First evidence of the formation of 5,8-endoperoxide from the oxidation of 1,4-disubstituted naphthalene by singlet oxygen

Article abstract of DOI:10.1039/a705716d

Bulky water-soluble 1,4-disubstituted naphthalenes react with singlet oxygen giving the usual 1,4- and the unexpected 5,8-endoperoxides and showing that the regioselectivity of the [4 + 2] cycloaddition of singlet oxygen depends on the steric hindrance of the substrate.

Full text of DOI:10.1039/a705716d

View Article Online / Journal Homepage / Table of Contents for this issue
First evidence of the formation of 5,8-endoperoxide from the oxidation of
1,4-disubstituted naphthalene by singlet oxygen
Christel Pierlot and Jean-Marie Aubry*
Equipe de Recherches sur les Radicaux Libres et l’Oxyge`ne Singulet (URA CNRS 351) Faculte´ de Pharmacie, 3 rue du Professeur
Laguesse BP 83 F59 006 Lille, France
Bulky water-soluble 1,4-disubstituted naphthalenes react
with singlet oxygen giving the usual 1,4- and the unexpected
5,8-endoperoxides and showing that the regioselectivity of
the [4 + 2] cycloaddition of singlet oxygen depends on the
steric hindrance of the substrate.
spectrum suggested that no other products of oxidation with
lower symmetry, such as hydroperoxides, were formed.
1H and 13C NMR chemical shift assignments for 2b¶ were in
agreement with the spectral data of other 1,4-naphthalene
endoperoxides such as 2c.10 The structure of the endoperoxide
3b∑ was confirmed by comparison with 2b. Thus, 3b has four
quaternary carbons in the aromatic region, in positions 1, 4, 9
and 10 (d, J = 139.2 and 133.5 Hz) whereas 2b has only two
carbons of this type, C9 and C10 (d, J = 142.2 Hz). Another
telling difference was observed for the carbons supporting the
peroxide bridge, which were the tertiary carbons C5 and C8 (d,
J = 75.7 Hz) for 3b compared to the quaternary carbons C1 and
C4 (d, J = 83.4 Hz) for 2b.
In contrast to anthracene and higher members of the acene
series,1 naphthalene does not react with singlet oxygen (1O2) to
form the 1,4-endoperoxide.2 However, the naphthalene 1,4-en-
doperoxide 2a has been obtained indirectly in three steps from
1,6-imino[10]annulene.3 When electron-donating substituents
are introduced on the 1,4 positions of the naphthalene core,
1O2 readily reacts leading regiospecifically to 1,4-endoper-
oxides.4–6 5,8-Endoperoxides have been solely observed when
at least one additional methyl group was placed on the 6,77 or
8 positions8,9 of the naphthalene core. In the course of our work
on water-soluble carriers of 1O2,6,10 we have discovered that the
photooxidation of the bulky non-ionic water-soluble 1,4-di-
substituted naphthalene, N,NA,NB,NAAA-tetrakis(2,3-dihydroxy-
propyl)2,2A-(naphthalene-1,4-diyldimethyl)bis(malonamide)
1b (Scheme 1) leads to the simultaneous formation of the 1,4-
and 5,8-endoperoxides (2b and 3b). This example clearly shows
that in addition to electronic effects, the steric hindrance of the
substituents can modify the regioselectivity of the [4 + 2]
cycloaddition of singlet oxygen on polycyclic aromatic
hydrocarbons.
When the oxidation of 1b was run over 2 h in H2O–MeOH
(77:13) instead of deuterated solvents, 2b was formed with a 13
times lower yield (3% compared to 39%), and the concentration
of 3b became too small to be measurable. This lower yield was
1
expected since the lifetime of O2 is 15 times shorter in H2O
than in D2O. Chemical oxidation of 1b by a chemical source of
1O2 (H2O2–MoO422),10 also led to the formation of 2b and 3b,
confirming the involvement of singlet oxygen in the formation
of both endoperoxides (Table 1).
Warming of the solutions of pure 2b and 3b obtained by
semi-preparative HPLC indicated that both endoperoxides
decompose giving the starting naphthalene 1b and oxygen
(Scheme 1) according to first order kinetics (Table 2).
Comparison of the half-lives of decomposition (t1/2) clearly
showed the greater stability of 3b since no significant
decomposition occurred at 5 °C, and it was found to be ten times
more stable than 2b at 25 and 37 °C. On the contrary, 2b was
particularly thermolabile since its half-life at 37 °C (t1/2 = 7
min) is three times shorter than that of other 1,4-endoperoxides
such as 2c. This phenomenon seems to be general, in fact we
observed similar half-lives for other bulky 1,4-disubstituted
endoperoxides such as 2d (t1/2 = 5 min at 37 °C). The poor
stability of these 1,4-endoperoxides could result from the steric
hindrance induced by the bulky substituents R in the ‘butterfly’
structure 2, which would favour dissociation into the starting
naphthalene derivative and oxygen.
NMR and HPLC analysis of 1b†‡ after dye-sensitized
photooxygenation§ at 25 °C in D2O–CD3OD (77:13) showed
a 50% conversion in 2 h, and indicated that, beside the usual
1,4-endoperoxide 2b, the unexpected 5,8-endoperoxide 3b was
also formed, with a time-independent ratio equal to 78:22
(Table 1). All spectra showed that the malonamide side chains
remained unchanged during oxidation and the symmetry of each
R
R
R
O
O
O
O
8
1
9
1O2
7
6
2
3
+
heat
R
10
4
5
R
R
1
2
3
kr2
1
?
1 + O2 —
2
3
(1)
(2)
a R = H
b R = CH2CH[CONHCH2CH(OH)CH2(OH)]2
c R = CH2CH2CO2Na
d R = CH2CH(CO2Na)2
kr3
1
?
1 + O2 —
Scheme 1
Table 2 Half-lives (t1/2) of dissociation at 5, 25 and 37 °C for endoperoxides
2b, 3b, 2c and 2d, and rate constant (kr) of their formation at 25 °C
Table 1 Product distribution after photooxidation of 1b
Conversion Regioselectivity
t1/2/min
Solvent
T/°C t/h (%)
2b:3b
kA/104 m21 s21
5 °C
25 °C
37 °C
(D2O–CD3OD = 77:13)
D2O–CD3OD (77:13)
H2O–MeOH (77:13)
D2O
25
25
25
50
25
2
2
1
1
2
50
3
26
40
20
78:22
100:( < 1)
69:31
( < 1):100
80:20
2b
3b
2c
2d
540
30
300
113
18
7
70
23
5
3
1
142
44
stable
stable
208
CD3OD
Chem. Commun., 1997
2289

View Article Online
The rate constants, kr2 [eqn. (1)] and kr3 [eqn. (2)], for the
chemical quenching of 1O2 by 1b were estimated by comparison
oxide, which dissociates as it is forms whereas the 5,8-endoper-
oxide remains unchanged.
The authors are grateful to Dr A. Defoin for helpful
discussions.
1
with the well known O2 acceptor 1c11 (Table 2). Thus, they
were separately photooxidized in deuterated water at 25 °C
under the same conditions. The kinetics of the disappearance of
the starting materials indicated that 1b (kr2 + kr3 = 4 3 104 m21
s21) is about 13 times less reactive than 1c (krNDP = 142 3 104
m21 s21). The constant regioselectivity observed at this
temperature permits us to determine the contribution of each
Footnotes and References
* E-mail: jmaubry@phare.univ-lille2.fr
† High performance liquid chromatographic analyses (HPLC) were carried
out with Waters pumps Model 600 with a reversed-phase column (Nova
Pack RP 18, 25 cm) using 15 to 50% MeOH over 20 min at 0.7 ml/min and
UV detection at 200 nm.
‡ Synthesis of 1b: A solution of tetraethyl 2,2A-(naphthalene-1,4-diyldime-
thyl)bis(malonate) (ref. 11) (2 g, 4.2 mmol) and 3-aminopropane-1,2-diol
(5.6 g, 73.3 mmol) in MeOH (30 ml) was stirred overnight under reflux.
After evaporation of the solvent, acetone was added to the residue and the
resulting suspension was filtered. The white precipitate was recristallised
pathway: kr2 = 3 3 104 m21 s21 and kr3 = 1 3 104 m21 s21
.
The poor reactivity of 1b towards 1O2 could be explained by the
electron-withdrawing effect of the amide function, which
decreases the electronic density on the 1,4-positions of the
naphthalene core, compared to the slight electron-donating
effect of the sodium carboxylate functions of 1c. However,
since these groups are separated from the binding site by an
ethylene linkage, it appears more likely that the difference of
reactivity is mainly a consequence of the steric hindrance due to
the side chains.6
from MeOH (50%); mp
= 195 °C; dH (300 MHz, D2O) 8.10 (dd,
J5,6 = J5,7 = 3.15, 2 H, H5), 7.64 (dd, J5,6 = J5,8 = 3.15, 2 H, H6), 7.26 (s,
2 H, H2); dC (300 MHz, D2O) 174.3 (s, CO), 136.0 and 134.4 (s, C1, C9),
129.9, 129.3 and 127.1 (s, C2, C5, C6), 72.5 [d, CH(OH)], 65.9 [d,
CH2(OH)], 57.2 [CH(C)3], 44.4 (d, CH2N), 35.3 (s, Ar-CH2); m/z 675
(MNa+), 663 (MH+).
To strengthen this hypothesis, we oxidized the tetra-(sodium
carboxylate) 1d which is more bulky than 1c, but which exhibits
a greater electronic density on the naphthalene ring due to the
presence of four sodium carboxylate functions instead of two.
Actually, the reactivity of 1d towards 1O2 is three times lower
than that of 1c. This result confirms the predominating
contribution of steric hindrance over the electron-donating
effects of the side chains of 1b and 1d.
It is noteworthy that the formation of the 5,8-endoperoxide
3b is straightforward only when the oxidation of 1b is carried
out in deuterated water. Thus, in deuterated methanol, only 4%
is obtained after 150 min at 25 °C (compared to 33% in
deuterated water). This result seems astonishing if we refer to
the lifetime of 1O2, which is 3.5 times longer in CD3OD
(tD = 230 ms) than in D2O (tD = 65 ms). In fact, contrary to a
generally accepted principle, it was reported recently that the
rate constants of singlet oxygen [4 + 2] cycloadditions (kr) are
strongly solvent-dependent. In particular, it was shown that
highly structured solvents such as formamide and water
considerably increase the reactivity of conjugated dienes and
polycyclic aromatic derivatives.2,12 Thus, the finding that 1b
exhibits a 15-fold higher rate constant in D2O than in CH3OD is
just a further example of this accelerating effect.
§ Photooxidation of 1b into 2b and 3b. A solution of 1b (30 mg) and
Methylene Blue (1.6 3 1025 m in 1 ml of deuterated water was irradiated
with a sodium lamp (150 W) under continuous bubbling of oxygen at
constant temperature. During the reaction, some Methylene Blue was
periodically added to compensate for its fading. HPLC analysis showed the
ratio of each product.
∑ Selected data for dH (300 MHz, D2O) 7.38 (m, 4 H, H5, H6), 6.87 (s, 2 H,
H2); dC (300 MHz, D2O) 174.2 (CO), 142.2 (C9), 140.5 (C2), 130.3 (C6),
123.6 (C5), 83.4 (C1), 72.5 [CH(OH)], 65.6 [CH2(OH)], 51.1 [CH(C)3], 44.3
(CH2N), 31.9 (CH2C1).
** Selected data for 3b: dH (300 MHz, D2O) 7.15 (s, 2 H, H2), 7.06 (dd,
J5,6 = J6,8 = 3.15, 2 H, H6), 6.10 (dd, J4,5 = J5,7 = 3.15, 2 H, H5); dC (300
MHz, D2O) 173.6 (CO), 139.2 (C9), 138.1 (C6), 133.5 (C1), 131.5 (C2), 75.7
(C5), 72.5 [CH(OH)], 65.6 [CH2(OH)], 58.0 [CH(C)3], 44.3 (CH2N), 33.2
(CH2C1).
1 J. Rigaudy, Pure Appl. Chem., 1968, 16, 169.
2 J. M. Aubry, B. Cazin, M. Rougee and R. V. Bensasson, J. Am. Chem.
Soc., 1995, 117, 9159.
3 M. Scha¨ffer-Ridder, U. Brocker and E. Vogel, Angew. Chem., Int. Ed.
Engl., 1976, 15, 228.
4 H. H. Wassermann and D. L. Larsen, J. Chem. Soc., Chem. Commum.,
1972, 253.
5 H. H. Hart and A. Oku, J. Chem. Soc., Chem. Commun., 1972, 254.
6 C. Pierlot, S. Hajjam, C. Barthe´le´my and J. M. Aubry, J. Photochem.
Photobiol. B., 1996, 36, 31.
Finally, it appears that the formation of the unusual
5,8-endoperoxide by reaction of 1O2 with 1,4-disubstituted
naphthalene is possible only when a number of stringent
requirements are met: (i) substituents must be a little electron-
donating to increase the electronic density of the whole
naphthalene core; (ii) they must be very bulky to make difficult
7 C. J. M. Van den Heuvel, H. Steinberg and T. J. de Boer, Recl. Trav.
Chim. Pays-Bas, 1980, 99, 109.
8 W. Adam, E. M. Peters, K. Peters, M. Prein and H. G. Von Schnering,
J. Am. Chem. Soc., 1995, 117, 6686.
1
the approach of O2 to the crowded ring and to impair the
formation of the resulting butterfly structure; (iii) the peroxida-
tion must be carried out in D2O in order to benefit from the rate-
accelerating effect of water on the [4 + 2] cycloaddition of 1O2
and from the enhancement of 1O2 lifetime in deuterated
solvents; (iv) a gentle warming of the reaction medium
increases the ratio of 5,8- to 1,4-endoperoxides with oxidation
time because of the greater thermolability of the 1,4-endoper-
9 W. Adam and M. Prein, Acc. Chem. Res., 1996, 9, 1625.
10 J. M. Aubry, B. Cazin and F. Duprat, J. Org. Chem., 1989, 54, 726.
11 C. S. Marvel and B. D. Wilson, J. Org. Chem., 1958, 23, 1483.
12 B. Cazin, J. M. Aubry and J. M. Rigaudy, J. Chem. Soc., Chem. Comm.,
1986, 952.
Received in Liverpool, UK, 4th August 1997; 7/05716D
2290
Chem. Commun., 1997