A hypervalent pentacoordinate boron compound with an N-B-N three-center four-electron bond

Article abstract of DOI:10.1021/jo1024656

For the purpose of synthesizing and characterizing hypervalent boron compounds with strong hypervalent interaction, we have prepared a boron compound with a tridentate ligand bearing two pyrimidine rings as nitrogen donors. X-ray analysis and molecular orbital calculations suggested that the boron compound was of hypervalent pentacoordinate structure with an N-B-N hypervalent bond. Thus, we have prepared the first hypervalent second row element compound with apical N coordination. A breakdown of energy contributions by DFT calculations revealed that the N-B-N bond energy of the pentacoordinate state ground state (13) was 2.8 kcal mol^-1. Implications were that the conjugation energy difference of 6.6 kcal mol^-1 (14.2-7.6 kcal mol^-1) with the tetracoordinate state was a crucial factor for shifting stability toward the pentacoordinate structure.

Full text of DOI:10.1021/jo1024656

ARTICLE
pubs.acs.org/joc
A Hypervalent Pentacoordinate Boron Compound with an N-B-N
Three-Center Four-Electron Bond
Yuichi Hirano, Satoshi Kojima, and Yohsuke Yamamoto*
Department of Chemistry, Graduate School of Science, Hiroshima University, 1-3-1 Kagamiyama, Higashi-Hiroshima 739-8526, Japan
S Supporting Information
b
ABSTRACT: For the purpose of synthesizing and characterizing
hypervalent boron compounds with strong hypervalent interaction,
we have prepared a boron compound with a tridentate ligand bearing
two pyrimidine rings as nitrogen donors. X-ray analysis and molecular
orbital calculations suggested that the boron compound was of hyper-
valent pentacoordinate structure with an N-B-N hypervalent bond.
Thus, we have prepared the first hypervalent second row element
compound with apical N coordination. A breakdown of energy con-
tributions by DFT calculations revealed that the N-B-N bond energy
of the pentacoordinate state ground state (13) was 2.8 kcal mol-1
.
Implications were that the conjugation energy difference of 6.6 kcal
mol-1 (14.2-7.6 kcal mol-1) with the tetracoordinate state was a crucial factor for shifting stability toward the pentacoordinate
structure.
’ INTRODUCTION
to the hypervalent boron in 2 to a more electron-withdrawing
ammonium group (Me-Nþ) did not lead to a shortening of the
B-O distance.8 We attributed these results to the rigid nature of
the tridentate ligand that limits bond shortening.
Hypervalent compounds are main group element compounds
in which the electron count of the central atom formally exceeds
the number defined by the octet rule. Main group elements in
and below the third row of the periodic table such as silicon,
phosphorus, sulfur, and iodine are capable of expanding their
valency to form such species, and many stable compounds with
atoms bearing 10 formal electrons are known.1 For second row
elements such as carbon and boron, the corresponding electron
state is found in the transition state of one of the fundamental
organic reactions, the SN2 reaction, and the general conception is
that these elements are reluctant to form stable hypervalent
compounds. Although some theoretical calculations indicate the
possibility of such compounds,2,3 synthesis and isolation have
posed an extremely difficult challenge.4
Recently, we developed a sterically rigid anthracene-type
tridentate ligand and applied it as a scaffold for the synthesis of
hypervalent carbon and boron compounds 1 and 2 (Figure 1).5
Characterization by X-ray crystallographic analyses and density
functional theory (DFT) calculations suggested that these com-
pounds were hypervalent species. That is, the compounds were
highly symmetric and the C-O and B-O distances were
considerably shorter than the sum of the van der Waals radii.6
Incidentally, the use of a similar rigid tridentate with two
methoxy coordinating groups but with thioxanthene as the
backbone was found useful for the synthesis of a compound that
could be regarded as a hexacoordinate carbon compound.7
However, the long C-O and B-O distances compared with
corresponding general covalent bonds suggested that the attrac-
tive interactions were weak. Exchanging the aromatic carbon para
Since the capability to form short hypervalent bonds would
require the apical donor groups to be attached to a more flexible
backbone, carbon compounds 3 and boron compounds 4 using a
van Koten-type tridentate ligand with oxygen donor sites were
examined (Figure 1).9 X-ray electron-density analyses and DFT
calculations suggested that these compounds (3 and 4) were also
hypervalent compounds. However, the C-O and B-O dis-
tances of the van Koten-type compounds 3 and 4 were longer
than those of the anthracene-type compounds 1 and 2, contrary
to our expectations. For this ligand system it was found that
pinacol derivative 4a assumed tricoordinate structure, although
the difference in the electronic nature of the boron atom
compared with 4b could be considered to be rather small
(Figure 2). As another flexible tridentate system, 2,6-bis-
(alkoxycarbonyl)-1-borylbenzenes were also examined.10 The
pinacol and catechol systems (5a and 5b, respectively) were
found to give pentacoordinate compounds with average B-O
bond lengths shorter than those of the van Koten ligand system,
but longer than those for the anthracene system. For this system,
the diarylboryl-substituted compound 5c was of the tetracoordi-
nate state in which only one of the two oxygen atoms interacts
with the central atom. These results show that for the flexible
tridentate system, when the electrophilicity of the boron atom is
weak or the nucleophilicity of the donor group is weak, the
Received: December 16, 2010
Published: March 04, 2011
r
2011 American Chemical Society
2123
dx.doi.org/10.1021/jo1024656 J. Org. Chem. 2011, 76, 2123–2131
|

The Journal of Organic Chemistry
ARTICLE
Figure 3. Working hypothesis for the pyrimidine ligand system.
Scheme 1. Synthesis of the Tridentate Ligand Precursor
Bearing Two Dimethylpyrimidine Rings
Figure 1. Pentacoordinate carbon and boron compounds of two ligand
systems.
Figure 2. Comparison of structures among boron compounds bearing
different substituents.
compound tends to be tricoordinate, and when the strength of
the two factors is intermediate, the compound becomes penta-
coordinate. Furthermore, when both the electrophilicity of the
boron atom and the nucleophilicity of the donor group are
strong, the coordination state tends to be tetracoordinate.
To raise the strength of the interaction between the central
atom and two donor atoms, other donor groups with higher
coordination ability were also examined. For boron, compounds
bearing two dialkylphosphino,11a dialkylamino,11b,12 or alkylthio
groups,5,13 and for carbon, compounds bearing aryloxides11c,d
were found to form unsymmetric tetracoordinate structures
according to X-ray analyses. Thus, all in all, for strong donors,
it seems that there is a preference for the formation of one strong
bond over two weak bonds as in the case of 5c. These results
implied that groups with donor ability that is intermediate in
strength between the weak alkoxy group and the strong alkyl-
amino or alkylthio group were required. Aromatic nitrogen
compounds seemed to be appropriate for this purpose as
previously applied for other systems.15,16 Since a symmetric
system seemed adequate for our purpose, which included
determination of the rotation barrier, we decided to introduce
pyrimidine rings. Another advantage of this system that we
envisaged was that the pentacoordinate state would have two
effective aryl-aryl π-conjugations between the pyrimidine rings
and the benzene ring while the tetracoordinate state would have
just one. Thus, the former would have larger conjugation energy
gain that might overcome the differences in bond energies
between the weaker hypervalent bond in the former and the
stronger single coordination bond in the latter (Figure 3).
Herein, we report on the synthesis, isolation, and character-
ization of a hypervalent pentacoordinate boron compound with
an N-B-N three-center four-electron bond designed according
to our working hypothesis. We also describe our estimation of
the N-B-N bond strength along with analysis of energy
contributions of the penta- and the tetracoordinate states based
upon theoretical calculations.
’ RESULT AND DISCUSSION
Synthesis of the Tridentate Ligand Precursor and Incor-
poration of the Boron Moiety. The precursor of the new
tridentate ligand with two dimethylpyrimidine rings was pre-
pared as described in Scheme 1. The underlining strategy for the
target compound was that the chlorine atom was expected to
improve crystallization and that the methyl groups of the
pyrimidine rings were expected to both increase reactivity for
2124
dx.doi.org/10.1021/jo1024656 |J. Org. Chem. 2011, 76, 2123–2131

The Journal of Organic Chemistry
ARTICLE
Scheme 2. Synthesis of Boron Compound 11 Bearing the Tridentate Ligand
Figure 4. ORTEP drawing of 11 with thermal ellipsoids at the 50% probability level. All hydrogen atoms have been omitted for clarity.
susceptible to moisture. Thus, the methoxy groups were ex-
changed with ethylene glycol to afford a 1,3,2-dioxaborolane
derivative 11 (Scheme 2). This product was still sensitive to
moisture but single crystals could be obtained.
Structure of the Boron Compound Bearing the Tridentate
Ligand. Single crystals suitable for X-ray structural analysis were
obtained as yellow plate crystals from a mixture of dry CH2Cl2
and n-hexane under a N2 atmosphere. The X-ray structure of 11
is shown in Figure 4, and selected structural parameters are
summarized in the left-hand column of Table 1.
Table 1. Selected Structural Parameters of the Structures
Determined by X-ray Structural Analysis and Calculations
(B3PW91/6-3lG(d))
X-ray (11)
calculation (12) calculation (13)
bond distance/Å
B1-N1
2.5371(12)
2.5371(12)
2.5918
2.5918
2.6034
2.6034
B1-N1*
bond angle/deg
The crystal structure of 11 shows the 1,3,2-dioxaborolane ring
to be planar indicating that the boron atom is sp2 hybridized and
this plane to be perpendicular to the phenyl group to which it is
attached. Furthermore, the compound was completely symme-
trical with respect to the plane of the central boryl ring. The
central boron atom and two nitrogen atoms of the pyrimidine
rings [N1-B1-N1*: 166.38(11)°] were nearly colinear. In this
geometry of the three atoms, the nitrogen atoms are oriented so
that it is possible for the lone pair of each nitrogen atom to
directly face the empty p-orbital of the central boron atom.
Furthermore, the two B-N distances [B1-N1, B1-N1*:
2.5371(12) Å] were substantially shorter than the sum of van
der Waals radii (3.62 Å). These results suggest that the com-
pound assumes trigonal bipyramidal (TBP) structure about the
central boron atom (B1) with O1, O2, and C2 as the three
equatorial atoms, and with two nitrogen atoms (N1 and N1*)
and the boron atom constituting an N-B-N three-center four-
electron bond. The dihedral angles between the central benzene
ring and the two pyrimidine rings [C3-C2-C5-N2, C3*-
C2*-C5*-N2*: 1.73°] were near 0°, indicating there is max-
imum conjugative interaction among the three rings. The direc-
tions of the minute deviations were counterclockwise with
respect to each other in a Cs-symmetry fashion, thereby creating
a slight curvature of the backbone consisting of the three
essentially coplanar rings.
N1-B1-N1*
C1-B1-N1
C1-B1-N1*
O1-B1-N1
O1-B1-N1*
166.38(11)
83.19(8)
83.19(8)
93.93(5)
93.93(5)
165.39
164.78
82.70
82.70
93.97
93.97
82.24
82.24
91.80
91.80
dihedral angle/deg
C3-C2-C5-N2
1.73(18)
1.73(18)
0.00
0.00
12.55
12.55
C3-C2*-C5*-N2*
the Stille coupling reaction and inhibit nucleophilic attack to the ring
upon attempted incorporation of the boryl moiety. 4-Chloroaniline
6 was treated with I2 to give 7 in 69% yield.16 Bromination of 7 by
treatment with CuBr2 and t-BuONO afforded 8 in 68% yield.17 The
Stille coupling reaction of diiodide 8 with tin derivative 9 proceeded
to yield 10 in 13% yield.18 Thus, the precursor 10 was prepared from
4-chloroaniline 6 in 6% total yield. The low yield of the final step
may have been caused by both the reactivity of the reagents, which
was still insufficient, and the thermal instability of the product in
solution at high temperature.
The introduction of the boryl moiety was attempted by
lithiation of 10 followed by treatment with some boron com-
pounds. However, this was found to be nontrivial and the
diarylboryl group, for instance, could not be incorporated due
to what we suspect to be steric repulsion by the methyl groups.
Fortunately, the rather small methyl borate reacted to give the
dimethoxyboryl derivative. However, this was found to be extremely
Insights on the N-B-N Bond of the Boron Compound.
To gain more insight on the N-B-N bond, the structure of 11
was optimized by ab initio molecular orbital calculations at the
2125
dx.doi.org/10.1021/jo1024656 |J. Org. Chem. 2011, 76, 2123–2131

The Journal of Organic Chemistry
ARTICLE
Table 2. Comparison of Structural Parameters between Transition State 12 and Global Minimum 13
HF/6-31G(d)
B3PW91/6-31G(d)
B3LYP/6-31G(d)
B3PW91/6-311þþG(df,dp)
B3LYP/6-311þþG(df,dp)
X
Y
Z
20
12.55
12.55
0.18
11.79
11.79
0.13
17.19
17.19
0.46
17.56
17.56
0.49
20
0.50
HF/6-31G(d) level and by hybrid density functional theory
(DFT) calculations at several levels of calculation using the
Gaussian 09 program,19 with results shown in Table 2.
1.73°. Otherwise, the structure of the conformer was practically
the same as the X-ray structure. As for the global minimum, the
dihedral angles were both 12.55° in a conrotatory manner. The
dioxaborolane plane was also rotated in the same direction in a
gear-like sense, all together making the conformer pseudo-C2
symmetric in contrast to the Cs-type structure of 11, deviations of
which were actually negligible. Disregarding these features,
conformer 13 had structural parameters very similar to those
of 12, as shown in Table 1. The reason for the twisting seems to
be the small but not negligible steric hindrance between the
dioxaborolane plane and the methyl groups of the pyrimidine
rings flanking this plane. A point to note is that the energy
difference between “transition state” 12 and global minimum 13
was found to be only 0.18 kcal mol-1, indicating that the energy
potential surface in the vicinity of the minimum is rather flat.
Thus, it is not unreasonable to assume that the discrepancy in
geometry between experiment and theory is brought about by
crystal lattice forces that cancel out the steric hindrance.
Optimization using the X-ray structure as the initial structure
with the MO method gave conformer 12 (HF/6-31G(d)) with
dihedral angles of 0° between the phenyl group and each
pyrimidine ring as a transition state structure with a single
negative frequency value. Further optimization in the direction
of the negative vibration of 12 (HF/6-31G(d)) yielded con-
former 13 (HF/6-31G(d)) with dihedral angles of 20° as the
global minimum. In addition to the large deviation of the dihedral
angle from the X-ray structure, the array of pyrimidine nitrogens
and the central boron was not appropriate for assuming hyper-
valent interaction and in this case the boron atom is better
described as a tricoordinate species. Thus, this MO method was
not sufficient for reproducing the experimental results. For the
hybrid methods, the B3PW91 and B3LPY functionals were used
in combinations with basis sets of 6-31G(d) and 6-311þþG(df,
pd) for optimization. Similar to the MO calculations, the con-
former with the two dihedral angles of 0° was found to be a
transition state for all four calculations. As for the global
minimum, the dihedral angles and the energy differences with
conformer 12 were basis set dependent. The dihedral angle
values were ca. 12° and 17° with the 6-31G(d) and 6-311þþG-
(df,pd) basis sets, respectively, and the corresponding energy
differences were ca. 0.15 and 0.48 kcal mol-1. Since the former
smaller basis set seemed to be sufficient for obtaining structural
parameters similar to those for the X-ray structure, we decided to
use it for further computations. Furthermore, the B3PW91
functional has been found to be more superior to the B3LYP
functional for reproducing structural parameters of X-ray struc-
tures for analogous hypervalent carbon and boron compounds
bearing neutral oxygen donors at the apical positions with only a
slight overestimation of the bond lengths (deviation: 0.1% to
1.2%).5,7-10 Thus, this functional was adopted in a combination
with the 6-31G(d) basis set. Selected structural parameters of the
two states deduced with the B3PW91/6-31G(d) method are
shown in Table 1 alongside the experimental data. The geometry
of 12, a transition state, was C2v symmetrical with two identical
B-N bond lengths (2.5918 Å), which were slightly longer
(2.2%) than the experimental data (2.5371(12) Å), and the
dihedral angles between the central benzene ring and each pyrimi-
dine ring were 0.00°, which is smaller than the experimental value of
The molecular orbitals (MOs) of 12, which is structurally close to
the X-ray structure but is a “transition state”, add support for the
presence of a three-center four-electron bond (Figure 5). HOMO-
16, HOMO-11, and LUMOþ12 can be regarded as the bonding,
nonbonding, and antibonding orbitals of the three-center four-
electron bond, respectively. HOMO-16, in particular, is suggestive
of attractive interaction by donation of lone pair electrons from one
nitrogen atom of each pyrimidine into the vacant p orbital on the
central boron atom. The corresponding orbitals of 13, the global
minimum, have energy levels that are the same as those for 12.
Furthermore, the appearances are essentially the same as those of
12, in spite of lower symmetry, with the only differences visually
showing in LUMOþ12, the antibonding orbital. Especially note-
worthy is that although the dihedral angle is deviated away (12°)
from the ideal angle of 0° as in 12, the bonding orbitals of 13 are still
suggestive of similar attractive interactions. Thus, small structural
deviations do not seem to change electronic features. Natural bond
orbital (NBO) analysis indicated that the N-B-N hypervalent
bond energy of 13 was 2.4 kcal mol-1
.
Elucidation of the Tetracoordinate Rotational Transition
State. We have previously used the variable-temperature NMR
coalescence method for measuring the rotation barrier of similar
pyrimidine rings in compounds bearing the N-S-N bond for
evaluating hypervalent interaction in these sulfur compounds,14
and the method seemed suitable for determining the rotation
2126
dx.doi.org/10.1021/jo1024656 |J. Org. Chem. 2011, 76, 2123–2131

The Journal of Organic Chemistry
ARTICLE
Figure 5. Molecular orbitals of 12 and 13 (threshold: 2.1 ꢀ 10-4 eV-3/2 for the bonding orbital, 1.4 ꢀ 10-4 eV-3/2 for the others): antibonding orbital
(left), LUMOþ12 (3.0 eV); nonbonding orbital (central), HOMO-11 (-8.4 eV); bonding orbital (right), HOMO-16 (-9.3 eV), respectively, for
both sets of orbitals.
Figure 7. Optimized structures of 13 (pentacoordinate, GS) and 14
(tetracoordinate, TS). Dihedral angles between the central benzene and
each pyrimidine ring are 12°, 12° for 13 and 90°, 0° for 14. Transition
state 14 is 6.0 kcal mol-1 higher in energy than 13.
The local maximum 14 had dihedral angles of 90° and 0°, as
would be expected. The calculated structure of 14 along with that
of 13 is shown in Figure 7.
Conformer 14 with the dihedral angles of 90° and 0° was
found to be a transition state structure with only one negative
frequency of -34.8 cm-1 with vibration vectors which coincide
with the direction of the rotation of the pyrimidine ring with
respect to the pyrimidine-benzene bond. Contrary to expecta-
tions, the degree of pyramidalization of the boron atom is small
and its appearance is closer to the tricoordinate state rather than
the tetracoordinate state. However, the orientation of the
pyramidalization is one in which the larger lobe of the vacant
Figure 6. Energy diagram for 11 with restricted rotation of the
pyrimidine ring, calculated at the B3PW91/6-31G(d) level.
barrier for compound 11. However, presumably due to the low
energy barrier for stereomutation, no separation of the pyrimi-
dine methyl signals, which are inequivalent in a frozen state,
could be observed even at low temperatures with a 600 MHz
NMR instrument. Thus, we resorted to theoretical calculations.
To get a schematic view of the rotation process, the dihedral
angle between the central benzene ring and one of the side
pyrimidine rings was varied from 0° to 90° with increments of 5°
and the structures were optimized at each degree by using
theoretical calculations at the B3PW91/6-31G(d) level. The
results are shown in Figure 6. The figure shows a near sigmoidal
curve, indicating that the rotation is rather straightforward with
no apparent complications.
orbital of the boron atom faces the pyrimidine nitrogen. This
implies that there is B-N interaction, although it may be weak.
Natural bond orbital (NBO) analysis indicated that the B-N
bond energy of 14 was 14.1 kcal mol-1, which is rather small but
much larger than that for the N-B-N bond (2.4 kcal mol-1) in
13. The difference in energy between 13 and 14, which corre-
sponds to the rotation barrier of one of the pyrimidine rings, was
calculated as 6.0 kcal mol-1. This low value is reasonable
considering the fact that the rotation barrier could not be
experimentally determined. There is also the possibility of a
tricoordinate boron conformer in which the dioxaborolane plane
is in conjugation with the phenyl ring as the transition state for
the rotation, though additional motions such as scission of the other
2127
dx.doi.org/10.1021/jo1024656 |J. Org. Chem. 2011, 76, 2123–2131

The Journal of Organic Chemistry
ARTICLE
Figure 10. Determination of stabilization energy in the tetracoordinate
structure.
have minimum N-B-N interaction and conjugation energy.
Thus, the energy difference between stable state 13 and 15
should correspond to the sum of N-B-N interaction and
conjugation energy. A one-point calculation at the B3PW91/6-
31G(d) level for 15 gave the value of 17.0 kcal mol-1 for the
energy difference with 13 (Figure 8).
Figure 8. Comparison of stabilization energy.
To estimate the contribution of just the conjugation energy by
the two pyrimidine-benzene coplanarizations, we have also
calculated the energy difference between species 130 and 150
which have the boryl moiety replaced by a hydrogen atom in 13
and 15 with other geometric features kept the same, respectively
(Figure 9). The difference was computed to be 14.2 kcal mol-1
,
thereby suggesting that the sum of the stabilization gain by
conjugation is 14.2 kcal mol-1 (7.1 kcal mol-1 per one
coplanarization), and N-B-N interaction which corresponds
to the difference is 2.8 kcal mol-1 (17.0-14.2 kcal mol-1). This
value is in good agreement with the value of 2.4 kcal mol-1 from
the NBO analysis.
In the case of the tetracoordinate 14, upon transformation
from 15, in addition to the gain of energy from one coplanariza-
tion and the formation of the B-N coordination bond, loss of
energy due to weakening of the other three bonds involving the
boron atom could be perceived as a major contribution to the
overall energy (Figure 10). Thus, the sum of these energies could
be regarded as 11.0 kcal mol-1 (Figure 8). Exchanging the boryl
group with a hydrogen atom in 14 to 140 and carrying out a one-
point calculation gave 7.6 kcal mol-1 as the difference in energy.
This would correspond to the conjugation energy, and the fact
that the conjugation energy of 130 was estimated to be essentially
2-fold of that of 140 implies that our approximations are justified.
The difference of 3.4 kcal mol-1 (11.0-7.6 kcal mol-1) can be
regarded as the sum of B-N bond energy gain and energy loss
upon deformation of the geometry of the boron atom.
The overall results are summarized in Figure 11. Implications
are that the energy of the N-B-N bond in the pentacoordinate
structure (2.8 kcal mol-1 by derivation or 2.4 kcal mol-1 by
NBO analysis) is weaker than that of a single B-N bond in the
tetracoordinate structure (14.1 kcal mol-1 by NBO analysis). By
itself this would mean that the formation of the hypervalent N-
B-N is unfavorable compared with tetracoordination by a single
B-N. However, owing to the large energy gain by π-conjugation
stabilization upon assuming pentacoordinate structure (14.2 kcal
Figure 9. Determination of stabilization energy in the pentacoordinate
structure.
B-N bond and B-C (phenyl) bond rotation are additionally
required. However, it was found by computational optimizations
that such a conformer would be less stable than 13 by 11.4 kcal mol-
1
and 14 by 5.4 kcal mol-1, respectively. In this conformer, the
pyrimidine-benzene dihedral angles were both 45° and the
dioxaborolane-benzene dihedral angle was 20°. Therefore, this
conformer can be excluded as a relevant state.
Estimation of the Energy Relationship between the Pen-
tacoordinate and Tetracoordinate States. A major factor
contributing to the stabilization of conformer 13 other than
N-B-N interaction is conjugation energy gained upon copla-
narization of the pyrimidine rings and the benzene ring.20 Thus,
the approximation of the conjugation energy would be important
to elucidate the reason why the pentacoordinate state was more
stable than the tetracoordinate state. To this end, we assumed
conformer 15, in which the dioxaborolane plane is fixed with the
same geometry as 13 and the two pyrimidine rings are oriented
perpendicular to the phenyl ring. This conformer is expected to
2128
dx.doi.org/10.1021/jo1024656 |J. Org. Chem. 2011, 76, 2123–2131

The Journal of Organic Chemistry
ARTICLE
magnitude of such bonds. The large difference in conjugation
energy between the pentacoordinate and tetracoordinate states
indicated that the utilization of conjugation energy gain is an
effective strategy for obtaining hypervalent species in favor of
prospectively stable tetracoordinate species. Attempts to experi-
mentally determine the hypervalent bond strength and to utilize
our findings in preparing analogues with other nitrogen hetero-
cycles are in progress.
’ EXPERIMENTAL SECTION
General Procedures. The 1H NMR (400 MHz) chemical shifts
(δ) are given in ppm downfield from Me4Si, determined by residual
CHCl3 (δ = 7.26 ppm) or CDHCl2 (δ = 5.30 ppm). The 11B NMR (127
MHz) chemical shifts (δ) are given in ppm downfield from the external
BF3 Et2O (δ = 0 ppm). The 13C NMR (100 MHz) chemical shifts (δ)
3
are given in ppm downfield from Me4Si, determined by CDCl3 (δ = 77.0
ppm). The 119Sn NMR (149 MHz) chemical shifts (δ) are given in
ppm downfield from the external Me4Sn (δ = 0 ppm).
THF, Et2O, and n-hexane were purified by passing through a Glass
Contour solvent dispensing system under N2 atmosphere. CH2Cl2 and
CH3CN were purified by distilling from CaH2 under a N2 atmosphere.
Column chromatography was carried out by using silica gel or reverse-
phase HPLC. 2-Chloro-4,6-dimethylpyrimidine was synthesized accord-
ing to a literature procedure.21
Synthesis of 4-Chloro-2,6-diiodoaniline (7). To a solution of
I2 (2.84 g, 11.2 mmol) in EtOH (40 mL) was added Ag2SO4 (3.49 g,
11.2 mmol) followed by 4-chloroaniline (6) (510 mg, 4.00 mmol) at
room temperature. After being stirred for 1 h at room temperature, the
suspension was concentrated by using a rotary evaporator and the
residue was dissolved in AcOEt (150 mL). The organic layer was washed
with sat. Na2S2O3 aq (2 ꢀ 100 mL) and dried over Na2SO4. The solvent
was removed by using a rotary evaporator. Compound 7 (1.04 g, 2.75
mmol, 69%) was obtained as a black solid. 1H NMR (CDCl3, δ) 7.62 (s,
2H), 4.61 (br s, 2H). Mp 128.9-129.7 °C (lit.22 mp 129 °C).
Synthesis of 1-Bromo-4-chloro-2,6-diiodobenzene (8). To a
solution of 7(1.04 g, 2.75 mmol) in dry CH3CN (50 mL) was added CuBr2
(3.07 g, 13.8 mmol) followed by t-BuONO (0.52 mL, 4.1 mmol) at room
temperature. After being stirred for 12 h at 50 °C, the reaction mixture was
cooled to room temperature and treated with sat. NaHCO3 aq (50 mL).
The reaction mixture was extracted with AcOEt (3 ꢀ 50 mL). The
combined organic layer was washed with sat. Na2S2O3 aq (100 mL) and
dried over Na2SO4. The solvent was removed by using a rotary evaporator.
The crude product was purified by silica gel column chromatography
(CH2Cl2, Rf 0.90). Compound 8 (830 mg, 1.87 mmol, 68%) was obtained
as a black solid. Colorless single crystals suitable for X-ray structural analysis
Figure 11. Comparison of stabilization energy between the pentacoor-
dinate and the tetracoordinate species.
mol-1) compared with that for the tetracoordinate structure (7.6
kcal mol-1) combined with deformation energy loss, the total
stabilization energy in the pentacoordinate structure has become
larger than that for the tetracoordinate structure, thereby making
the pentacoordinate structure more favorable.
’ CONCLUSION
To gain further knowledge of hypervalent compounds of
second row elements, we have synthesized a new tridentate
bearing two pyrimidine rings, and have utilized it for the
synthesis, isolation, and characterization of the first hypervalent
pentacoordinate boron compound with an N-B-N three-
center four-electron bond, which also happens to be the first of
its kind for all second row elements. This was achieved although
the coordinating sites were less geometrically constricted than
the anthracene backbone system. In this structure, the two
pyrimidine rings and the central benzene ring were found to be
essentially coplanar, whereas theoretical calculations at the
B3PW91/6-31G(d) level predicted a slightly distorted ground
state with pyrimidine-benzene dihedral angles of 12.55°. How-
ever, the potential surface within this angle range was shallow
with an energy difference between optimized structures with
dihedral angles of 0° and 12.55° being only 0.18 kcal mol-1. This
suggested, at least for our system, that small deviations in the
structure do not divert the electronic property of the species. By
breaking down the energies of stabilization of the stable penta-
coordinate conformer into conjugation energy and bond energy
of the N-B-N, and combining calculations, we could estimate
the N-B-N bond energy to be 2.8 kcal mol-1. Although the
approximation was made by indirect means, it was in good agreement
with the value of 2.4 kcal mol-1 derived from NBO analysis, and
thus the elucidated value should give an indication of the
1
were obtained by recrystallization from EtOH in the open air. H NMR
(CDCl3, δ) 7.84 (s, 2H). 13C NMR (CDCl3, δ) 139.5 (CH), 134.6 (C),
133.8 (C), 99.6 (C). HRMS (APPI, positive) calcd (m/z) for [M
(C6H2BrClI2)]þ• 441.71124, found 441.71124. Elemental analysis calcd
(%) for C6H2BrClI2 with 0.5C2H6O (residual solvent from crystallization):
C 17.73, H 1.06. Found: C 17.57, H 1.02. Mp 123.6-124.4 °C.
Synthesis of 4,6-Dimethyl-2-(tri-n-butylstannyl)pyrimidine
(9). To a solution of (i-Pr)2NH (8.40 mL, 68.0 mmol) in dry THF
(70 mL) was added dropwise n-BuLi (1.58 M in n-hexane, 38.6 mL, 61.0
mmol) at 0 °C. After 30 min of stirring, (n-Bu)3SnH (15.5 mL, 57.0
mmol) was added dropwise to the solution. After being stirred for 10
min, the mixture was cooled to -78 °C. The solution was slowly
transferred with vigorous stirring to a solution of 2-chloro-4,6-dimethyl-
pyrimidine (7.65 g, 54.0 mmol) in dry THF (70 mL) at -78 °C. After
being stirred for 2 h, the solution was allowed to warm to 0 °C. After 2 h
of stirring, the reaction mixture was quenched with H2O (50 mL). The
reaction mixture was extracted with AcOEt (3 ꢀ 200 mL). The
combined organic layer was washed with brine (400 mL) and dried
2129
dx.doi.org/10.1021/jo1024656 |J. Org. Chem. 2011, 76, 2123–2131

The Journal of Organic Chemistry
ARTICLE
over Na2SO4. The solution was filtered through a pad of silica gel and the
silica gel was washed with AcOEt (400 mL). The filtrate was concen-
trated by using a rotary evaporator. The crude product was purified by
silica gel column chromatography (AcOEt: n-hexane = 1:10, Rf 0.25).
Compound 9 (13.6 g, 46.8 mmol, 87%) was obtained as a colorless oil.
1H NMR (CDCl3, δ) 6.82 (s, 1H), 2.41 (s, 6H), 1.58 (quint, 3J = 8 Hz,
6H), 1.33 (sext, 3J = 8 Hz, 6H), 1.14 (t, 3J = 8 Hz, 6H), 0.88 (t, 3J = 8 Hz,
9H). 13C NMR (CDCl3, δ) 187.1 (C), 163.5 (C), 118.1 (CH), 29.0
(CH2), 27.2 (CH2), 24.1 (CH3), 13.7 (CH2), 10.2 (CH3). 119Sn NMR
(CDCl3, δ) -74.2. Elemental analysis calcd (%) for C18H34N2Sn: C
54.43, H 8.63, N 7.05. Found: C 54.31, H 8.66, N 7.03.
Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ,
UK; fax þ44 1223 336033; e-mail deposit@ccdc.cam.ac.uk).
Theoretical Calculations. The MO and DFT calculations were
performed with the Gaussian 09 program package.19 The final geometry
optimizations were performed with DFT/B3PW91,25 with the 6-31G-
(d) basis set for all atoms. No symmetry constraints were imposed
during these optimizations.
The energy diagram for 11 with restricted rotation of the pyrimidine
ring (Figure 6) was obtained by plotting the energies of optimized
structures with dihedral angles varying from 0° to 90° in 5° increments.
Structures of 12, 13, and 14 were found by optimization of initial
structures with pyrimidyl-phenyl with two dihedral angles of 0°, 0°
(12-DFT), 15°, 15° (13-DFT), and 90°, 0° (14-DFT), respectively.
Synthesis of 1-Bromo-4-chloro-2,6-bis(4,6-dimethyl-2-pyri-
midinyl)benzene (10). A mixture of 8 (222 mg, 0.501 mmol), 9 (421
mg, 1.06 mmol), PdCl2(PPh3)2 (72 mg, 0.10 mmol), and LiCl (142 mg,
3.35 mmol) in p-xylene (2.5 mL) was heated under reflux for 23 h and
then cooled to room temperature. The reaction mixture was filtered
through Celite and the Celite was washed with AcOEt (5 mL). The
filtrate was concentrated by using a rotary evaporator. The crude
product was purified by column chromatography (silica gel, AcOEt: n-
hexane = 1:1, Rf 0.25) followed by reverse-phase HPLC (RT = 23).
Compound 10 (27 mg, 0.067 mmol, 13%) was obtained as a white solid. 1H
NMR (CDCl3, δ) 7.60 (s, 2H), 7.04 (s, 2H), 2.55 (s, 12H). 13C NMR
(CDCl3, δ) 166.5 (C), 164.8 (C), 142.7 (C), 132.8 (C), 130.4 (CH), 119.0
(C), 118.4 (CH), 23.7 (CH3). HRMS (APPI, positive) calcd (m/z) for
[M(C18H16BrClN4) þ H]þ 403.03196, found 403.03238. Elemental
analysis calcd (%) for C6H2BrClI2 with 0.25H2O: C 52.96, H 4.07, N
13.73. Found: C 52.97, H 3.90, N 13.70. Mp 160.7-161.5 °C dec.
Synthesis of 1-Chloro-3,5-bis(4,6-dimethyl-2-pyrimidinyl)-
4-(1,3,2-dioxaborolan-2-yl)benzene (11). To a solution of 10
(245 mg, 0.607 mmol) in dry Et2O (50 mL) was added dropwise t-BuLi
(1.55 M in n-pentane, 0.80 mL, 1.3 mmol) at -78 °C. After the mixture
was stirred for 2 h, B(OMe)3 (0.45 mL, 3.9 mmol) was added dropwise.
After being stirred for 20 h, the reaction mixture was allowed to warm to
room temperature. The solvent was removed by using a rotary evapora-
tor and CH3CN (3 mL) was added to the residue. The suspension was
filtered through a pad of Wakogel 50C18 and the silica gel was washed
with CH3CN. The filtrate was concentrated by using a rotary evaporator.
The crude product was roughly purified by reverse-phase HPLC (RT =
35). The boronic ester derivative (15 mg, 0.041 mmol, 7%) was obtained
as a white solid. However, since the boronic ester derivative easily
hydrolyzed to the boronic acid in the open air, further derivation was
carried out. To a mixture of the hydrolysate and Na2SO4 (15 mg, 0.10
mmol) were added glycol (0.39 M in dry CH2Cl2, 0.17 mL, 0.062 mmol)
and dry CH2Cl2 (3 mL) at room temperature. After being stirred for 72 h
at room temperature, the reaction mixture was filtered through a
membrane filter. To the filtrate was slowly added dry n-hexane (10
mL). Yellow single crystals of 11 suitable for X-ray structural analysis
’ ASSOCIATED CONTENT
S
Supporting Information. Spectroscopic and crystallo-
b
graphic data for new compounds, and atomic coordinates for
the calculated structures and CIF files for 8 and 11. This material
is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
yyama@sci.hiroshima-u.ac.jp
’ ACKNOWLEDGMENT
This work was supported by Grants-in-Aid for Scientific
Research (Nos. 17350021 and 21350029) from the Ministry of
Education, Culture, Sports, Science and Technology of the
Japanese Government. The authors thank Mr. Hitoshi Fujitaka
and Dr. Daisuke Kajiya of the Natural Science Center for Basic
Research and Development (N-BARD), Hiroshima University
for VT-NMR (600 MHz) measurements and HRMS measure-
ments, respectively. The computations were performed using the
Research Center for Computational Science, Okazaki, Japan and
Information Media Center, Hiroshima University.
’ REFERENCES
(1) Akiba, K.-y. Chemistry of Hypervalent Compounds; Wiley-VCH:
New York, 1999.
(2) For theoretical calculations on hypervalent nitrogen see:
(a) Ewig, C. S.; Van Wazer, J. R. J. Am. Chem. Soc. 1989, 111, 4172–
4178. (b) Michels, H. H.; Montgomery, J. A., Jr. J. Chem. Phys. 1990,
93, 1805–1813. (c) Bettinger, H. F.; Schleyer, P. v. R.; Schaefer, H. F., III.
J. Am. Chem. Soc. 1998, 120, 11439–11448. (d) Ponec, R.; Roithova, J.
Theor. Chem. Acc. 2001, 105, 383–392.
(3) For theoretical calculations on hypervalent carbon see:
(a) Gutsev, G. L. Chem. Phys. 1992, 166, 57–68. (b) Pierrefixe, S. C.
A. H.; Poater, J.; Im, C.; Bickelhaupt, F. M. Chem.—Eur. J. 2008,
14, 6901–6911. (c) Pierrefixe, S. C. A. H.; van Stralen, S. J. M.; van
Stralen, J. N. P.; Fonseca Guerra, C.; Bickelhaupt, F. M. Angew. Chem.,
Int. Ed. 2009, 48, 6469–6471. (d) Kikuchi, Y.; Ishii, M.; Akiba, K.-y.;
Nakai, H. Chem. Phys. Lett. 2008, 460, 37–41. (e) Atsumi, T.; Abe, T.;
Akiba, K.-y.; Nakai, H. Bull. Chem. Soc. Jpn. 2010, 83, 892–899.
(4) (a) Forbus, T. R., Jr.; Martin, J. C. J. Am. Chem. Soc. 1979,
101, 5057–5059. (b) Martin, J. C. Science 1983, 221, 509–514. (c) Lee,
D. Y.; Martin, J. C. J. Am. Chem. Soc. 1984, 106, 5745–5746. (d) Forbus,
T. R., Jr.; Martin, J. C. Heteroatom Chem. 1993, 4, 129–136. (e) Hojo,
M.; Ichi, T.; Shibato, K. J. Org. Chem. 1985, 50, 1478–1482.
1
were obtained from the solvent mixture under a N2 atmosphere. H
NMR (CD2Cl2, δ) 8.32 (s, 2H), 7.01 (s, 2H), 4.22 (s, 4H), 2.55 (s,
12H). 11B NMR (CD2Cl2, -80 °C, δ) 16 (br). HRMS (ESI, positive)
calcd (m/z) for [M (C20H20BClN4O2) þ H]þ 395.14406, found
395.14399. Mp 273.0-274.0 °C dec.
X-ray Crystal Structural Analyses. Crystals suitable for X-ray
structure determination were mounted on a CCD diffractometer and
irradiated with graphite monochromated Mo KR radiation (λ = 0.71073
Å) at 173 K for data collection. The structure was solved by a direct
method using the SIR-97 program.23 Refinement on F2 was carried out
using full-matrix least-squares with the SHELX-97 program.24 All non-
hydrogen atoms were refined by using anisotropic thermal parameters.
Selected structural parameters are summarized in Table 1. Crystallo-
graphic data have been deposited with the Cambridge Crystallographic
Data Centre; deposition nos. CCDC-804980 and 804981 for compounds 8
and 11, respectively. Copies of the data can be obtained free of charge via
http://www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge
(5) (a) Akiba, K.-y.; Yamashita, M.; Yamamoto, Y.; Nagase, S. J. Am.
Chem. Soc. 1999, 121, 10644–10645. (b) Yamashita, M.; Yamamoto, Y.;
Akiba, K.-y.; Nagase, S. Angew. Chem., Int. Ed. 2000, 39, 4055–4058.
2130
dx.doi.org/10.1021/jo1024656 |J. Org. Chem. 2011, 76, 2123–2131

The Journal of Organic Chemistry
ARTICLE
(c) Yamashita, M.; Yamamoto, Y.; Akiba, K.-y.; Hashizume, D.; Iwasaki,
F.; Takagi, N.; Nagase, S. J. Am. Chem. Soc. 2005, 127, 4354–4371.
(6) Emsley, J. The Elements, 3rd ed.; Oxford University Press:
Oxford, UK, 1998.
(7) Yamaguchi, T.; Yamamoto, Y.; Kinoshita, D.; Akiba, K.-y.;
Zhang, Y.; Reed, C. A.; Hashizume, D.; Iwasaki, F. J. Am. Chem. Soc.
2008, 130, 6894–6895.
(8) Yano, T.; Yamaguchi, T.; Yamamoto, Y. Chem. Lett. 2009,
38, 794–795.
(9) (a) Akiba, K.-y.; Moriyama, Y.; Mizozoe, M.; Inohara, H.; Nishii,
T.; Yamamoto, Y.; Minoura, M.; Hashizume, D.; Iwasaki, F.; Takagi, N.;
Ishimura, K.; Nagase, S. J. Am. Chem. Soc. 2005, 127, 5893–5901.
(b) Nakatsuji, J.-y.; Moriyama, Y.; Matsukawa, S.; Yamamoto, Y.; Akiba,
K.-y. Main Group Chem. 2006, 5, 277–285.
(10) Nakatsuji, J.-y.; Yamamoto, Y. Bull. Chem. Soc. Jpn. 2010,
83, 767–776.
(11) (a) Yamashita, M.; Watanabe, K.; Yamamoto, Y.; Akiba, K.-y.
Chem. Lett. 2001, 1104–1105. (b) Yamashita, M.; Kamura, K.; Yama-
moto, Y.; Akiba, K.-y. Chem.—Eur. J. 2002, 8, 2976–2979. (c) Yama-
shita, M.; Mita, Y.; Yamamoto, Y.; Akiba, K.-y. Chem.—Eur. J. 2003,
8, 3655–3659. (d) Yamaguchi, T.; Yamamoto, Y. Chem. Lett. 2007,
36, 1438–1439.
(12) (a) Toyota, S.; Futawaka, T.; Ikeda, H.; Oki, M J. Chem. Soc.,
Chem. Commun. 1995, 2499–2500. (b) Toyota, S.; Futawaka, T.;
Asakura, M.; Ikeda, H.; Oki, M. Organometallics 1998, 17, 4155–4163.
(13) Toyota, S.; Uemitsu, N.; Oki, M. Heteroatom Chem. 2004,
15, 241–245.
(14) (a) Akiba, K.-y.; Ohsugi, M.; Iwasaki, H.; Ohkata, K. J. Am.
Chem. Soc. 1988, 110, 5576–5578. (b) Ohkata, K.; Ohsugi, M.;
Yamamoto, K.; Ohsawa, M.; Akiba, K J. Am. Chem. Soc. 1996,
118, 6355–6369. (c) Yamauchi, Y.; Akiba, K.-y.; Nakai, H. Chem. Lett.
2007, 36, 1120–1121. (d) Atsumi, T.; Abe, T.; Akiba, K.-y.; Nakai, H.
Bull. Chem. Soc. Jpn. 2010, 83, 520–529.
(15) Hirano, Y.; Saiki, Y.; Taji, H.; Matsukawa, S.; Yamamoto, Y.
Heterocycles 2008, 76, 1585–1592.
(16) Charrier, N.; Demont, E.; Dunsdon, R.; Maile, G.; Naylor, A.;
O’Brien, A.; Redshaw, S.; Theobald, P.; Vesey, D.; Walter, D. Synthesis
2006, 3467–3477.
(17) Sun, C.; Lin, X.; Weinreb, S. M. J. Org. Chem. 2006, 71, 3159–
3166.
(18) Wilkinson, A. J.; Pushmann, H.; Howard, J. A. K.; Foster, C. E.;
Williams, J. A. G. Inorg. Chem. 2006, 45, 8685–8699.
(19) Frisch, M. J.; et al. Gaussian 09, Revision A.02; Gaussian, Inc.,
Wallingford, CT, 2009.
(20) Hydrogen bonding between a pyrimidine nitrogen and an
ortho proton of the phenyl group can also be envisioned, thus giving
rise to two more interactions in the coplanar conformation of 13.
However, since this hydrogen bonding includes interaction involving a
C-H bond, the H-N distance is long (2.32 Å), and the orientation of
the C-H bond is essentially orthogonal to the direction of the supposed
sp2 lone pair of the nitrogen atom of the pyrimidine ring; this contribu-
tion should be insignificant compared with the stabilization gained by
aryl-aryl π-conjugation.
(21) Vlꢀad, G.; Horvꢀath, I. T. J. Org. Chem. 2002, 67, 6550–6552.
(22) Dains, F. B.; Vaughan, T. H.; Janney, W. M. J. Am. Chem. Soc.
1918, 40, 930–936.
(23) Altomare, A.; Burla, M. C.; Camalli, M.; Cascarano, G. L.;
Giacovazzo, C.; Guagliardi, A.; Moliterni, A. G. G.; Polidori, G.; Spagna,
R. J. Appl. Crystallogr. 1999, 32, 115–119.
(24) Sheldrick, G. M. SHELX-97, Program for the Refinement of
Crystal Structures; University of Gottingen: Gottingen, Germany, 1997.
(25) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3098–3100. (b) Becke,
A. D. J. Chem. Phys. 1993, 98, 5648–5652. (c) Perdew, J. P.; Wang, Y.
Phys. Rev. B 1992, 45, 13244–13249.
2131
dx.doi.org/10.1021/jo1024656 |J. Org. Chem. 2011, 76, 2123–2131