strain argument is used to explain the stability
of boroxine‚ligand adducts.19
As arylboroxines find increased utility in material science
applications, such as covalent organic frameworks20 and nano-
scale molecular scaffolds,21 a firm understanding of fundamental
arylboroxine solution chemistry, including the rich ligand
chemistry, would facilitate progress. Although the ligand
facilitated trimerization of arylboronic acids has been qualita-
tively reported, there have been no quantitative evaluations of
this two-step process. As pointed out earlier, Tokunaga et al.
investigated the thermodynamics of arylboroxine construction
from monomeric arylboronic acids but did not address adduct
formation. This computational and NMR-based study fills a gap
in the literature of arylboroxines by systematically and quan-
titatively investigating the role of ligation in arylboroxine ring
construction.
To complement our NMR studies, we have chosen to use
three implicit solvents of varying polarity in our calculations:
acetone, dichloromethane and chloroform. A selection of
different p-phenyl substituents are used in the calculations (X
) H, CH3, OCH3, F, Cl, C(O)OCH3, C(O)CH3, CHO, CN, CF3).
The six calculated compounds (monomer, dimer, trimer, and
their 1:1 adducts) for each substituent are shown in Figure 2.
In our previous computational study,22 we examined each
individual step in the two-step reaction sequence: boroxine
* Corresponding author. E-mail: jkua@sandiego.edu.
10.1021/jp062055e CCC: $33.50 © 2006 American Chemical Society
Published on Web 06/15/2006

Formation of Triphenylboroxine‚Amine Adducts
J. Phys. Chem. A, Vol. 110, No. 26, 2006 8159
Figure 1. Two-step reaction sequence of boroxine construction followed by adduct formation.
Figure 2. Compounds in this computational study.
In this study we have considered only 1:1 adducts because
formation of 1:2 adducts were found to be much less favorable
in our previous study.22 For our calculations we have used the
computationally less expensive NH3 as the amine base. The
NMR experiments use pyridine. A comparison between NH3
and pyridine with respect to the thermodynamics for the
formation of 1:1 adducts with arylboroxines can also be found
in our previous study. In short, we found that trends were the
same for both ligands, except that pyridine has a weaker binding
energy to arylboroxine than NH3.
Charges are allowed to develop on the surface according to the
electrostatic potential of the solute and ꢀ; then the polarized
reaction field of the solvent acts back on the quantum mechan-
ical description of the solute. The wave function of the complex
is relaxed self-consistently with the reaction field to solve the
PB equations. Although the forces on the quantum mechanical
solute atoms due to the solvent can be calculated in the presence
of the solvent, in this work, the solvation energy was calculated
at the optimized gas-phase geometry. This is because there is
little change between the gas-phase and implicit solvent
optimized geometries. The difference in energy between the
Materials and Methods
unsolvated and solvated structures is designated Esolv
.
The analytical Hessian was calculated for each optimized
geometry in the gas phase. The DFT gas-phase energy was then
corrected for zero-point vibrations. The temperature-dependent
enthalpy correction term is straightforward to calculate from
statistical mechanics. Assuming that the translational and
rotational corrections are a constant times kT, that low frequency
vibrational modes will generally cancel out when calculating
enthalpy differences, and that the vibrational frequencies do not
change appreciably in solution, we can calculate H298K. The sum
of the zero point energy and enthalpy corrections to 298 K are
collectively designated Ecorr. The calculated values of Eelec, Esolv
and Ecorr are available in the Supporting Information. The
corresponding free-energy corrections in solution are much less
reliable.30-32 Changes in free-energy terms for translation and
rotation are poorly defined in solution, particularly as the size
Computational Methods. All calculations were carried out
using Jaguar 5.523 at the B3LYP24-27 flavor of density functional
theory with a 6-311+G* basis set. We chose to run our
calculations at a similar level of theory and basis set to
complement our previous study.22 The electronic energy of the
optimized gas-phase structures is designated Eelec. The Poisson-
Boltzmann (PB) continuum approximation28,29 was used to
describe the effect of solvent. In this approximation, a smooth
solvent-accessible surface of the solute is calculated by rolling
a sphere of radius Rsolv over the van der Waals surface. The
solvent is represented as a polarizable continuum surrounding
the molecule with dielectric constant ꢀ. The parameters used
for the dielectric constant and probe radius are ꢀ ) 20.7 and
Rsolv ) 2.43 Å for acetone, ꢀ ) 10.0 and Rsolv ) 2.33 Å for
dichloromethane, ꢀ ) 4.8 and Rsolv ) 2.50 Å for chloroform.

8160 J. Phys. Chem. A, Vol. 110, No. 26, 2006
Kua et al.
Figure 3. Dimerization and trimerization reactions of phenylboronic acid.
of the molecule increases. Additional corrections to the free
energy for concentration differentials among species (to obtain
the chemical potential) can be significant, especially if the
solubility varies among the different species in solution.
Furthermore, because the reactions being studied are in solution,
the free energy being accounted for comes from two different
sources: thermal corrections and implicit solvent. Neither of
these parameters is easily separable, nor do they constitute all
the required parts of the free energy under our approximations
of the system.
Our reported reaction energies are as follows: (1) ∆Eelec is
the difference in electronic energy between reactants and
products in a given reaction. (2) ∆Hgas is calculated by adding
zero point energy and thermal corrections to 298 K to the
electronic energies. (3) ∆Hsoln is calculated by adding the free
energy due to solvation (for each solvent) to ∆Hgas. It is
important to note that even though the solvation energy
contribution is to some extent a free-energy correction, it
certainly does not account for all of the free energy. Hence, we
will retain the symbol ∆H and refer to this quantity as the
solution phase enthalpy in our results and discussion.
Materials and Instrumentation. NMR data were acquired
on a 400 MHz Varian Mercury system. All para-substituted
phenylboronic acids were obtained from Frontier Scientific and
used as received. The acetone-d6 was purchased from Cambridge
Isotope Labs in single-use ampules to minimize hydration.
Pyridine was purchased from Acros Organics and was stored
over 4 Å molecular sieves. Solvent chemical shifts for 1H spec-
tra are referenced to the deuterated solvent used in the
experiment.
overall equilibrium constants for the two-step sequence outlined
in Figure 1 were calculated using the following expression:
[H2O]3[borox‚pyr]
K )
[BA]3[pyr]
where [pyr] is the concentration of free pyridine, [borox‚pyr]
is the concentration of arylboroxine‚pyridine, and [BA] is the
concentration of the boronic acid monomer.
The downfield doublets associated with the boronic acid
monomer and arylboroxine aromatic protons were well separated
and therefore integrated separately. Because pyridine is in fast
exchange on the NMR time scale, the aromatic protons cor-
responding to the arylboroxine species and the pyridine reso-
nances are a weighted average of the bound and free species.
However, under conditions where the ratio of pyridine/boronic
acid was varied from 0.33:1 to 2:1, respectively, all arylboroxine
was found to be bound. This observation was independently
confirmed by monitoring the chemical shift of the arylboroxine/
arylboroxine‚pyridine’s downfield aromatic doublet as a function
of pyridine added. In the 0.33:1 to 2:1 pyridine/boronic acid
range, the chemical shift of the downfield doublet did not shift,
indicating that all arylboroxine present in solution was bound
by pyridine. The integration of the arylboroxine‚pyridine was
then used to determine the concentration of bound versus
unbound pyridine. In several cases, one of pyridine’s three
aromatic resonances overlapped with the downfield doublet of
the boronic acid or arylboroxine‚pryidine. In these instances,
the two remaining pyridine resonances were used to deconvolute
the overlap and isolate the portion of the integration originating
from the boronic acid or arylboroxine‚pyridine species.
General Method for NMR Titration Measurements. Each
para-substituted phenylboronic acid was added to a NMR tube
as to give approximately 0.02 mmol. To the solid was added
acetone-d6 (0.5 mL), and a baseline spectrum was taken prior
to the addition of the ligand. Pyridine (0.5 µL) was then added,
Results and Discussion
Formation of Dimers and Trimers from Boronic Acid.
The dimerization and trimerization reactions starting from
boronic acid monomer 1 are shown in Figure 3. The dimer and
arylboroxine trimer are labeled 2 and 3, respectively. Calculated
energies for the dimerization and trimerization reactions are
compiled in Tables 1 and 2. The substituents have been sorted
with respect to their Hammett parameters with the exception
of CN and CF3. We swapped the position of CN and CF3 to
group CN with the other π-acceptors. A graphical view of ∆H
for the dimerization and trimerization reactions is represented
by the left two columns (labeled 2x and 3x, respectively) in
Figure 4.
1
and the tube shaken and rested for 10 min before a second H
NMR spectrum was acquired. The pyridine addition was
repeated five times (total pyridine added was 3 µL), resulting
in six 1H NMR spectra for each para-substituted phenylboronic
acid. The pyridine/boronic acid molar ratios were in the range
of 0.33:1 to 2:1, respectively.
Assignment of the aromatic protons in 4-methoxyphenylbo-
ronic acid was made via a ROESY experiment.
The NMR spectra obtained after pyridine addition were inte-
grated to determine concentrations of water, pyridine, mono-
meric boronic acid, arylboroxine, and pyridine‚arylboroxine. The

Formation of Triphenylboroxine‚Amine Adducts
J. Phys. Chem. A, Vol. 110, No. 26, 2006 8161
TABLE 1: Energetics (kcal/mol) of Dimer Formation
Formation of 1:1 Adducts. Because trigonal planar boron
can act as a Lewis acid, the addition of a Lewis base such as
NH3 may lead to the formation of 1:1 adducts. In the adduct, a
new B-N bond is formed and boron adopts a tetrahedral
environment. The base donates electron density to the tetrahedral
boron, which in the Lewis structure carries a formal negative
charge (nitrogen has the formal positive charge). Calculated
energies for the addition of one equivalent of NH3 to monomer,
dimer and trimer to form 1:1 adducts are compiled in Tables
3-5. The adducts are labeled 1‚NH3, 2‚NH3 and 3‚NH3 for the
monomer, dimer and trimer, respectively. A graphical view of
∆Hsoln for 1:1 adduct formation is represented by the middle
three columns (labeled 1+NH3, 2+NH3, 3+NH3, respectively)
in Figure 4.
X
∆Eelec ∆Hgas ∆Hsoln(acetone) ∆Hsoln(CH2Cl2) ∆Hsoln(CHCl3)
OCH3
CH3
H
F
Cl
2.54 2.20
2.77 1.22
2.76 1.86
2.57 1.64
2.65 1.16
2.93
2.04
2.36
2.01
1.54
3.74
3.73
3.64
3.39
3.70
2.57
1.95
2.38
2.06
1.56
3.70
3.73
3.61
3.36
3.80
2.29
1.83
2.19
1.83
1.18
3.65
3.42
3.36
3.29
3.70
C(O)OCH3 3.09 3.35
C(O)CH3 3.08 3.34
CHO
CN
3.07 3.32
3.09 3.35
3.18 3.44
CF3
TABLE 2: Energetics of Trimer Formation
X
∆Eelec ∆Hgas ∆Hsoln(acetone) ∆Hsoln(CH2Cl2) ∆Hsoln(CHCl3)
OCH3
CH3
H
F
Cl
12.45 9.69
13.11 8.56
13.21 10.47
12.58 9.82
12.91 9.01
2.82
2.23
3.65
3.28
2.49
6.76
6.66
6.78
6.47
4.76
3.28
2.69
4.36
3.91
3.09
7.39
7.39
7.40
7.15
5.53
4.08
3.57
5.06
4.63
3.76
8.31
8.00
8.15
8.07
6.35
Three clear trends can be observed. The first trend is that the
change in enthalpy is most favorable for the addition of NH3 to
the trimer and least favorable for adding NH3 to the monomer.
This comes as no surprise because the trimer has the greatest
Lewis acidity and the monomer the least. ∆H is positive for
the formation of 1‚NH3, but for the more acidic dimer and
trimer, ∆H is negative, indicating that forming the 1:1 adduct
in solution is enthalpically favorable for dimers and trimers.
Note that ∆S is negative for adduct formation; thus, -T∆S will
be positive and the entropic contribution disfavors adduct
formation.
C(O)OCH3 14.46 13.44
C(O)CH3 14.67 13.67
CHO
CN
14.76 13.75
14.75 13.75
14.58 11.80
CF3
All ∆H values are positive for both the dimerization and
trimerization reactions. There is an average 88% increase
(slightly less than double) in ∆H comparing the trimerization
to the dimerization reaction for all substituents. This increase
is most dramatic for π-accepting substituents in chloroform, the
least polar of the solvents in our study. We find that:
(a) Substituents that are overall electron-donating, and
π-donors in particular, result in ∆H values that are less positive
than those for the unsubstituted case (X ) H).
(b) The halogens (X ) F, Cl), although overall electron-
withdrawing, are good π-donors and have ∆H values less
positive than X ) H.
(c) Electron-withdrawing substituents, π-acceptors in par-
ticular, have ∆H values more positive than X ) H.
The second trend is that the exothermicity of adduct formation
increases according to overall electron-withdrawing capability
of the para-substituent. This holds true in all three solvents. This
is not surprising because electron-donating groups are expected
to destabilize the buildup of negative charge on the boron due
to dative covalent bonding from the lone pair of ammonia.
Electron-withdrawing groups, on the other hand, act to stabilize
the adduct. The distinction between σ and π effects is not
important once the adduct (with four-coordinate tetrahedral
boron) is formed from a thermodynamic standpoint (although
it may have a kinetic effect). Hence, we see that for adduct
formation, the halogens act primarily in their electron-withdraw-
ing capacity, stabilizing the build-up of negative charge on
boron. Both F and Cl have ∆H values more exothermic than
H. Because we have arranged the substituents according to their
Hammett parameters (with the exception of CF3 and CN), the
middle three columns in Figure 4 highlight the trend toward
increasing exothermicity with electron-withdrawing ability.
(d) Comparing the different solvents we find that as the
polarity of the solvent decreases, the change in solvation
enthalpy becomes increasingly endothermic.
The dimerization reaction has equal numbers of reactants and
products; thus we expect that the contribution of ∆S will be
small (but not necessarily negligible given that the ∆H values
are also relative small). The trimerization reaction releases three
water molecules, so we expect ∆S to be positive and -T∆S to
be negative. Although we have not calculated free energies in
solution, we expect ∆G values to still be positive, but smaller
in magnitude than ∆H. The equilibrium constant measured via
NMR spectroscopy (in CDCl3) by Tokunaga et al.18 suggest
that the trimerization reaction is indeed thermodynamically
unfavorable. Similar results were noted in our NMR experi-
ments; however, our two-step solution measurements were
performed in acetone instead of chloroform. Solvent appears
to play an important role in the equilibrium, and therefore no
direct comparison can be made between Tokunaga’s NMR
results and those presented here. This will be discussed in the
next section.
In summary, we find that substitution of π-electron-
withdrawing groups in the para-position of the phenyl ring
further destabilize the dimer and trimer with respect to its
monomers, whereas the opposite is observed from π-electron
donors. We have previously compared and found good relative
agreement between a subset of our calculated ∆Eelec values with
-RT ln Keq from experiment.22
The third trend is that, as the polarity of the solvent decreases,
∆Hsoln is less exothermic for forming 2‚NH3 and 3‚NH3 and
more endothermic for forming 1‚NH3. Because addition of NH3
leads to a more polar compound (with formal B- and N+
charges), we would expect greater stabilization of the adduct
with more polar solvents. This is indeed what we find in our
calculations; however, this is not what is observed experimen-
tally where more polar solvents (e.g., acetone) disfavor adduct
formation relative to less polar solvents (e.g., chloroform). The
1:1 adduct with pyridine (experiment) is also less polar than
with NH3 (calculation).
Figure 5 shows the equilibrium distribution of 4-methoxy-
phenylboronic acid in acetone and chloroform. In both experi-
ments one-third of an equivalent of pyridine (relative to the total
molar quantity of boronic acid) was added to a solution of
4-methoxyphenylboronic acid. Because pyridine is known to
be in fast exchange on the NMR time scale, the downfield
doublet assigned to the arylboroxine (∼7.8 ppm in acetone-d6
and ∼8.0 ppm in CDCl3) actually represents a weighted average

8162 J. Phys. Chem. A, Vol. 110, No. 26, 2006
Kua et al.
Figure 4. Calculated ∆Hsoln (kcal/mol) for all reactions studied.
of ligated and unligated arylboroxine. Under the conditions of
this experiment, arylboroxine was found to be fully bound by
pyridine (see NMR Methods section). In acetone, a 4:1 ratio of
boronic acid:arylboroxine‚pyridine is observed. In chloroform,
the major species is not the boronic acid but rather the
arylboroxine‚pyridine with an observed ratio of 1:2 boronic acid:
arylboroxine‚pyridine.
observations. The first issue relates to the inherent solubility of
water in the two solvents and the ability of this dissolved water
to participate in the ring-opening reaction. Practically speaking,
acetone is much more difficult to obtain and maintain anhydrous
and at low concentrations of substrate (∼30 mM) trace water
(as well as the extruded water from the condensation reaction)
may be significant.
There are two factors that may play a role in causing the
discrepancy between the computational results and experimental
The second factor relates to the solubility of the starting
boronic acid. At the concentrations used in this study, the

Formation of Triphenylboroxine‚Amine Adducts
J. Phys. Chem. A, Vol. 110, No. 26, 2006 8163
TABLE 3: Energetics (kcal/mol) of Formation of 1‚NH3
X
∆Eelec ∆Hgas ∆Hsoln(acetone) ∆Hsoln(CH2Cl2) ∆Hsoln(CHCl3)
OCH3
CH3
H
F
Cl
0.58 2.37
0.14 1.94
-0.32 1.51
-0.67 1.18
-1.00 0.87
3.88
3.17
2.60
2.06
1.78
1.93
1.75
1.35
0.64
0.72
3.51
2.89
2.36
1.78
1.43
1.54
1.61
0.97
0.37
0.46
3.24
2.58
2.11
1.58
1.26
1.45
1.20
0.77
0.25
0.30
C(O)OCH3 -1.27 1.17
C(O)CH3 -1.42 1.05
CHO
CN
-1.94 0.55
-2.32 0.20
-1.96 -0.08
CF3
TABLE 4: Energetics (kcal/mol) of Formation of 2‚NH3
X
∆Eelec ∆Hgas ∆Hsoln(acetone) ∆Hsoln(CH2Cl2) ∆Hsoln(CHCl3)
OCH3
CH3
H
F
Cl
-0.83 0.76
-1.98 0.27
-2.59 -0.36
-2.91 -0.64
-3.41 -1.10
-0.89
-1.38
-1.70
-2.17
-2.72
-3.96
-4.00
-4.55
-4.77
-4.99
-0.76
-1.25
-1.84
-2.13
-2.69
-3.82
-3.88
-4.54
-4.77
-5.16
-0.20
-1.06
-1.44
-1.69
-2.28
-3.39
-3.48
-4.10
-4.42
-4.83
C(O)OCH3 -4.05 -1.77
C(O)CH3 -4.04 -1.75
CHO
CN
CF3
Figure 5. 1H NMR spectra of 4-methoxyphenylboronic acid in the
presence of pyridine comparing the equilibrium distribution of species
as a function of solvent.
-4.83 -2.50
-5.40 -3.06
-5.18 -3.46
TABLE 5: Energetics (kcal/mol) of Formation of 3‚NH3
its monomers; they have the most endothermic ∆H values for
forming 3 (Table 2, second column in Figure 4). The halogens,
on the other hand, as π-donors, are less endothermic toward
trimer formation, and their overall electron-withdrawing capabil-
ity (though weaker than that of the π-acceptors in this study)
leads to significant stabilization of 3‚NH3.
X
∆Eelec ∆Hgas ∆Hsoln(acetone) ∆Hsoln(CH2Cl2) ∆Hsoln(CHCl3)
OCH3
CH3
H
F
Cl
-4.46 -2.36
-5.60 -3.38
-6.45 -4.25
-6.93 -4.72
-7.51 -5.22
-3.68
-4.52
-6.58
-7.29
-8.27
-9.51
-9.68
-10.31
-10.97
-10.98
-3.70
-4.33
-6.47
-7.12
-8.08
-9.31
-9.40
-10.24
-10.90
-10.85
-3.45
-4.09
-6.07
-6.80
-7.61
-8.78
-8.99
-9.95
-10.44
-10.44
C(O)OCH3 -8.45 -6.20
C(O)CH3 -8.63 -6.34
CHO
CN
CF3
The overall two-step equilibrium (Figure 1) associated with
1
arylboroxine formation and ligation was also examined by H
-9.53 -7.20
-10.10 -7.73
-9.78 -7.99
NMR spectroscopy. To our knowledge this is the first experi-
mental study to quantify the thermodynamics of the two-step
equilibrium process.
Figure 7 shows a representative titration experiment for
4-methoxyphenylboronic acid in acetone. The top spectrum
(Figure 7, top panel), in the absence of the pyridine ligand,
shows a mixture of 4-methoxyphenylboronic acid (labeled with
circles) and 4-methoxyphenylboroxine (labeled with triangles).
In the absence of pyridine the monomeric boronic acid
dominates. Of particular interest are the downfield sets of
aromatic protons (8.2-7.5 ppm) as the chemical shifts associated
with these protons are diagnostic of the structure (i.e.,
arylboroxine‚pyridine, arylboroxine, boronic acid). The boronic
acid -OH resonance appears as a sharp singlet at approximately
6.8 ppm in the top spectrum.
The middle spectrum results from the addition of 1 equiv of
pyridine relative to 3 molar equiv of monomeric boronic acid.
The presence of the ligand drives the formation of the aryl-
boroxine ring and the downfield doublet associated with the
arylboroxine moves upfield. Using this experimental approach,
equilibrium values and the corresponding ∆G values were
determined for a series of para-substituted compounds. The
results are presented in Table 7.
The free energy trends in the NMR data are in general
agreement with the trends observed in the computational portion
of the study. The most electron-donating substituent, OCH3, has
the least negative free energy change. H and CH3 are very close.
The electron-withdrawing substituents have a more negative free
energy change than X ) H. Note that the computational results
are in terms of enthalpy changes rather than free energy changes;
entropic and solubility effects were not included. Therefore, it
is not unexpected that there are some differences in the trend.
In the computational study, X ) CH3 (being mildly electron
4-substituted-phenylboronic acid compounds are sparingly
soluble in chloroform whereas in acetone, homogeneous solu-
tions are formed. The addition of pyridine generally clarifies
the chloroform solutions; i.e., arylboroxine‚pyridine is soluble
in chloroform. We think that the difference in solubility disfavors
the reverse reaction (hydrolysis of arylboroxine back to mono-
mer) because of an additional entropic cost due to phase
separation of the sparingly soluble monomer from the solvent.
Our calculations do not take this into account.
In all cases (except one: X ) OCH3 in chloroform), the
exothermicity of ∆H for adduct formation is larger in magnitude
than the endothermicity of ∆H for trimerization; i.e., the
presence of NH3 favors stabilization of the arylboroxine ring
via adduct formation. In the absence of NH3, formation of the
arylboroxine ring is disfavored by an increasingly endothermic
∆H. This is also the trend observed in the experimental data.
Only after adding a suitable ligand, in our case pyridine, is the
equilibrium shifted substantially toward the arylboroxine prod-
ucts. The net result from adding these two ∆H values is
represented graphically by the rightmost column in Figure 4
(labeled 3x+NH3). This net ∆H value is obtained by adding
∆H values from the second (labeled 3x) and fifth (labeled
3+NH3) columns. The overall net reaction is shown in Figure
6 and the net ∆H values are compiled in Table 6.
The most exothermic net ∆H values come from the CF3 and
Cl substituents. CF3 is very strongly electron withdrawing,
thus stabilizing the build-up of negative charge on boron in
the adduct. Although the π-electron-withdrawing substituents
(C(O)OCH3, C(O)CH3, CHO, CN) also strongly stabilize adduct
formation, they decrease the stability of the trimer relative to

8164 J. Phys. Chem. A, Vol. 110, No. 26, 2006
Kua et al.
Figure 6. Net reaction for forming 3‚NH3 when monomers and NH3 are present.
with respect to its monomers (favorable for π-donors, unfavor-
able for π-acceptors) and stabilization of 3‚NH3 (favorable for
overall electron-withdrawing substituents).
Relative Importance of 2‚NH3 as an Intermediate. How
does the reaction proceed from monomers and free NH3 in
solution to formation of 3‚NH3, the thermodynamic sink in this
system? Our results suggest that forming the trimer in the
absence of base is highly unfavorable, as 3 is the highest energy
stable intermediate among all possible species. Instead, our
calculations point to 2‚NH3 as the most likely intermediate prior
to formation of 3‚NH3. The formation of 2‚NH3 from monomers
and free NH3 in solution is shown in Figure 8. The net ∆H
values for this reaction are represented by the second rightmost
column in Figure 4 (labeled 2x+NH3) and are the sum of ∆H
values from Tables 1 and 4 (or the first and fourth columns of
Figure 4 labeled 2x and 2+NH3, respectively). These net values
are shown in Table 8.
Figure 7. 1H NMR stack plot in acetone-d6 showing the equilibrium
distribution of species for 4-methoxyphenylboronic acid as a function
of pyridine added.
The net ∆H values are slightly endothermic for electron-
donating substituents and X ) H. For electron-withdrawing
substituents, these values are marginally exothermic with two
exceptions (X ) F, C(O)OCH3 in chloroform). Note that ∆S is
negative in this reaction; thus, -T∆S will be positive and the
entropic contribution is likely to lead to positive ∆G values for
all substituents except those with the most exothermic ∆H values
(X ) CF3, CN, Cl). Thus the dimer adduct is unlikely to be
observed unless the substituent is strongly electron withdrawing
(and preferably through the σ-framework). Preliminary 1H and
19F NMR data obtained for p-(trifluoromethyl)phenylboronic
acid may support this hypothesis (see Supporting Information
where an additional set of signals is tentatively assigned to the
dimer adduct).
TABLE 6: Net ∆H (kcal/mol) for Forming 3‚NH3 When
Monomers and NH3 Are Present
X
∆Hsoln(acetone)
∆Hsoln(CH2Cl2)
∆Hsoln(CHCl3)
OMe
CH3
H
F
-0.86
-2.29
-2.93
-4.01
-5.78
-2.75
-3.02
-3.53
-4.50
-6.22
-0.42
-1.64
-2.11
-3.21
-4.99
-1.92
-2.01
-2.84
-3.75
-5.32
0.63
-0.52
-1.01
-2.17
-3.85
-0.47
-0.99
-1.80
-2.37
-4.09
Cl
C(O)OCH3
C(O)CH3
CHO
CN
CF3
There are two possible ways to get to 2‚NH3 from monomers
and free NH3: (1) dimerization of the monomers first, followed
by addition of NH3 to the dimer; (2) addition of NH3 to the
monomer first, followed by addition of the second monomer.
These two routes are shown in Figure 9. To compare the
favorability of these two routes, we can compare the relative
∆H values for the first step in each route, i.e., comparing the
results in Table 1 to those in Table 3, or the first column in
Figure 4 (labeled 2x) with the third column (labeled 1+NH3).
Both of these first steps are endothermic.
TABLE 7: NMR-Based Thermodynamic Data Acquired in
acetone-d6 for the Two-Step Equilibrium Process Shown in
Figure 1
net equilibrium constant
(step 1 and 2 from Figure 1)
∆G
(kcal/mol)
X
OCH3
H
CH3
C(O)CH3
CF3
CN
1.4
2.8
3.3
10.0
11.0
16.0
-0.2
-0.6
-0.7
-1.3
-1.4
-1.7
We find that ∆H is less endothermic for dimerization as the
first step for electron-donating substituents and X ) H. On the
other hand, the electron-withdrawing substituents have a less
endothermic ∆H compared to X ) H for adding NH3 to the
monomer as the first step. This suggests that the choice of
substituent can play a role in which pathway is taken: electron
donors favor the top pathway in Figure 9 and electron acceptors
prefer the bottom pathway.
donating) was less exothermic than X ) H, although the
differences in experimental ∆G and computational ∆H compar-
ing both substituents were both very small. The other difference
is that X ) CF3 was found to be more exothermic than X )
CN computationally.
In summary, overall electron-withdrawing ability is important
in stabilizing formation of 3‚NH3. Formation of this adduct was
found to be overall exothermic in almost all cases. The value
of ∆H represents a balance between the relative energy of 3
The argument above only considered the enthalpic contribu-
tion of the first step of the reaction. The entropic contribution

Formation of Triphenylboroxine‚Amine Adducts
J. Phys. Chem. A, Vol. 110, No. 26, 2006 8165
Figure 8. Net reaction for forming 2‚NH3 from monomers and free NH3.
Figure 9. Routes to formation of 2‚NH3 from monomers and free NH3.
TABLE 8: Net ∆H (kcal/mol) for Forming 2‚NH3 When
independent confirmation that electron-withdrawing substituents
do drive the two-step equilibrium toward adduct formation, in
agreement with our computational results.
Monomers and NH3 Are Present
X
∆Hsoln(acetone)
∆Hsoln(CH2Cl2)
∆Hsoln(CHCl3)
We also find, from our calculations of ∆H, that the most
important intermediate prior to forming 3‚NH3 is 2‚NH3.
Formation of 2‚NH3 can proceed via two pathways. Electron-
donating substituents favor dimerization of two monomers
before addition of NH3, and electron-withdrawing substituents
favor formation of a 1‚NH3 before addition of the second
monomer. Entropic effects may possibly change these results.
As the solvent increases in polarity from chloroform to
dichloromethane to acetone, formation of trimer in the absence
of NH3 decreases in endothermicity, whereas formation of the
more polar adducts in the presence of NH3 is further stabilized.
As entropic effects and phase separation are not included in
our calculations, it is not surprising that our calculated trend of
∆H is different from that observed by NMR.
OCH3
CH3
H
F
Cl
C(O)OCH3
C(O)CH3
CHO
CN
2.04
0.66
0.66
-0.16
-1.18
-0.22
-0.27
-0.91
-1.38
-1.29
1.81
0.70
0.54
-0.07
-1.13
-0.12
-0.15
-0.93
-1.41
-1.36
2.09
0.77
0.75
0.14
-1.10
0.26
-0.06
-0.74
-1.13
-1.13
CF3
is expected to further destabilize formation of the monomer 1:1
adduct (bottom pathway) and have less of an effect on
dimerization; the latter having equal numbers of moles of
reactants and products. The activation barrier in each case may
also have a significant effect on the pathway taken. We are
currently working on optimizing all the low-lying transition
states and calculating the barriers in this system. Preliminary
estimates suggest that the barriers for each step in both pathways
are low (in the range 6-14 kcal/mol), with addition of base to
form an adduct having lower barriers than coupling of monomers
to form dimer. NMR data are supportive of low barriers as (1)
the addition of base proceeds completely to the thermodynamic
sink, therefore suggesting equlilibration at room temperature,
and (2) pyridine is in fast exchange on the NMR time scale.
These points mesh well with the low dissociation/association
activation barrier (approximately 10 kcal/mol) reported for the
phenylboroxine•pyridine system.15
We are in the process of calculating activation energy barriers
between these intermediates, and determining the geometric
structure of the relevant transition states.
Acknowledgment. This research was supported by a Camille
and Henry Dreyfus Foundation Start-up Award (J.K.), University
of San Diego startup funds (J.K.), awards from Research
Corporation (P.M.I., J.K.) and the American Chemical Society
Petroleum Research Fund (40382-GB4) (P.M.I.). NSF MRI
CHE-0417731 (P.M.I) is also acknowledged.
Supporting Information Available: Calculated values of
1
E
elec, Esolv and Ecorr, and preliminary H and 19F NMR data
supporting the existence of a pyridine‚boronic acid dimer adduct.
This material is available free of charge via the Internet at http://
pubs.acs.org.
Conclusions
From our DFT calculations we find that in the absence of
base, formation of dimers and trimers is thermodynamically
unfavorable. Para substituents that are π-acceptors have the
greatest destabilizing effect for forming dimers and trimers. For
formation of adducts 1‚NH3, 2‚NH3 and 3‚NH3 from 1, 2 and
3, respectively, with free NH3, we find that the electron-with-
drawing ability of a substituent correlates well with its ability
to stabilize the adduct. Although the formation of 1‚NH3 is
calculated to be endothermic, formation of 2‚NH3 and 3‚NH3
are calculated to be exothermic. Our NMR data provided
References and Notes
(1) Hall, D. G. Boronic Acids; Wiley-VCH: New York, 2005.
(2) Morgan, A. B.; Jurs, J. L.; Tour, J. M. Polym. Prepr. (Am. Chem.
Soc., DiV. Polym. Chem.) 1999, 40, 553.
(3) Mehta, M. A.; Fujinami, T. Chem. Lett. 1997, 915.
(4) Yang, Y.; Inoue, T.; Fujinami, T.; Mehta, M. A. J. Appl. Polym.
Sci. 2002, 84, 17-21.
(5) Forsyth, M.; Sun, J.; Zhou, F.; MacFarlane, D. R. Electrochim.
Acta 2003, 48, 2129.
(6) Miyaura, N.; Suzuki, A. Chem. ReV. 1995, 95, 2457.

8166 J. Phys. Chem. A, Vol. 110, No. 26, 2006
Kua et al.
(7) Alcaraz, G.; Euzenat, L.; Mongin, O.; Katan, C.; Ledoux, I.; Zyss,
J.; Blanchard-Desce, M.; Vaultier, M. Chem. Commun. 2003, 22, 2766.
(8) Fielder, W. L.; Chamberlain, M. M.; Brown, C. A. J. Org. Chem.
1961, 26, 2154.
(9) Branch, G. E. K.; Yabroff, D. L. J. Am. Chem. Soc. 1932, 54, 2569.
(10) Mariategui, J. F.; Niedenzu, K. J. Organomet. Chem. 1989, 369,
137.
(11) Beckett, M. A.; Strickland, G. C.; Varma, K. S.; Hibbs, D. E.;
Hursthouse, M. B.; Malik, K. M. A. J. Organomet. Chem. 1997, 535, 33.
(12) Ritchey, J. M. Synthesis and Properties of Addition Complexes of
Boroxines and Other Selected Boron-Containing Systems. Ph.D. Thesis.
University of Colorado, 1968.
(13) Snyder, H. R.; Konecky, M. S.; Lennarz, W. J. J. Am. Chem. Soc.
1958, 80, 3611.
(14) Das, M. K.; Mariategui, J. F.; Niedenzu, K. Inorg. Chem. 1987,
26, 3114.
(15) Wu, Q. G.; Wu, G.; Brancaleon, L.; Wang, S. Organometallics
1999, 18, 2553.
(16) Vargas, G.; Hernandez, I.; Hopfl, H.; Ochoa, M.-E.; Castillo, D.;
Farfan, N.; Santillan, R.; Gomez, E. Inorg. Chem. 2004, 43, 8490.
(17) Beckett, M. A.; Brassington, D. S.; Owen, P.; Hursthouse, M. B.;
Light, M. E.; Malik, K. M. A.; Varma, K. S. J. Organomet. Chem. 1999,
585, 7.
(18) Tokunaga, Y.; Ueno, H.; Shimomura, Y.; Seo, T. Heterocycles
2002, 57, 787.
(19) Beckmann, J.; Dakternieks, D.; Duthie, A.; Lim, A. E. K.; Tiekink,
E. R. T. J. Organomet. Chem. 2001, 633, 149.
(20) Cote, A. P.; Benin, A. I.; Ockwig, N.; O’Keefe, M.; Matzger, A.
J.; Yaghi, O. M. Science 2005, 310, 1166.
(21) Perrtu, E. K.; Arnold, M.; Iovine, P. M. Tetrahedron Lett. 2005,
46, 8753.
(22) Kua, J.; Iovine, P. M. J. Phys. Chem. A 2005, 109, 8938.
(23) Jaguar V5.5; Schrodinger, LLC: Portland, OR, 2003.
(24) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(25) Becke, A. D. Phys. ReV. A 1988, 38, 3098.
(26) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200.
(27) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(28) Tannor, D. J.; Marten, B.; Murphy. R.; Friesner, R. A.; Sitkoff,
D.; Nicholls, A.; Ringnalda, M.; Goddard, W. A., III; Honig, B. J. Am.
Chem. Soc. 1994, 116, 11875.
(29) Marten, B.; Kim, K.; Cortis, C.; Friesner, R. A.; Murphy, R. B.;
Ringnalda, M. N.; Sitkoff, D.; Honig, B. J. Phys. Chem. 1996, 100, 11775.
(30) Wiberg, K. B.; Bailey, W. F. J. Am. Chem. Soc. 2001, 123, 8231.
(31) Nielsen, R. J.; Keith, J. M.; Stoltz, B. M.; Goddard, W. A., III. J.
Am. Chem. Soc. 2004, 126, 7967.
(32) Florian, J.; Warshel, A. J. Phys. Chem. B 1998, 102, 719.