Chemiluminescence in autoxidation of phosphonate carbanions. Phospha-1,2-dioxetanes as the most likely high-energy intermediates

Article abstract of DOI:10.1021/jo030046l

Autoxidation of the phosphonate carbanions derived from 9-phosphono-10-hydroacridanes provided chemiluminescence ascribed to the excited acridone anion. The intramolecular CIEEL (chemically initiated electron exchange luminescence) mechanism can be applied to this chemiluminescence because of the much higher emission efficiency compared to that of 9-phosphono-10-methylacridanes. The effect of the phosphonate substituents on the emission efficiency and especially on the rates of the chemiluminescence decay can be interpreted to originate from the valence deviation of the phosphorus atoms, which is connected with the substituent effect on the geometrical selectivity in the olefination reaction of the phosphonate carbanions. The enhanced chemiluminescence in the presence of the fluorophores was also detected in autoxidation of the carbanions of diethyl diphenylmethylphosphonate and fluorenylphosphonate. Although the evidence is circumstantial, these results strongly support the belief that phospha-1,2-dioxetane is the most likely high-energy intermediate generating the excited molecules.

Full text of DOI:10.1021/jo030046l

Ch em ilu m in escen ce in Au toxid a tion of P h osp h on a te Ca r ba n ion s.
P h osp h a -1,2-d ioxeta n es a s th e Most Lik ely High -En er gy
In ter m ed ia tes
J iro Motoyoshiya,* Tsuneaki Ikeda, Shuji Tsuboi, Tatsuya Kusaura, Yoshino Takeuchi,
Satoko Hayashi, Sachiko Yoshioka, Yutaka Takaguchi, and Hiromu Aoyama
Department of Chemistry, Faculty of Textile Science & Technology, Shinshu University, Ueda,
Nagano 386-8567, J apan
jmotoyo@giptc.shinshu-u.ac.jp
Received February 3, 2003
Autoxidation of the phosphonate carbanions derived from 9-phosphono-10-hydroacridanes provided
chemiluminescence ascribed to the excited acridone anion. The intramolecular CIEEL (chemically
initiated electron exchange luminescence) mechanism can be applied to this chemiluminescence
because of the much higher emission efficiency compared to that of 9-phosphono-10-methylacridanes.
The effect of the phosphonate substituents on the emission efficiency and especially on the rates of
the chemiluminescence decay can be interpreted to originate from the valence deviation of the
phosphorus atoms, which is connected with the substituent effect on the geometrical selectivity in
the olefination reaction of the phosphonate carbanions. The enhanced chemiluminescence in the
presence of the fluorophores was also detected in autoxidation of the carbanions of diethyl
diphenylmethylphosphonate and fluorenylphosphonate. Although the evidence is circumstantial,
these results strongly support the belief that phospha-1,2-dioxetane is the most likely high-energy
intermediate generating the excited molecules.
In tr od u ction
gested as the pertinent intermediates. Recently, we
reported the observation of chemiluminescence during
the oxygenation of phosphonate carbanions.4 This chemi-
luminescence reaction is also expected to involve phos-
pha-1,2-dioxetanes as the key high-energy intermediates.
Elucidation of such intermediates will be significant for
further understanding of the Horner-Wadsworth-Em-
mons (HWE) reaction,5 a very important olefination
reaction of phosphonate carbanions, whose intermediate
has been investigated less than that of the Wittig
reaction.6 This study was conducted to explore the
chemiluminescent autoxidation of phosphonate carban-
ions from the viewpoint of chemiluminescence, thus
providing strong support for phospha-1,2-dioxetanes as
the high-energy intermediates as well as supplying a new
aspect for reactions of phosphonate carbanions.
Phospha-1,2-dioxetanes (1) have been proposed 1 to be
the most likely intermediates in the oxygenation of
phosphonium ylides, sometimes called the oxy-Wittig
reaction, which generates the corresponding carbonyl
compounds and phosphine oxides. In a structural simi-
larity to the well-known 1,2-dioxetanes (2)2,3 providing a
peculiar phenomenon, chemiluminescence, the phospha-
1,2-dioxetanes are also expected to display chemilumi-
nescence along with decomposition, where the electroni-
cally excited carbonyl compounds would be generated
similarly to thermal decomposition of 1,2-dioxetanes.
Resu lts a n d Discu ssion
Indeed, light emission has been observed during the
singlet oxygenation of suitable phosphonium ylides1b,c and
phosphazines,1d where phospha-1,2-dioxetanes were sug-
The autoxidation of phosphonate carbanions derived
from 9-phosphono-10-hydroacridanes (3) and t-BuOK in
aprotic solvents provided chemiluminescence that lasted
long enough to be spectroscopically detected, and whose
(1) (a) J ohnson, A. W. Ylid Chemistry; Academic Press: New York,
1966; Chapter 3, pp 16-125. (b) Suzuki, N.; Wakatsuki, S.; Izawa, Y.
Tetrahedron Lett. 1980, 21, 2513-2514. (c) Akasaka, T.; Sato, R.; Ando,
W. J . Am. Chem. Soc. 1985, 107, 5539-5540. (d) Akasaka, T.; Sato,
R.; Miyama, Y.; Ando, W. Tetrahedron Lett. 1985, 26, 843-846.
(2) Adam, W. Four-Membered Ring Peroxides: 1,2-Dioxetanes and
R-Peroxylactones, The Chemistry of Peroxides; Patai, S., Ed.; J ohn Wiley
& Sons, Ltd.: New York, 1983; Chapter 24, pp 829-920.
(3) Turro, N. J . Modern Molecular Photochemistry; The Benjamin/
Cummings Publishing Co., Inc.: California, 1978; Chapter 14, pp 579-
614.
(4) (a) Motoyoshiya, J .; Isono, Y.; Hayashi, S.; Kanzaki, Y.; Hayashi,
S. Tetrahedron Lett. 1994, 35, 5875-5878. (b) Motoyoshiya J .; Tsuboi,
S.; Kokin, K.; Takaguchi, Y.; Hayashi, S.; Aoyama, H. Heterocycl.
Commun. 1998, 4, 25-29. (c) Motoyoshiya, J .; Takaguchi, Y.; Aoyama,
H. Phosphorus, Sulfur Silicon 2002, 177, 1977-1979.
(5) (a) Horner, L.; Hoffmann, H.; Wippel, H. G.; Klahre, G. Chem.
Ber. 1959, 92, 2499-2505. (b) Wadsworth, W. S., J r.; Emmons, W. D.
J . Am. Chem. Soc. 1961, 83, 1733-1738.
(6) Maryanoff, B. E.; Reitz, A. B. Chem. Rev. 1989, 89, 863-927.
10.1021/jo030046l CCC: $25.00 © 2003 American Chemical Society
Published on Web 06/24/2003
5950
J . Org. Chem. 2003, 68, 5950-5955

Chemiluminescence of Phosphonate Carbanions
F IGURE 1. ESR spectra of the acridanyl ketyl radicals generated from 9-diethylphosphono-10-methylacridane (4b) and
9-ethoxycarbonyl-10-methylacridane and their computer-simulated spectra.
emission spectra were in complete agreement with the
fluorescence spectrum of the anion of acridone (5) inde-
pendently generated under basic conditions. However,
the chief product in these chemiluminescence reactions
was not 5 but weakly fluorescent 9-phosphonoacridine
(7). For instance, 5 and 7a (R ) OMe) was obtained in
8% and 85% isolated yields, respectively, from the
thorough autoxidation of 3a (R ) OMe) with t-BuOK in
were made to detect a ketyl radical adjacent to the
phosphorus atom. The ESR spectrum of the radical
DMSO. On the other hand, similar reactions of 4 also
showed chemiluminescence due to the fluorescence of
N-methylacridone (6) as previously reported.4 Although
one could see this weak, pale blue light emission with
the naked eye in the dark when the reaction was carried
out in high concentration, it was too weak to be detected
spectroscopically. In the 31P NMR spectrum of 4f (R )
OPh) treated with t-BuOK in THF, the signal for 4f at
δ +14.67 ppm became smaller with the increase in a new
signal at δ -9.54 ppm ascribed to the diphenyl phosphate
anion with time. To scavenge the phosphorus fragment,
a thorough autoxidation of 4f was carried out in the
presence of t-BuOK in THF to give 6 and potassium
diphenyl phosphate in 89% and 80% isolated yields,
respectively, which were those expected from the oxy-
Wittig type of reaction.
generated in the colored solution prepared from 4b (R )
OEt) and t-BuOK in aerated DMSO at room temperature
was successfully recorded as shown in Figure 1, in which
the g value was estimated to be 2.0027, and the whole
spectrum was split in two. A part of the spectrum is
similar to that of the 9-methyl-10-carbethoxyacridanyl
ketyl radical (Figure 1) and the 9-phenylsulfenyl-10-
methylacridanyl ketyl radical.8 The hyperfine coupling
constants listed in Figure 1 were estimated by computer
simulation, referring to the data documented by J anzen
et al.9 The value, 19.5G, found for the coupling constant
for the interaction of the unpaired electron and the
phosphorus atom is very small compared to the values
previously reported,10 and therefore, this radical is highly
delocalized over the aromatic rings. The light emission
from the solution of 4b was quenched at once when
Because radical species have sometimes been observed
during autoxidation of stabilized carbanions,7 attempts
(8) Motoyoshiya, J .; Inoue, H.; Takaguchi, Y.; Aoyama, H. Heteroat.
Chem. 2002, 13, 252-257.
(9) Happ, W. J .; J anzen, E. J . J . Org. Chem. 1970, 35, 96-102.
(7) Scott, G. Atmospheric Oxidation and Antioxidants; Elsevier: The
Netherlands, 1965; Chapter 3.
J . Org. Chem, Vol. 68, No. 15, 2003 5951

Motoyoshiya et al.
TABLE 1. Ch em ilu m in escen ce Qu a n tu m Yield s a n d
Kin etic P a r a m eter s
SCHEME 1
efficiency when the phenolic proton of the firefly luciferin
was replaced by a methyl group. As depicted in Scheme
1 similar circumstances are furnished in the present
chemiluminescence if the phospha-1,2-dioxetanes could
be postulated for the reaction intermediates as well as
the electron acceptors such as the 1,2-dioxetanes, because
an amide anion in 3 would have a much lower oxidation
potential than the lone pair of the methylated nitrogen
in 4. An increase in Φs by an electron-attracting 2,2,2-
trifluoroethyl (TFE) group was also observed, which is
probably due to the promotion of ring closure (vide infra)
as well as the CIEEL process.
a
b
Determined by UV spectrum. Estimated by the photon-
counting method using luminol chemiluminescence as standard
(einstein/mol). c Reference 12. Arbitrary unit standarized from
d
the value kb for 4a . e Arbitrary unit standardized from the value
for 4a determined from the curves in Figure 1. The higher values
were regarded as ka and the lower ones as kb.
Of interest was the drastic change in the emission
profiles as shown in Figure 2. The maximum intensity
appeared at about 60 s for 3a , 3b, and 3e, whereas that
for 3d and 3e occurred faster. Especially, the maximum
of 3d occurred within 1 s. The curves in Figure 2 could
be matched by an equation for the combined exponential
contribution of the two parameters, ka and kb, which is
adapted to sequential reaction process. Thus, the emis-
sion intensity I at any time t is expressed by the following
equation:
radical trapping agents such as powdered sulfur or
p-benzoquinone were added to the solution, whereas the
weak emission generated from the ionic reaction of
9-diethylphosphonoacridine (7b, R ) OEt) and alkaline
hydrogen peroxide was not interrupted. Thus, the present
chemiluminescent autoxidation of the phosphonate car-
banions proceeds via interaction of atmospheric oxygen
and the acridanyl ketyl radical, probably generated by
an electron transfer to molecular oxygen.
The remarkable effect of the phosphonate substituents
on the geometric selectivity in the HWE reaction has been
well documented.11 This effect is believed to originate
from the various rates for oxaphosphetane ring forma-
tion, namely, the acceleration of the ring formation
results in an increase in the ratio of thermodynamically
less favored Z-olefins. To explore the substituent effect
on the chemiluminescence, a kinetic study based on
chemiluminescence decay was done on the reactions of 3
and 4. The collected data of the yields of the emitter (Φr)
determined by UV spectrum and singlet excited yields
(Φs) are listed in Table 1. There is a Φs value for 3
conspicuously much larger than that for 4,12 which
explains the contribution of an intramolecular CIEEL
(chemically initiated electron exchange luminescence)
process as proposed for the firefly luciferin biolumines-
cence:13 namely, strong support for the CIEEL mecha-
nism was due to the drastic decrease in the emission
I ) M{exp(-kat) - exp(-kbt)}
where M is proportional to the relative maximum inten-
sity (rel Imax). Such a sequential reaction can be seen in
the peroxyoxalate chemiluminescence,14 in which the key
species are presumed to be cyclic peroxides such as 1,2-
dioxetanones. In the present system, if either kinetic
parameter, ka or kb, corresponds to a ring-closing step
producing a phospha-1,2-dioxetane, the structure of the
phosphonate substituents should affect the rates, as has
been found in the HWE reaction.15,16 Thus, an enhance-
ment of the rates is expected for 3c, 3d , 4c, and 4d ,
because the formation of the phospha-1,2-dioxetane ring
would be promoted by the electronegative substituents
or the strained five-membered cyclic moieties.17 According
to the technique used for kinetic analysis of the peroxy-
(13) Koo, J .-Y.; Schmidt, S. P.; Schuster, D. B. Proc. Natl. Acad. Sci.
U.S.A. 1978, 75, 30-33.
(10) (a) Baban, J . A.; Roberts, B. P. J . Chem. Soc., Chem. Commun.
1979, 373-374. (b) Baban, J . A.; Cooksey, C. J .; Roberts, B. P. J . Chem.
Soc., Perkin Trans. 2 1979, 781-787. (c) Alberti, A.; Griller, D.; Nazran,
A. S.; Pedulli, G. F. J . Org. Chem. 1986, 51, 3959-3963.
(11) For a review, see: (a) Motoyoshiya, J . Trends Org. Chem. 1998,
7, 63-73. (b) Ando, K. Synth. Org. Chem. J pn. 2000, 58, 869-876.
(12) The chemiluminescence quantum yields for 4 were reinvesti-
gated to find that the values were lower by 1 order of magnitude than
those reported in ref 4b.
(14) (a) Hadd, A. G.; Robinson, A. L.; Rowlen, K. L.; Birks, J . W. J .
Org. Chem. 1998, 63, 3023-3031. (b) Stevani, C. V.; Lima, D. F.;
Toscano, V. G.; Baader, W. J . J . Chem. Soc., Perkin Trans. 2 1996,
989-995.
(15) Larsen, R. O.; Aksnes, G.Phosphorus Sulfur 1983, 15, 229-
237.
(16) Motoyoshiya, J .; Kusaura, T.; Kokin, K.; Yokoya, S.; Takaguchi,
Y Narita, S.; Aoyama, H. Tetrahedron 2001, 57, 1715-1721.
(17) Breuer, E.; Bannet, D. M. Tetrahedron 1978, 34, 997-1002.
5952 J . Org. Chem., Vol. 68, No. 15, 2003

Chemiluminescence of Phosphonate Carbanions
F IGURE 2. Emission profiles for the chemiluminescence reactions of 3a -e.
TABLE 2. En er gy Differ en ce betw een Oxya n ion s a n d
P h osp h a -1,2-d ioxeta n es
SCHEME 2
oxalate chemiluminescence reaction,11 computational
work fitting the emission curves allowed us to estimate
the relative values of both parameters ka and kb in each
reaction, as well as relative maximum emission intensity
(Imax.). These data are shown in Table 1. As expected, a
drastic increase in ka by some hundred times compared
to the others was found in the reaction of 3d . An
appreciable acceleration in 4c and 4d was also found,
but that for 3c was negligible. Such a large enhancement
of the rate in 3d was also observed when the reactants
and base concentrations were changed; therefore, ka does
not depend on the initial carbanion formation or the
sequential oxygenation but on the valence transformation
of the phosphorus atoms. This suggests that ka is the rate
corresponding to the ring-closure step. Accordingly, the
light emission synchronizes with the ring formation;
namely, the high-energy species providing chemilumi-
nescence should be the phospha-1,2-dioxetanes. There-
fore, we can illustrate the whole reaction path involving
the acridanyl ketyl radical (8) and phospha-1,2-dioxe-
tanes (9) as depicted in Scheme 2.
a
Calculated at the RHF/6-31G(d) level of theory.
with those for 3a and 3e was evaluated, which resulted
in the acceleration of the ring formation, as was observed
above.
The autoxidation of other phosphonate carbanions
derived from diethyl diphenylmethylphosphonate (10)
and diethyl fluorenylphosphonate (11) showed only a
faint light emission in the absence of the fluorescent
additives, but DBA (9,10-dibromoanthracene, a triplet
energy acceptor) or DPA (9,10-diphenylanthracene, a
singlet energy acceptor) enhanced the chemiluminescence
despite the lower energy for the excited benzophenone
(ET1 ) 68-69 kcal/mol) and fluorenone (ET1 ) 53 kcal/
mol, ET2 ) 60 kcal/mol, ES1 ) 63.2 kcal/mol) compared
with that for triplet DBA (ET1 ) 71 kcal/mol) and that
for singlet DPA (ES1 ) 72.9 kcal/mol).18 The linear
correlation in the Stern-Volmer plot of the double
reciprocal of the DBA concentration and the chemilumi-
nescence quantum yields establishes a bimolecular pro-
cess between the excited species and DBA (Figure 3).
A few examples of such an upward energy transfer in
the DBA-enhanced chemiluminescence from excited ben-
zophenone have been reported previously,8,19 but a fur-
ther study is needed to clarify this energy transfer.20 The
The energetic advantage for the five-membered cyclic
phosphonate during the ring formation was also sup-
ported by calculations using the ab initio method at the
RHF/6-31G(d) level of theory. Table 2 lists the energies
and the valence angles of the oxyanions and the corre-
sponding phospha-1,2-dioxetanes of 3a , 3d , and 3e. The
smaller valence angle for the oxyanion of 3d compared
wtih those of 3a and 3e leads to easy liberation of its
strain energy by transformation into the phospha-1,2-
dioxetane. Thus, a much larger energy gap between the
oxyanion and the phospha-1,2-dioxetane for 3d compared
(18) Richardson, W. H.; Thomson, S. A. J . Org. Chem. 1985, 50,
1803-1810.
J . Org. Chem, Vol. 68, No. 15, 2003 5953

Motoyoshiya et al.
easy phospha-1,2-dioxetane ring formation results in an
increase in the rate of the light emission. On the other
hand, this effect in the HWE reaction provides diversity
in the geometric selectivity.
Exp er im en ta l Section
Gen er a l Meth od s. When required, all solvents were puri-
fied by standard techniques. Column chromatography was
performed on silica gel. All NMR spectra were recorded in
CDCl3. The chemiluminescence quantum yields were mea-
sured using a photomultiplier tube connected to a photon-
counting system (vide infra).
9-Dim eth ylp h osp h on o-10-h yd r oa cr id a n e (3a ). A mix-
ture of N-hydroacridinium bromide (2.00 g, 7.68 mmol) and
trimethyl phosphite (1.81 mL, 15.4 mmol) was stirred for 2 h
at room temperature. The product was collected, washed with
petroleum ether, and recrystallized from benzene to give white
crystals (0.80 g, 53%): mp 166-168 °C; 1H NMR δ 3.54 (d,
6H, J PH )11 Hz), 4.58 (d, 1H, J PH ) 25 Hz), 6.67-7.22 (m,
8H); 13C NMR δ 43.03 (d, J PC ) 140 Hz), 53.59 (d, J PC ) 8
Hz), 114.03, 114.40, 120.77, 128.27, 130.08, 140.28; 31P NMR
δ 24.81; MS (m/z) 289 [M+] (22.6), 180 (100). Anal. Calcd for
F IGURE 3. Linear relationship in double-reciprocal plot of
1/[fluorophores] vs 1/ΦCL.
C
15H16NO3P: C, 62.28; H, 5.58; N, 4.84. Found: C, 62.40; H,
5.57; N, 4.65.
9-Dieth ylp h osp h on o-10-h yd r oa cr id a n e (3b). A mixture
of N-hydroacridinium bromide (5.00 g, 19.2 mmol) and triethyl
phosphite (6.60 mL, 38.40 mmol) was heated for 10 min at 80
°C. The product was collected, washed with petroleum ether
and recrystallized from benzene to yield a white powder (3.42
estimation and quantitative comparison of the quantum
yields to argue the spin selectivity are therefore not as
meaningful. Nevertheless, Figure 3 would be instructive
to determine the excited species roughly. No enhance-
ment by DPA in the chemiluminescence reaction of 10
suggests the predominant formation of triplet benzophe-
none, which is in agreement with the thermal decomposi-
tion of 3,3-diphenyl-4-methoxy-1,2-dioxetane.21 On the
other hand, an almost equal measure of DBA enhance-
ment in the reaction of 10 and 11 is contrary to the
finding of the preferential formation of the singlet excited
fluorenone in the thermal decomposition of 3,3-(9′,9′-
fluorenyl)-4-methoxy-1,2-dioxetane.21 This is probably
due to an entirely different type of the counterparts
formed in the reactions, namely, 1,2-dioxetanes produce
two carbonyl compounds,22 whereas the present reaction
forms carbonyl compounds and phosphate ions. These
results show that chemiluminescence is not only a
significant reaction for the acridanyl phosphonates but
a general event in the autoxidation of the phosphonate
carbanions if the reaction conditions are adjusted.
1
g, 70%): mp 190-192 °C; H NMR δ 1.16 (t, 6H, J ) 7 Hz),
3.82-3.90 (m, 4H), 4.54 (d, 1H, J PH ) 25 Hz), 6.64-7.24 (m,
8H); 13C NMR δ 16.75 (d, J PC ) 5 Hz), 43.79 (d, J PC ) 139
Hz), 63.17 (d, J PC ) 7 Hz), 114.28, 115.26, 120.97, 128.47,
130.59, 140.63; 31P NMR δ 22.45; MS (m/z) 317 [M+] (18.2),
180(100). Anal. Calcd for C17H20NO3P: C, 64.35; H, 6.35; N,
4.41. Found: C, 64.06; H, 6.46, N, 4.27.
9-Bis(2′,2′,2′-t r iflu or oet h yl)p h osp h on o-10-h yd r oa cr i-
d a n e (3c). A solution of N-hydroacridinium bromide (3.00 g,
11.5 mmol) and tris(2,2,2-trifluoroethyl) phosphite (5.68 g, 17.3
mmol) in dimethyl formamide (DMF) (10 mL) was heated for
12 h under reflux. Removal of DMF under reduced pressure,
addition of brine, extraction with chloroform, and removal of
the solvent after drying over anhydrous Na2SO4 gave the crude
product, which was purified by column chromatography (elu-
ant AcOEt/hexane ) 3/2) to afford white crystals (2.14 g,
1
44%): mp 147-150 °C; H NMR δ 3.96-4.03, 4.11-4.19 (m,
4H), 4.76 (d, 1H, J PH ) 25 Hz), 6.27 (s, 1H), 6.72-6.74, 6.92-
6.96, 7.16-7.26 (m, 8H); 13C NMR δ 43.28 (d, J PC ) 140 Hz),
62.61 (dq, J FC ) 38 Hz, J PC ) 7 Hz), 122.88 (dq, J FC ) 276 Hz,
J PC ) 8 Hz), 112.75, 114.35, 121.18, 129.03, 130.02, 140.70
(Ar C); 31P NMR δ 24.67; MS (m/z) 425 [M+] (8.0), 180 (100).
Anal. Calcd for C17H20NO3P: C, 64.35; H, 6.35; N, 4.41.
Found: C, 64.06; H, 6.46, N, 4.27.
Con clu sion
9-(4′,5′-Dim eth yl-2′-oxo-2′λ5-[1′,3′,2′]d ioxa p h osp h ola n -
2′-yl)-10-h yd r oa cr id a n e (3d ). A mixture of N-hydroacri-
dinium bromide (2.64 g, 10.1 mmol) and 4,5-dimethyl-2-
methoxy-1,3,2-dioxaphospholane (3.04 g, 20.3 mmol), prepared
from the reaction of methyl dichlorophosphite and 2,3-butane-
diol (mixture of stereoisomers) in the presence of triethylamine
in benzene, was stirred for 1.5 h at room temperature. The
product was collected, washed with petroleum ether, and
recrystallized from acetonitrile to give white crystals (2.37 g,
74%): mp 208-214 °C; 1H NMR δ 0.77, 0.93, 1.18, 1.19 (d,
6H, J PH ) 6 Hz), 2.96-3.04, 3.72-3.81, 4.05-4.12, 4.67-4.72
(m, 2H), 4.79, 4.84 (d, 1H, J PH ) 25 Hz), 6.69-6.72, 6.89-
6.93, 7.12-7.16, 7.28-7.34 (m, 8H), 8.55 (s, 1H); 13C NMR δ
14.75, 16.20, 17.18, 17.83, 42.84, 43.23 (d, J PC ) 129 Hz), 77.43,
78.32, 80.38, 83.32, 113.81, 121.06, 128.49, 130.33, 130.78,
140.30; 31P NMR δ 39.00, 39.26, 39.45 (mixture of stereoiso-
mers); MS (m/z) 315 [M+] (2.4), 180 (100). Anal. Calcd for
The results in the present study are indicative that
phospha-1,2-dioxetanes are the most likely intermedi-
ates, although the evidence is circumstantial. The sub-
stituent effect on chemiluminescence decay was under-
stood as a result of the different rates for the ring-closure
step in the oxy-Wittig type of the reactions; namely, the
(19) Kazakov, V. P.; Voloshin, A. I.; Yakovlev, V. N.; Maistrenko,
G. Ya.; Khusainova, I. A.; Shishlov, N. M. High Energy Chem. 1996,
30, 330-334.
(20) (a) Quinga, E. M. Y.; Mendenhall, G. D. J . Am. Chem. Soc. 1983,
105, 6520-6521. (b) A possible explanation for this energy transfer is
the formation of the higher excited states of the ketones.
(21) Baumstark, A. L.; Wilson, T.; Landis, M. E.; Bartlett, F. D.
Tetrahedron Lett. 1976, 28, 2397-2400.
(22) Richardson, W. H.; Thomson. S. A. J . Org. Chem. 1985, 50,
1803-1810.
5954 J . Org. Chem., Vol. 68, No. 15, 2003

Chemiluminescence of Phosphonate Carbanions
C17H18NO3P: C, 64.76; H, 5.75; N, 4.44. Found: C, 64.54; H,
5.76; N, 4.32.
(29), 206 (51). Anal. Calcd for C15H14NO3P: C, 62.72; H, 4.88;
N, 4.88. Found: C, 62.60; H, 4.91; N, 5.27.
9-(5′,5′-Dim eth yl-2′-oxo-2′λ5-[1′,3′,2′]dioxaph osph or in an -
2′-yl)-10-h yd r oa cr id a n e (3e). This was prepared similarly
to the above procedure using N-hydroacridinium bromide (2.00
g, 7.68 mmol) and 5,5-dimethyl-2-methoxy-1,3,2-dioxaphos-
phorinane (1.89 g, 11.5 mmol), obtained from the reaction of
methyl dichlorophosphite and 2,2-dimethyl-1,3-propnediol in
the presence of triethylamine in benzene. The product was
recrystallized from acetonitrile (1.68 g, 67%): mp 187-188 °C;
1H NMR δ 0.53, 0.93 (s, 6H), 3.48-3.53, 3.92-3.97 (m, 4H),
4.70 (d, 1H, J PH ) 26 Hz), 6.69-6.71, 6.89-6.91, 7.10-7.14,
7.29-7.31 (m, 8H); 13C NMR δ 18.83, 19.73 (J PC ) 7 Hz), 41.36
(d, J PC ) 137 Hz), 74.51 (d, J PC ) 8 Hz), 112.05, 112.51, 119.13,
126.54, 128.57, 138.34; 31P NMR δ 16.71; MS (m/z) 329 [M+]
(2.5), 59 (100). Anal. Calcd for C18H20NO3P: C, 65.65; H, 6.12;
N, 4.25. Found: C, 65.72; H, 6.08; N, 4.07.
9-Dim eth ylp h osp h on o-10-m eth yla cr id a n e (4a ), 9-d i-
eth ylp h osp h on o-10-m eth yla cr id a n e (4b), 9-bis(2′,2′,2′-
t r iflu or oet h yl)p h osp h on o-10-m et h yla cr id a n e (4c), and
9-d ip h en ylp h osp h on o-10-m eth yla cr id a n e (4f) are known
compounds and were prepared by the established procedure.4b
9-(4′,5′-Dim eth yl-2′-oxo-2′λ5-[1′,3′,2′]d ioxa p h osp h ola n -
2′-yl)-10-m eth yla cr id a n e (4d ). This was prepared similarly
to the above procedure using N-methylacridinium iodide (1.60
g, 5.00 mmol) and 4,5-dimethyl-2-methoxy-1,3,2-dioxaphos-
pholane (1.60 g, 5.00 mmol, mixture of stereoisomers). The
product was recrystallized from benzene/hexane (0.37 g,
49%): mp 201-202 °C; 1H NMR δ 0.66, 0.85, 1.17 (d, 6H, J )
6 Hz), 2.85-2.92, 4.02-4.08, 4.61-4.68 (m, 2H), 3.35 (s, 3H),
4.78 (d, 1H, J PH ) 22 Hz), 6.86-6.98, 7.22-7.35 (m, 8H); 13C
NMR δ 15.11, 17.49, 18.05, 33.54, 43.92, 45.20 (d, J PC ) 129
Hz), 80.41, 83.66, 112.89, 118.50, 121.28, 128.84, 130.40,
143.25; 31P NMR δ 39.49, 39.71, 40.42 (mixture of stereoiso-
mers); MS. m/z 329 [M+] (4.3), 180 (100). Anal. Calcd for
9-Dieth ylp h osp h on oa cr id in e (7b). This was prepared
similarly to the above procedure using 3b (1.72 g, 5.42 mmol)
and chloranil (1.60 g, 6.51 mmol). The crude product was
recrystallized from ether (0.70 g, 41%): mp 82-84 °C; 1H NMR
(60 MHz) δ 1.28 (t, 6H, J ) 7.2 Hz), 4.18 (dq, 4H, J ) 7.2 Hz,
J PH ) 14.4 Hz), 7.41-7.86 (m, 4H), 9.11-9.27 (m, 4H); MS
(m/z) 315 [M+] (100), 286 (29), 206 (51). Anal. Calcd for C17H18
-
NO3P: C, 64.76; H, 5.75; N, 4.44. Found: C, 64.64; H, 5.66;
N, 4.81.
Ca libr a tion of P h otom u ltip lier . It is known23 that au-
toxidation of a luminol solution having an absorbance of 1.0
produces photons of 1.62 × 10-6 mol/L in the reaction with
excess t-BuOK in t-BuOH; therefore, the correction value (G)
for the photomultiplier is defined by
Vlumi
G ) (1.62 × 10-6)Abslumi
Nlumi
where Abslumi is the absorbance, Vlumi is the volume of the
luminol solution, and Nlumi is the count number determined
by an average of at least five measurements. Because a luminol
solution of a certain concentration suitable for the capacity of
the photomultiplier used in this study was required, the
absorbance at 359.5 nm of a stock luminol solution (1.6 × 10-7
M) in distilled DMSO was determined to be 1.22 × 10-3 from
a calibration curve of the luminol solution. To this luminol
solution (2.0 mL) in a quartz cell was added t-BuOK solution
(0.3 mL, 2.0 × 10-2 M) in t-BuOH, and the photons generated
immediately were counted. The G value was calculated ac-
cording to the above equation.
Mea su r em en t of CL Qu a n tu m Yield s. The total CL
quantum yields (ΦCL) were calculated according to
Rphoto
ΦCL) GN
C
18H20NO3P: C, 65.65; H, 6.12; N, 4.25. Found: C, 65.63; H,
(
)
M
6.31; N, 4.48.
9-(5′,5′-Dim eth yl-2′-oxo-2′λ5-[1′,3′,2′]d ioxa p h osp h or a n -
2′-yl)-10-m eth yla cr id a n e (4e). This was prepared similarly
to the above procedure using N-methylacridinium iodide (1.12
g, 3.50 mmol) and 5,5-dimethyl-2-methoxy-1,3,2-dioxaphos-
phorinane (2.22 g, 12.5 mmol). The product was recrystallized
from benzene/hexane (0.90 g, 75%): mp 242-243 °C; 1H NMR
δ 0.52, 0.95 (s, 6H), 3.38 (s, 3H), 3.39-3.43, 3.89-3.94 (m, 4H),
4.67 (d, 1H, J PH ) 26 Hz), 6.87-6.98, 7.23-7.36 (m, 8H); 13C
NMR δ 21.06, 21.95, 32.60 (d, J PC ) 7 Hz), 33.58, 44.92 (d,
J PC ) 136 Hz), 76.71-76.99 (m), 112.77, 118.63, 121.31, 128.76,
130.37, 143,23; 31P NMR δ 16.20; MS (m/z) 343 [M+] (3.6), 194
(100). Anal. Calcd for C19H22NO3P: C, 66.46; H, 6.46; N, 4.08.
Found: C, 66.51; H, 6.48; N, 4.28.
where N is the average number of the counts, Rphoto the
correction factor given by the manufacturer for the fluores-
cence emission maximum of the fluorophores, and M is the
moles of phosphonoacridanes. For a typical run, a solution (1.0
mL) containing 3 or 4 (1.00 × 10-3 M) in distilled DMSO was
placed in a 1 × 1 cm2 quartz cuvette in front of the photomul-
tiplier in exactly the same geometry. Photon-counting was
initiated simultaneously with the injection of a solution (0.5
mL) of t-BuOK (2.00 × 10-1 M) in DMSO into the cuvette, and
the data collection was continued for 512 s. The total CL
quantum yields were calculated according to the above equa-
tion.
Ack n ow led gm en t. This work was partially sup-
ported by a Grant-in-Aid for the 21st Century COE
Research (14B1-30) by the Ministry of Education,
Culture, Sports, Science and Technology of J apan.
9-Dim et h ylp h osp h on oa cr id in e (7a ). A solution of 3a
(3.80 g, 13.1 mmol) and chloranil (3.88 g, 15.8 mmol) in
benzene (150 mL) was heated under reflux for 4 h. After
removal of the solvent, purification by column chromatography
(eluant hexane/AcOEt ) 6/1) gave the product, which was
recrystallized from chloroform/hexane (2.44 g, 65%): mp 164-
J O030046L
1
166; H NMR (60 MHz) δ 3.85 (d, 6H, J PH ) 11.4 Hz), 7.27-
(23) Lee, J .; Seliger, H. H. Photochem. Photobiol. 1965, 4, 1015-
1048.
7.94 (m, 4H), 8.18-8.40 (m, 4H); MS (m/z) 315 [M+] (100), 286
J . Org. Chem, Vol. 68, No. 15, 2003 5955