Full Paper
at all. Further experiments revealed a radical intermediate for
the stepwise naphthalenes 5a–z. Therefore, we thermolyzed
the endoperoxides 5a–z in p-xylene (8), in which a well-stabi-
lized radical can be formed by hydrogen atom abstraction,
which, on the other hand, gives a stable and nonvolatile prod-
uct 9 (Scheme 7).
havior for the diverse substitution patterns from À113 up to
1
+50 JKÀ1 molÀ1 (Table 2) but still give high O2 yields (Table 1).
This seems to be contradictory at the first glance; however,
the thermodynamic properties were determined in pure
chloroform (spectroscopic grade), whereas the singlet oxygen
yields were determined according to literature by using NMR
experiments in [D]-chloroform with excess of tetramethylethy-
lene (TME) in a concentration of solvent regime to reduce loss
1
by physical deactivation.[23,39] This could influence the O2 yield
by substitution of solvent molecules in the solvent cage.[61]
TME could already develop an OÀO···p interaction[42] and thus
drive the mechanism towards a concerted pathway. Hence, the
values determined in this way must be understood as maxi-
1
mum O2 yields or transfer rates.
By using B3LYP/6-31G*, the calculated DS# values (T: 296 K,
scaling factor 1.0015,[62] normal pressure) range from +4.4 up
to +17.5 JKÀ1 molÀ1, only, and were calculated for a strictly sin-
glet (concerted) pathway. A stepwise mechanism was calculat-
ed for benzene endoperoxides under irradiation very recent-
ly.[63] The discrepancy between our calculated and the experi-
mental data underline again that also the thermolysis does not
proceed exclusively in a concerted manner.
Scheme 7. Thermolysis of endoperoxide 5 in p-xylene (8).
We were able to isolate the parental naphthalene 4a–z in
quantitative yield. From the literature and our own studies we
1
know that the naphthalene endoperoxides 5a–z yield O2 in
about 80%.[23] Furthermore, we isolated the hydroperoxide 9
in ꢁ25% yield, which is also formed by air autoxidation very
slowly.[56] Additionally, the oxidation can be accelerated by cat-
alysts[57] or through a radical pathway with cerium(IV) ammoni-
um nitrate.[58] Thus, it is clear that the formation of 9 can only
occur under radical conditions. From the quantity of p-xylene
for our experiments (100 mmol, 10.62 g) we isolated 1.4 mg
(13.2 mmol) of 9 before thermolysis (0.02%). The hydroperoxide
9 is neither formed under identical conditions without the
Conclusion
We have synthesized a broad number of new naphthalenes
and their corresponding endoperoxides, which we then isolat-
ed and fully characterized. The kinetics of the photooxygena-
tions and thermolysis have been investigated in detail and re-
vealed interesting influences of functional groups. We demon-
strated that our new endoperoxides are more suitable donors
1
presence of an endoperoxide, nor with O2 under irradiation,
as revealed in an additional competition experiment.
In summary, the naphthalene endoperoxides should thermo-
lyze mainly through a retro [4++2] cycloaddition as well as less
effectively through a stepwise radical mechanism. Furthermore,
the formation of the hydroperoxide 9 in this way seems to be
important as they are found in some applications,[59] but the
catalytic synthesis is difficult and results in a broad mixture of
products.[57] Finally, we determined the value of DG# (at 296 K).
1
for O2 because they exhibit prolonged half-lives and improved
storing properties. Additionally, longer half-lives reduce the
total amount of physical deactivation and increase the singlet
1
oxygen yield. Because O2 is constantly liberated over a longer
1
time period, slow acceptors can also be reacted with these O2
donors.
Moreover, we determined the thermodynamic properties
through experiments and calculations. We showed that the O2
cleavage does not proceed by a concerted pathway exclusive-
ly, but predominantly. As well as the the transition state of the
concerted pathway, we detected radical intermediates in the
stepwise reaction through trapping experiments.
The singlet oxygen donors shown herein should play an im-
portant role in “dark oxygenations” for light-sensitive accept-
ors, and also with some more modifications, for medical appli-
cations, which are currently under investigation in our labora-
tories.
With increasing positive DG#, the k value drops. More impor-
À1
tant is the value of DS#; from the literature it is known that
1
thermolysis reactions will release high amounts of O2 if the
DS# value is small or negative.[60] This requires a concerted
pathway.[21] A stepwise mechanism would lead to lower 1O2
yields along with a more positive DS# value,[23] earlier shown
for anthracenes in our group.[36]
A very interesting work about this topic deals with 1,4- and
9,10-anthracene endoperoxides. The first one features a low
DS# value together with high h and decays over a concerted
pathway, whereas the latter one reacts in the opposite
manner.[21]
Here, we found the first evidences for a partly concerted
and partly stepwise mechanism. The endoperoxides 5a–z liber-
ate oxygen in ꢁ80% in its singlet state and ꢁ20% in the trip-
let state proven by 1O2- and 3O2 scavengers, respectively. A
comparison of the experimental and calculated DS# shows an
interesting result. The experimental data, which were ensured
by at least four independent measurements, show different be-
Experimental Section
Determination of the photooxygenation kinetics (k1)
The exact mass of naphthalene 4 (7.811–25.572 mg, Metler Toledo
MX5, D=1 mg) was dissolved in chloroform (100 mL) to obtain an
accurate concentration of 510À4 m. This solution was diluted to
5·10À5 m and stored in the dark. After dissolving in chloroform
Chem. Eur. J. 2015, 21, 8569 – 8577
8575
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim