R. Dumeunier, A. Huber / Tetrahedron Letters 55 (2014) 4410–4414
4413
Cl
O
SC(S)O(CH2)2Ph
CO2Me
DLP (0.09eq)
Cl
Cl
S
S
O(CH2)2Ph
Cl
Cl
+
+
O
( )
8
(
)
8
DCE (1M), 80° C
89%
Cl
Cl
S
S
18
Cl
Ph(CH2)2O(S)CS
DLP (0.09eq)
pMeO-Ph
Cl
O(CH2)2Ph
Si
Cl
Si
DCE (1M), 80° C
58%
pMeO-Ph
19
Figure 9.
sensitive to hydrolysis, already on TLC plates, and probably would
not stand a chromatography column). Straightforward substitution
with the commercially available xanthate salt 9 afforded 15 and
with the xanthate salt 2917 afforded solid, shelf-stable 16 and 17
with good yields after purification. For a cheaper and large scale
access, 25 can be prepared from the Zincke reaction of p-cresol
with CCl4 in the presence of AlCl3,18 and 26 can be prepared by
the abnormal Reimer–Tiemann reaction of p-cresol with chloro-
form and NaOH.19
The use of 5% of AcOH in converting 13 to 16 turned out to be
crucial for deciding the outcome of the reaction. Indeed, our first
attempt to convert 13 to 16 led unexpectedly to aromatized xan-
thate 30 in excellent yield (Fig. 6).
Even though the quantitative NMR of xanthate salt 29, prepared
as reported in the literature,17 showed excellent purity (>99%), we
could only explain this result by the catalytic decomposition of the
desired xanthate 16 by traces of a strong base, putatively residual
tBuOK or PhCH2CH2OK, coming from the preparation of 29. In
order to avoid this side reaction, we then ran the same substitution
in the presence of 5% of acetic acid and successfully got the desired
xanthate 16 with yields varying between 77% and 89%. With 16 in
hand, its instability towards catalytic amounts of a strong base
could be confirmed by running the catalytic decomposition on pur-
pose (Fig. 7).
To our great delight, the direct use of 15, 16 and 17 as surro-
gates of 12, 18 and 19 proved successful, and reacted as hoped with
unactivated olefins with acceptable yields (Table 1). As can be seen
from entry 6, the formation of a 6-membered ring from sabinene
supports our initial thoughts that the intermediate radical (formed
by addition of trichloromethyl radical to sabinene) is allowed
enough time to rearrange towards the thermodynamic product,
as sabinene has been known to be used exactly for this purpose
(under kinetic conditions, a five membered ring would have been
obtained).20
To answer whether the reaction goes via in situ intermediacy of
S-di/tri-chloromethyl xanthates such as 18 and 19, we wanted to
observe if those species are actually present during the reaction.
Cycle a was then performed independently by radical initiated
fragmentation-recombination (Fig. 8) in the absence of an olefin
to deliver 18 with an isolated yield of 74% after purification, and
from 17 to deliver dichloromethyl- 19 with 60% yield.
With the now available analytical details of both 18 and 19, we
can ascertain that none of 18 could be detected while monitoring
the reactions of 16 with olefins. Xanthate 18 is then not accumulat-
ing, and if formed at all, it would then be under steady-state con-
ditions, below our LC/MS detection thresholds. This is not true for
xanthate 19, which was seen accumulating, for example up to 50%
during the reaction of 17 with an olefin (entry 8, Table 1), before
being consumed in turn to the desired product. This makes close
to certain the fact that a significant part at least of 17 was chan-
nelled through cycles a then b when reacted with an olefin.
Without much surprise, cycle b works out also very well when
ran independently. Both 18 and 19 reacted directly with olefins
and delivered adducts faster, and with better yields than did their
surrogates 16 and 17 (Fig. 9).
Even though by these experiments we have shown that both
cycles a and b are efficient when ran independently, we still ignore
what fraction, if any, of the starting materials passes through cycle c.
Conclusion and perspectives
S-Trichloromethyl xanthate 18 and S-dichloromethyl xanthate
19 were prepared and isolated here for the first time. Both add effi-
ciently to olefins in a typical chain transfer reaction. But more
importantly, their isolation is not necessary as their respective pre-
cursors, 15, 16 and 17, are giving the same products when reacted
with olefins. Our objective to prepare trichloromethyl xanthate to
allow for intermediate radicals to rearrange was also demonstrated
to be efficient, as from the reaction of 16 with sabinene. From this
result, we are confident that 16 can be used to deliver thermody-
namic products of special interest to us for the preparation of bio-
logically active ingredients, or be used in radical-polar crossover
reactions to allow for intermediate cationic rearrangements. This
will be reported in due time.
Clarification that cycle c is operating would also be important.
Evaluating efficiency of cycle c could come for example by identi-
fying a xanthate carried by toluene, that would transfer RÅ effi-
ciently to an olefin whereas the direct RSC(@S)OR2 would not, for
example if RÅ is less stable than radical 23. If possible, and if we
can experimentally favour cycle c, consequences for xanthate
chemistry might be quite large.
Indeed, the novel aromatizing fragmentation concept described
in this Letter might break the propagation constraint, which is, that
in order to get the reversible propagation forward (as in Fig. 1), the
radical RÅ transferred to the olefin has to be at least as, and preferably
more stable than, intermediate radical 4 coming from its addition to
the olefin. As can be seen in Figure 4, the radical 22 formed by break-
ing of the CAS bond of 16 is not the one that is transferred to the ole-
fin. It is a relatively stable radical (secondary, bis-allylic) and we can
assume that it will be generated quite easily in most cases, but due to
the highly exergonic, aromatizing fragmentation to toluene, it might
produce in turn significantly less stable RÅ. If a couple of obvious
kinetic conditions are met, the propagation would still be brought
forward via cycle c, even with unstable RÅ. How much of this hypoth-
esis is false will be evaluated in a near future.
Acknowledgments
The authors would like to acknowledge Dr Myriem El Qacemi
and Dr Alain De Mesmaeker for insightful discussions.
References and notes