Chemiluminescent autoxidation of α-silyl carbanions derived from 9-silyl-10-methylacridanes

Article abstract of DOI:10.3987/COM-04-10094

The autoxidation of α-silyl carbanions derived from 9-trimethysilyl- and 9-triphenylsilyl-10-methylacridanes (1a and 1b) produced the fluorescent N-methylacridone (NMA) accompanied by weak chemiluminescence. Differences in the chemiluminescence quantum yields and the time course of the light emission were detected between the reactions of 1a and 1b, which was due to the substituent on the silicon atom. Additionally, the 9-trimethylsilyl-10-methylacridinium salt also underwent a chemiluminescent reaction when reacted with alkaline hydrogen peroxide. It is probable that these chemiluminescent reactions would proceed via the Peterson-type reaction and involve the dioxasiletanide-like species during the reaction pathway.

Full text of DOI:10.3987/COM-04-10094

HETEROCYCLES, Vol. 63, No. 7, 2004, pp. 1667 - 1672
Received, 2nd April, 2004, Accepted, 27th May, 2004, Published online, 1st June, 2004
α
CHEMILUMINESCENT AUTOXIDATION OF -SILYL
CARBANIONS DERIVED FROM 9-SILYL-10-METHYLACRIDANES
Jiro Motoyoshiya,* Kunihiko Tokutake, Motoki Kuroe, Sachiko Yoshioka,
Yoshinori Nishii, and Hiromu Aoyama
Department of Chemistry, Faculty of Textile Science and Technology,
Shinshu University, Ueda, Nagano 386-8567, Japan
E-mail: jmotoyo@giptc.shinshu-u.ac.jp
Abstract – The autoxidation of α-silyl carbanions derived from 9-trimethysilyl-
and 9-triphenylsilyl-10-methylacridanes (1a and 1b) produced the fluorescent N-
methylacridone (NMA) accompanied by weak chemiluminescence. Differences
in the chemiluminescence quantum yields and the time course of the light
emission were detected between the reactions of 1a and 1b, which was due to the
substituent on the silicon atom. Additionally, the 9-trimethylsilyl-10-
methylacridinium salt also underwent a chemiluminescent reaction when reacted
with alkaline hydrogen peroxide. It is probable that these chemiluminescent
reactions would proceed via the Peterson-type reaction and involve the
dioxasiletanide-like species during the reaction pathway.
One of the most important features of silicon in organic chemistry is the strong affinity of an oxyanion for
the silyl moiety, and such a peculiar chemical property allows for the synthetically versatile reaction
1
called the Peterson reaction; the interaction of α-silyl carbanions with carbonyl compounds producing
olefins as the silicon variation of the Wittig-type reactions. On the other hand, the interaction between α-
silyl carbanions and molecular oxygen is, to our knowledge, only been slightly investigated, though the
2
autoxidation of α-phosphoryl carbanions is well studied. If the autoxidation of α-silyl carbanions
forming the corresponding carbonyl compounds proceeds via the Peterson-type reaction, 3 either cyclic or
linear peroxides should be involved during the reaction process, which might provide chemiluminescence
from the excited fluorescent molecules generated by decomposition of the peroxides. In fact, we have

found that the autoxidation of the phosphonate carbanions derived from 9-phosphorylacridanes was
4
chemiluminescent. Therefore, it is of significance to determine the behavior of α-silyl carbanions
toward oxygen and whether or not such a reaction is chemiluminescent. In this paper, we report our
results of the autoxidation of 9-silyl-10-methylacridanides and related reactions from the viewpoint of
chemiluminescence.
5
6
The autoxidation of the 9-trimethylsilyl- and 9-triphenylsilyl-10-methylacridanes (1a) and (1b) ,
prepared by reduction of N-methylacridinium iodide followed by silylation of 9-lithio-10-
methylacridanide with trimethylchlorosilane or triphenylchlorosilane, was carried out in the presence of
potassium tert-butoxide (t-BuOK) in aerated DMSO to give N-methylacridone (NMA) as the sole
fluorescent product in 55 and 52 % yield, respectively (Scheme 1). In the reaction of 1b, triphenylsilanol
was isolated in 64% yield along with NMA. 7 These products were those expected from the Peterson-type
reaction, namely, an α-silyl carbanion initially reacted with molecular oxygen to give the oxyanion (2),
and the final products would be formed by the cleavage of the four-membered peroxide, 1,2,3-
dioxasiletanide (3) generated by a Wittig-type reaction, or a linear silylperoxide (4) generated by a Brook-
type rearrangement as shown in Scheme 1.
SiR3
LiAlH4
1. LDA, THF
2. R3SiCl
ether
N
Me
N
Me
N
Me
I
1a: R=Me
1b: R=Ph
O
OOSiR3
R3Si
O
R3Si
O O
t-BuOK, O2
or
DMSO
N
Me
N
Me
N
Me
3
4
2
*
O
+
[ R3SiOH ]
N
Me
*
NMA
+
hν
NMA
Scheme 1

Of interest is the weak light emission observed when a concentrated solution of 1a or 1b was treated with
t-BuOK in aerated DMSO or DMF. In these reactions, the light emission was very weak, so that the
emitter could not be spectroscopically identified. As NMA was the sole fluorescent product of these
reactions, it is the most likely candidate for the emitter. The chemiluminescence quantum yields (ΦCL)
were estimated to be 5.0 x 10-7 and 1.6 x 10-6 einstein/mol for the reactions of 1a and 1b in aerated
DMSO,8 respectively, by a photon-counting method with a photomultiplier tube compared with the light
emission from a luminol chemiluminescence as a standard. These values were very low, but ΦCL for 1b
was by three times greater than for 1a. The more distinct difference was found in the time course of the
light emission as shown in Figure 1. These results show that the light emission from 1b lasted longer than
that from 1a, namely, the light forming process of 1b proceeds more slowly but effectively than that of 1a
due to the substituent effect of the silyl groups. A remarkable substituent effect in the Peterson reactions
of α-silyl amides and esters was reported, in which a large difference in the geometrical selectivity was
detected when the reagents having trimethylsilyl or triphenylsilyl were used in the reaction with
9
aldehydes. The difference both in the electronic factor and in the size of the silyl groups would lead to
the different kinetics between the reactions of 1a and 1b.
SiR3
t-BuOK, O2
600
+
hν
NMA
in DMSO
N
500
Me
1a: R= Me
1b: R= Ph
400
300
200
100
0
1b
1a
0
200
400
600
800
1000
1200 sec.
Figure 1 Time course of chemiluminescence reaction of 1a and 1b.
To explore the properties of the silylperoxide (5), one of the probable intermediates in the present
chemiluminescent autoxidation, its precursor, a hydroperoxide (6), was prepared by the reaction of N-

methylacridinium iodide and aqueous hydrogen peroxide in methanol as shown in Scheme 2. Thus an
10
inseparable mixture of the hydroperoxide (6) and the alcohol (7) was obtained in the ratio of 52 : 48,
which was subsequently treated with trimethylchlorosilane and triethylamine to give NMA in 12 % yield,
while no change in the mixture was found in the absence of trimethylchlorosilane. Therefore, the
silylperoxide (5) easily decomposes into NMA even under weak basic conditions, but this decomposition
reaction was not luminescent.
OOH
OH
aq. H2O2
+
N
Me
N
N
Me
Me
I
6
7
Me3Si Cl
Et3N
OOSiMe3
NMA
N
Me
5
Scheme 2
11
9-Trimethylsilylacridinium trifluoromethylsulfate (8) was prepared by a 4-step synthetic procedure via
9-trimethylsilylacridine (9) (Scheme 3) and examined for its reaction. A weak and instantaneous light
emission was observed with the quantitative formation of NMA, when 8 was treated with alkaline
hydrogen peroxide in the dark. The reaction path of this chemiluminescent reaction should be similar to
that for the autoxidation of 1, because the oxyanion (2) should also be formed by the attack of the
hydroperoxide anion on the 9-carbon of 8. As the similar reaction of N-methylacridinium iodide lacking a
silyl group gave no chemiluminescence, the silyl group of 8 plays a crucial role in the light formation
process.
SiMe3
DDQ
LiAlH4
1. LDA
2. Me3SiCl
N
N
N
H
H
H
Br
SiMe3
SiMe3
MeOSO2CF3
NMA
+
aq. H2O2 / OH
2
N
N
hν
Me
CF3SO3
9
8
Scheme 3

The chemiluminescent reactions presented in this study might proceed via two pathways 3 as described in
Scheme 1. Although current evidence for specifying the reaction pathway is insufficient, the substituent
effect observed in this study seems to suggest the involvement of the 1,2,3-dioxasiletanide-like species,
because the difference in the electronic and steric factors of the trimethylsilyl and triphenylsilyl would be
large enough to affect the rates of the formation of the four-membered ring. In addition, the cyclic
peroxide would be more suitable for the high-energy species producing the excited molecules rather than
the linear peroxide. It is also probable that a 1,2,3-oxasiletanide-like species is a transition state (Scheme
4), because it has been proposed that a 1,2-dioxetane is not needed as an intermediate for the
chemiluminescence reactions of 9-acridinepercarboxylates, but a transition state could also supply
sufficient energy for generation of the excited acridone anion.12 The ΦCL’s estimated in the present
reactions are much lower than those for the chemiluminescent autoxidation of the 9-phosphorylacrydanyl
4
anions, which would be due to an apparently less strained structure of a 1,2,3-dioxasiletanide than that
of a phospha-1,2-dioxetane as found when compared with an oxaphosphetane and an oxasiletanide . 13
δ -
O
R3Si
R3Si
O
O O
δ -
*
[ R3SiOH ]
+
NMA
N
N
Me
Me
2
Scheme 4
ACKNOWLEDGEMENTS
This work was partially supported by a Grant-in-Aid for the 21st Century COE program by the Ministry
of Education, Culture, Sports, Science, and Technology.
REFERENCES AND NOTES
1. (a) D. J. Peterson, J. Org. Chem., 1968, 33, 780. (b) D. J. Ager, Synthesis, 1984, 384.
2. (a) H. Zimmer, R. E. Koenigkramer, R. L. Cepulis, and D. M. Nene, J. Org. Chem., 1980, 45,
2018. (b) E. Vedejs, H. Mastalerz, G. P. Meier, and D. W. Powell, J. Org. Chem., 1981, 46, 5253.
(c) M. Mikolajczyk, W. Midura, and S. Grzejszczak, Tetrahedron Lett., 1983, 24, 2489.
3. (a) L. F. van Staden, D. Gravestock, and D. J. Ager, Chem. Soc. Rev., 2002, 31, 195. (b) T.
Kawashima, K. Naganuma, and R. Okazaki, Chem. Lett., 1999, 1139.

4. (a) J. Motoyoshiya, Y. Isono, S. Hayashi, Y. Kanzaki, and S. Hayashi, Tetrahedron Lett., 1994,
35, 5875. (b) J. Motoyoshiya, T. Ikeda, S. Tsuboi, T. Kusaura, Y. Takeuchi, S. Hayashi,
S. Yoshioka, Y. Takaguchi, and H. Aoyama, J. Org. Chem., 2003, 68, 5950.
5. S. Fukuzumi, T. Kitano, and K. Mochida, Chem. Lett., 1989, 2177. Selected data for 1a: mp 155-
156 oC (benzene-hexane). 1H NMR (400 MHz, CDCl3) δ – 0.09 (s, 9H), 3.22 (s, 3H), 3.38 (s,
1H), 6.73 (d, 2H, J = 7.36 Hz), 6.82-6.89 (m, 4H), 7.05-7.09 (m, 2H). 13C NMR (100 MHz,
CDCl3) δ –3.09, 33.24, 39.53, 112.05, 120.91, 125.94, 127.04, 127.89, 143.01. m/z 251 (M+).
Anal. Calcd for C17H12NSi: C, 76.35; H, 7.91; N, 5.24. Found: C, 76.03, H, 8.03; N, 5.26.
6. H. Gilman and G. D. Lichtenwalter, J. Org. Chem., 1958, 23, 1586. Selected data for 1b: mp 136-
138 oC (benzene-hexane). 1H NMR (400 MHz, CDCl3) δ 2.53 (s, 3H), 4.36 (s, 1H), 6.51-7.33 (m,
23H). 13C NMR (100 MHz, CDCl3) δ 32.58, 38.35, 112.23, 120.97, 127.62, 127.97, 128.36,
129.66, 133.34, 135.42, 136.98, 144.09. m/z 453 (M+). Anal. Calcd for C32H27NSi: C, 84.72; H, 6.00;
N, 3.09. Found: C, 84.84, H, 5.68; N, 2.95.
7. A solution of 1b (1.0 g, 3.7 mmol) in THF (50 mL) was bubbled with oxygen gas for 10 min, and
t-BuOK (0.49 g, 4.4 mmol) was then added to the solution. After being stirred for 2 h at room
temperature and under reflux for 1 h in an oxygen atmosphere, the solvent was removed in vacuo. The
residue was extracted with benzene. The triphenylsilanol was determined by comparison of an
authentic sample using liquid chromatography. NMA was isolated by purification using column
chromatography (0.24 g, 52 %).
8. A solution (2.0 mL) of 1a (1.0 x 10-4 M) in DMSO and a solution (1.0 mL) of t-BuOK (2.0 x 10-2 M)
in t-BuOH was mixed in air and the photons immediately generated from the reaction were counted
by a photomultiplier tube (R464, Hamamatsu Photonics K.K) in the range of 300-650 nm.
9. S. Kojima, H. Inai, T. Hidaka, and K. Ohkata, Chem. Commun., 2000, 1795.
10. Selected data for the mixture of 3 and 4: 1H-NMR (400 MHz, CDCl3) δ 3.50 (s, 1.44H, CH3 for 4),
3.53 (s, 1.56H, CH3 for 3), 5.56 (s, 0.48H, 9-H for 4), 6.06 (s, 0.52H, 9-H for 3). The alcohol (4) was
independently prepared by the reaction of N-acridinium iodide with saturated K2CO3 solution.
11. Selected data for 9: 1H NMR (400 MHz, CDCl3) δ 0.00 (s, 9H), 7.56 (t, 2H, J = 7.60 Hz), 7.77 (dd,
13
2H, J = 8.40, 7.60 Hz), 8.01 (d, 2H, J = 8.40 Hz), 8.25 (d, 2H, J = 8.40 Hz). C NMR (100 MHz,
CDCl3) δ 4.05, 99.67, 114.80, 125.75, 127.59, 128.64, 131.49, 144.15. Selected data for 7: 1H NMR
(400 MHz, CDCl3) δ 1.02 (s, 9H), 5.09 (s, 3H), 8.22-9.04 (m, 8H).
12. E. H. White, D. F. Roswell, A. C. Dupont, and A. A. Wilson, J. Am. Chem. Soc., 1987, 109, 5189.
13. A. Pelter, D. Buss, E. Colclough, and B. Singaram, Tetrahedron, 1993, 49, 7077.