10.1002/adsc.201800058
Advanced Synthesis & Catalysis
The overall H-D exchange of the isohexides at
160 °C can be explained by two possible pathways:
1) keto-enol tautomerization after the
dehydrogenation step (Scheme 2) or alternatively 2)
by a C–H bond activation as proposed by Sajiki and
co-workers for alkanes.[13a] Considering the latter
route, the isomerization of the isohexides could also
proceed by C–H activation under these conditions
(Scheme 1, bottom). In this case the molecule is
bound with the oxygen of the hydroxy group to the
ruthenium surface which is known to be highly
oxophilic.[18] Then a heterolytic cleavage of the C–H
bond results in a prochiral carbenium ion and a
surface hydride.
1
Although also a homolytic cleavage of the C–H
bond is possible, working in the protic environment
of water, a heterolytic cleavage is more likely. The
carbenium ion can than rotate along the C–O bond as
well as the Ru–O bond. Thus, an inversion of the
stereo centre can be obtained providing that an
additional surface hydride is available for the
isomerization. This explains the need of additional
molecular hydrogen although it is overall not
consumed in the reaction. The dehydrogenation/re-
hydrogenation mechanism can occur in a similar way
but the ketone intermediate is formed which can
desorb from the catalyst’s surface as mentioned in the
literature[4b] in contrast to the direct C–H activation as
a concerted mechanism that would not result in a
stable intermediated that could desorb.
In order to elucidate whether the isomerization as
well as the H-D exchange proceeds solely through a
dehydrogenation/re-hydrogenation (for the deuterium
exchange with keto-enol tautomerization as
intermediate step) experiments with dimethyl
isosorbide (DMI) instead of the free sugar alcohol
were performed. First, the mere isomerization was
investigated. Thus, the reaction was performed in
water instead of deuterium oxide. Although one
would expect no isomerization to occur, a broad
product spectrum was obtained. By GC-MS
approximately half of the products could be identified
(Figure S7-S20). Besides residual DMI, the
monoethers were the main products and also one
isohexide, most likely isosorbide, could be detected.
Consequently, ether cleavage by hydrogenolysis
occurs to a large extend. Since methane was detected
in the gas phase via GC, the cleavage proceeds
through a hydrogenolysis of the C-O bonds.
Interestingly, three monoethers could be identified as
well. However, DMI itself can only be cleaved into
two different monoethers in case solely
hydrogenolysis occurs. Therefore, they have to be
formed partially by isomerization. It could not be
identified whether the isomerization happened at the
ether or the hydroxy moiety. Although the
isomerization at the hydroxy group is more likely, the
Figure 2. H NMR (close-up) of pure dimethyl isosorbide
(DMI) as well as of the product solutions after
isomerization at the indicated temperatures for 3 h. D2O
was used as the solvent and the deuterium source.
isomerization at the ether group cannot be ruled out.
Furthermore, a compound possessing the same molar
mass as DMI could be identified in the product
spectrum. A re-etherification after the cleavage by
hydrogenolysis that forms methane and subsequent
isomerization is highly unlikely. Hence, DMI itself
had to be isomerized as well. This cannot be
explained
by
a
mere
dehydrogenation/re-
hydrogenation mechanism due to the ether
functionalities that cannot be dehydrogenated. Thus, a
direct hydride exchange has to be involved. The
mechanism can be assumed in analogy to the
aforementioned C–H activation mechanism for the
isomerization of isohexides via a carbenium ion
(Scheme 3). Therefore, the general isomerization is
not only possible via dehydrogenation/re-
hydrogenation but also by a direct C–H activation at
temperatures of at least 160 °C.
In the literature almost no H-D exchange was
observed at carbon atoms bearing an ether
functionality using a Pd/C catalyst.[19] Since we
observed isomerization of DMI under the applied
conditions, also the H-D exchange was investigated
using DMI as substrate and Ru/C as catalyst. Again
without molecular hydrogen present only DMI could
be detected in the NMR as well as by HPLC
analytics. For the reactions with molecular hydrogen
80, 120 and 160 °C were screened as reaction
temperatures. While at the first two temperatures no
changes in the 1H NMR spectrum (Figure 2) could be
observed, basically all signals vanished at 160 °C.
Interestingly, since methoxy groups were cleaved
only partially by hydrogenolysis, also the signals for
the terminal methyl group at the ether moiety
vanished because of H-D exchange. Thus, with a
Ru/C catalyst, higher hydrogen pressure as well as
higher temperatures a full H-D exchange is possible
4
This article is protected by copyright. All rights reserved.