Synthesis and structure of pentacoordinate hypervalent boron compounds bearing a 1,8-Dimethoxy-10-methylacridinium Skeleton

Article abstract of DOI:10.1246/cl.2009.794

Pentacoordinate hypervalent boron compounds 4a and 4b with a newly prepared 1,8-dimethoxy-10-methylacridinyl ligand were synthesized. X-ray crystallography revealed that the distances between the central boron and both oxygen atoms of the tridentate ligan

Full text of DOI:10.1246/cl.2009.794

794
Chemistry Letters Vol.38, No.8 (2009)
Synthesis and Structure of Pentacoordinate Hypervalent Boron Compounds
Bearing a 1,8-Dimethoxy-10-methylacridinium Skeleton
Tomoyuki Yano, Torahiko Yamaguchi, and Yohsuke Yamamotoꢀ
Department of Chemistry, Graduate School of Science, Hiroshima University,
1-3-1 Kagamiyama, Higashi-hiroshima 739-8526
(Received May 14, 2009; CL-090477; E-mail: yyama@sci.hiroshima-u.ac.jp)
Pentacoordinate hypervalent boron compounds 4a and 4b
R
R
Mes
Mes
with a newly prepared 1,8-dimethoxy-10-methylacridinyl ligand
were synthesized. X-ray crystallography revealed that the dis-
tances between the central boron and both oxygen atoms of
B
N
B
N
MeO
OMe MeO
OMe
˚
the tridentate ligand were in the range of 2.37–2.53 A, showing
the pentacoordinate structure. Cyclic voltammetry of 4b showed
that the potential of the reduction (E1=2 ¼ ꢁ0:57 V in CH2Cl2,
N
Me
3
OTf−
Me
Me
4
5
vs. SCE) is much lower than that of the corresponding Gabbaı’s
¨
Chart 2.
system 5 (ꢁ0:28 V), probably due to the difference in coplanar-
ity between the boryl group and the acridinyl skeleton in 4b and
5 (twist angle = 44.8ꢂ in 5 and 88.2ꢂ in 4b).
OMe
OMe
OMe
1. n-BuLi
1. t-BuLi
2. ClC(O)OMe
I
2. I2
N
OMe
OMe
Hypervalent compounds,1 which have more than eight elec-
trons in their valence shell, are widely known in the chemistry of
heavier main group elements but are very rare in second-row
elements such as carbon2a–2d and boron.2b,3 Recently, we report-
ed two series of hypervalent boron compounds 12b,3a and 23b
using a sterically rigid anthracene ligand or a relatively flexible
van Koten-type ligand (Chart 1). However, we found that the
interaction between the central boron atom and the coordinating
oxygen atoms varies with the rigidity of the skeleton and with
the substituents on the central boron. For example, with the
same substituent (catecholate), B–O distances in 1a [2.379(2),
N
I
Me
7 y. 65%
Me
6
MeO
O
OMe
MeO
LiAlH4
N
N
Me
Me
8 y. 51%
9 y. 84%
Scheme 1.
˚
˚
2.441(2) A: av. 2.41 A] are significantly shorter than those in
˚
˚
2a [2.527(9), 2.660(10) A (1st crystal), 2.496(10), 2.702(10) A
new acridinium tridentate ligand skeleton 3 and the boron com-
pounds 4 bearing the skeleton because of our additional interests
in the redox behavior in comparison with the corresponding
˚
(2nd crystal): av. 2.60 A], indicating that steric rigidity plays
an important role in the strength of the interaction between B
and O, although the substituent on the oxygen is different. In ad-
dition, the structure itself is different between 2a and 2b [B–O
Gabbaı’s system 54 (Chart 2). It is expected that 1,8-di-
¨
methoxy-9-lithio-10-methylacridan, which is the precursor of
4, would react with boron reagents easier than that of 1, 1,8-di-
methoxy-9-lithioanthracene, due to a steric reason.
˚
distance: 3.024(3), 3.155(3) A; no coordination between B and
O], indicating that the electronic and/or steric effects of the sub-
stituents of the central boron is large for the weakly coordinating
system. Because we were not successful in preparing 1b bearing
a pinacolate substituent, we have been interested in creating an-
other sterically rigid tridentate ligand system. Here we report a
As shown in Scheme 1, 1,8-dimethoxy-10-methylacridan
(9), which is the precursor of the new tridentate ligand, was syn-
thesized. Bis(3-methoxyphenyl)methylamine (6) was synthe-
sized by the Pd-catalyzed cross coupling reaction of 3-iodo-
anisole with m-anisidine followed by methylation with NaH
and MeI. The amine 6 was selectively dilithiated by refluxing
in n-hexane, and bis(2-iodo-3-methoxyphenyl)methylamine (7)
was obtained after treatment with I2. The treatment of 7 with
t-BuLi and ClC(O)OMe afforded 1,8-dimethoxy-10-methylacri-
done (8). The reduction of 8 with LiAlH4 afforded the acridan 9.
The introduction of boryl groups is illustrated in Scheme 2.
The reaction of the acridan 9 with n-BuLi followed by treatment
with the corresponding boron reagent afforded the desired 10a
and 10b. Without isolation of unstable 10a and 10b, the reaction
R
R
R
R
MeO
B
OMe
TolO
B
OTol
Tol = C6H4(p-Me)
2a : BR2 = CatB
2b : BR2 = PinB
1a : BR2 = CatB
1b : BR2 = PinB
ꢁ
with Ph3CþBF4 gave the boron compounds 4a and 4b in very
CatB =
O
O
PinB =
O
O
B
B
low yields, but 4a and 4b could be isolated by recrystallization
1
from THF/CH3CN.5 The H NMR spectra of 4a and 4b show
a symmetrical pattern of the tridentate ligand (one OMe group
Chart 1.
Copyright Ó 2009 The Chemical Society of Japan

Chemistry Letters Vol.38, No.8 (2009)
795
Table 1. Selected structural parameters for 1a, 2a, 2b, 4a, and
4b
R
R
BR
MeO
2OMe
1. n-BuLi
MeO
B
OMe
H
1a2b
2a3b
2b3b
4a
4b
2. a) CatBCl
Ph3C+BF4
−
b) PinBOiPr
+
N
Average
˚
2.41
2.60
3.09
2.41
2.51
9
N
−
B–O/A
O–B–O/ꢂ
BF4
Me
Me
167.1
160.5a
145.0
165.2
147.9
4a : BR2 = CatB y. 1%
4b : BR2 = PinB y. 10%
10a : BR2 = CatB
10b : BR2 = PinB
aAverage of two independent molecules.
Scheme 2.
ꢁ0:98 V vs. SCE). Furthermore, the potential of the first reduc-
tion of 5 is distinctly more positive than that of Mes3B and 10-
methylacridinium, but the peak potential of the first reduction of
4b (Ep ¼ ꢁ0:60 V vs. SCE) is same to that of 1,8-dimethoxy-10-
methylacridinium (Ep ¼ ꢁ0:60 V vs. SCE), which is the side-
product of the synthesis of 4 and shows an irreversible redox
wave. Although the difference can be due to the presence of elec-
tron-donating methoxy groups in 4 and the difference in the sub-
stituent on the central boron, we think that the difference in
coplanarity between the boryl group and the acridinyl skeleton
(twist angle = 44.8ꢂ in 5 and 88.2ꢂ in 4b) should be one of
the reasons, because the conjugation between the central boron
and the ꢀ system in the acridinium skeleton should reduce the
reductive potential. The distortion in 4b is caused by the steric
repulsion between the boryl group and the two methoxy groups
of the ligand. The introduction of other substituents on the boron
atom and examination of the structure and the reduction poten-
tial are in progress.
ꢁ
.
Figureꢁ1. ORTEP drawings (50% elliposoid) of 4a BF4 and
.
4b PF6 (solvent is omitted).
(6H), one NMe group (3H), three aryl-H (2H ꢃ 3)). The results
are similar to the cases of the anthracene system 1 and the flex-
ible van Koten-type system 2.
Although the conversion from 9 to 4 is not efficient, espe-
cially in 4a which is notꢁquite stable, we managed to obtain
single crystals of 4a BF4 and 4b PF6ꢁ, which was obtained
.
.
References and Notes
by a counter anion exchange using KþPF6ꢁ, suitable for X-ray
1
a) J. I. Musher, Angew. Chem., Int. Ed. Engl. 1969, 8, 54. b) K. Akiba,
Chemistry of Hypervalent Compounds, Wiley-VCH, New York, 1999.
a) K. Akiba, M. Yamashita, Y. Yamamoto, S. Nagase, J. Am. Chem. Soc.
1999, 121, 10644. b) M. Yamashita, Y. Yamamoto, K. Akiba, D.
Hashizume, F. Iwasaki, N. Takagi, S. Nagase, J. Am. Chem. Soc. 2005,
127, 4354. c) K. Akiba, Y. Moriyama, M. Mizozoe, H. Inohara, T. Nishii,
Y. Yamamoto, M. Minoura, D. Hashizume, F. Iwasaki, N. Takagi, K.
Ishimura, S. Nagase, J. Am. Chem. Soc. 2005, 127, 5893. d) T. Yamaguchi,
Y. Yamamoto, D. Kinoshita, K. Akiba, Y. Zhang, C. A. Reed, D.
Hashizume, F. Iwasaki, J. Am. Chem. Soc. 2008, 130, 6894.
ꢁ
ꢁ
.
.
analysis. The ORTEP drawings of 4a BF4 and 4b PF6 are il-
lustrated in Figure 1.6 The boron atoms of 4a and 4b are almost
planar as indicated by the sum of the bond angles around the
boron atom being almost 360ꢂ. The B–O1 and B–O2 distances
between the central boron atom and the oxygen atoms of the
two methoxy groups in 4a were 2.375(8) and 2.437(8) A, respec-
tively. Because the distances and the O1–B–O2 angle of 4a
[165.2(4)ꢂ] are almost the same as those of 1a [B–O1(O2) =
2.379(2), 2.441(2) A, O1–B–O2 = 167.10(7) ] where the B–
O attractive interaction was confirmed by experimental electron
density distribution analysis, the structure of 4a should be re-
garded as pentacoordinate.
The corresponding B–O distances in 4b were 2.525(9) and
˚
2.501(9) A, respectively, which are slightly longer than those
2
˚
3
a) M. Yamashita, Y. Yamamoto, K. Akiba, S. Nagase, Angew. Chem., Int.
Ed. 2000, 39, 4055. b) J. Nakatsuji, Y. Moriyama, S. Matsukawa, Y.
Yamamoto, K. Akiba, Main Group Chem. 2006, 5, 277. c) D. Y. Lee,
J. C. Martin, J. Am. Chem. Soc. 1984, 106, 5745.
2b
ꢂ
˚
4
5
C.-W. Chiu, F. P. Gabbaı, Angew. Chem., Int. Ed. 2007, 46, 1723.
¨
ꢁ
1H NMR (400 MHz, CDCl3): ꢁ 3.78 (s, 6H), 4.90 (s,
.
Data for 4a BF4
:
3H), 7.12 (d, 2H, 3J ¼ 8 Hz), 7.20–7.22 (m, 2H), 7.34–7.36 (m, 2H),
8.19 (d, 2H, 3J ¼ 8 Hz), 8.33 (t, 2H, 3J ¼ 8 Hz); HRMS m=z: calcd fꢁor
in 4a but are still much shorter than the sum of the van der Waals
7
þ
.
C
22H19BNO4 ([M] ), 372.1407; found: 372.1411. Data for 4b BF4
:
˚
radius of B and O (3.48 A). In the case of the flexible van Koten-
type ligand system 2, 2b is concluded to be a tricoordinate based
1H NMR (400 MHz, CD3CN): ꢁ 1.53 (s, 12H), 4.22 (s, 6H), 4.58 (s, 3H),
7.33 (d, 2H, 3J ¼ 8 Hz), 7.96 (d, 2H, 3J ¼ 9 Hz), 8.25 (dd, 2H, 3J ¼ 8
and 9 Hz); HRMS m=z: calcd for C22H28BNO4 (½M þ Hꢄþ), 381.2111;
˚
on the long B–O distances [3.024(3) and 3.155(3) A]. The B–O1
found: 381.2115.
and B–O2 distances in 4b are much shorter than those in 2b be-
cause of the steric rigidity of the tridentate ligand 4 (Table 1).
Therefore, it is concluded that in a sterically rigid acridinium li-
gand the electronic and/or steric effects of the substituents of the
central boron are not as large as those in a van Koten-type ligand.
Although pure 4a could not be obtained for the study of re-
dox behavior, the cyclic voltammetry of 4b in CH2Cl2 could be
measured. A single reversible redox wave (E1=2 ¼ ꢁ0:57 V in
CH2Cl2, vs. SCE) and a subsequent irreversible wave (Ep ¼
ꢁ1:48 V vs. SCE) were observed, and the potentials of reduc-
ꢁ
.
.
6
Crystal Data for 4a BF4 CH3CN: C24H22B2F4N2O4, Mr: 500.06, mono-
clinic, P21=n (No. 14), a ¼ 11:2300ð7Þ, b ¼ 11:4590ð7Þ, c ¼ 18:1700ð12Þ
A, V ¼ 2337:5ð3Þ A , Z ¼ 4, Dcalcd ¼ 1:421 g cmꢁ3, R ¼ 0:1021 (I >
3
˚
˚
2ꢂðIÞ), Rw ¼ 0:3754 (all data), GOF ¼ 1:113 for 4460 reflections and
357 parameters (CCDC-733927). Crystal Data for 4b PF6ꢁ: C22H27BF6-
.
NO4P, Mr: 525.23, monoclinic, P21=c (No. 14), a ¼ 7:1230ð6Þ,
3
˚
˚
b ¼ 27:711ð2Þ, c ¼ 11:8230ð12Þ A, V ¼ 2320:8ð4Þ A , Z ¼ 4, Dcalcd
¼
1:503 g cmꢁ3, R ¼ 0:0917 (I > 2ꢂðIÞ), Rw ¼ 0:3431 (all data), GOF ¼
1:130 for 4311 reflections and 323 paꢁrameters (CCDC-733928). The data
were collected at 200 K (for 4a BF4 ) or 173 K (for 4b PF6ꢁ) using a
.
.
Mac Science DIP 2030 imaging plate equipped with graphite-monochro-
˚
tions are much lower than that of corresponding Gabbaı’s system
¨
5,4 which shows two reversible redox waves (E1=2 ¼ ꢁ0:28,
mated Mo Kꢃ radiation (ꢄ ¼ 0:71073 A).
7
J. Emsley, The Elements, 3rd ed., Oxford University Press, 1998.
Published on the web (Advance View) July 4, 2009; doi:10.1246/cl.2009.794