Angewandte
Chemie
isolated from the hydrogenation of LA (1608C, 10 MPa H2,
18 h, no additives) in an amount corresponding to 80% of the
combined mass of the employed metal precursor and ligand.
The 31P{1H} NMR spectroscopic data confirms the presence of
the triphos ligand in the typical tripodal facial coordination
mode at the ruthenium center (31P{1H}: triplet, d = 34.6 ppm,
1P; doublet, d = 27.0 ppm, 2P; JPP = 30 Hz).
The formation of a ruthenium hydride species is proven by
a singlet resonance at d = À7.35 ppm in the 1H{31P} NMR
spectrum. The intensity of the signal indicates the presence of
two terminal hydride ligands in the complex. It was possible
to fully characterize the complex by multinuclear NMR
spectroscopy as [Ru(triphos)(CO)(H)2].[22]
Although the homogeneous nature of the catalyst is
advantageous for molecular control of the reaction pathway,
it is often perceived as an inherent limitation for the
implementation of a production process. This can be over-
come, however, by integration of reaction and separation
steps into a continuous process that allows for inherent
recycling of the catalyst. The melting point of 31–338C and
high thermal stability of LA results in it being suitable as a
reaction medium for homogeneously catalyzed processes
without addition of any solvent. The solvent-free conditions
greatly facilitate the isolation and downstream processing of
the individual products. This has been demonstrated for
2-MTHF, which forms a two-phase mixture with the water
produced during the dehydration reactions if conversion and
selectivity are high enough.
recyclability of the catalyst system was demonstrated in
repetitive batch mode on a laboratory scale. The yields of
isolated 2-MTHF were 88, 79, 84, and 81% in four consec-
utive batches with the same catalyst additive mixture, thus
indicating that deactivation of the catalyst was negligible up
to this point.
The vapor stream from the flash distillation, which
consists of the products water, 2-MTHF, and the side product
pentanol, is condensed and fed into the decanter of a
heteroazeotropic distillation system. The 2-MTHF, which is
recovered in high purity at the bottom of the respective
column, contains pentanol as a side product and can be used
directly if the alcohol impurities can be tolerated in the
application. A single additional rectification column is
sufficient to obtain pure 2-MTHF, if needed. Exploiting the
same heat source for catalyst recycling and product recovery
can significantly reduce the energy demand of the process,
which is on the order of 3% of the energy content of the
recovered 2-MTHF. A preliminary economic evaluation
revealed that the production costs are dominated mainly by
the costs of the raw materials and catalyst.
This finding indicates that aspects of economy of scale are
not as important as for example in biomass-to-liquids (BTL)
concepts. Thus, relatively small production units could be
envisaged to address infrastructure limitations with biore-
newable feedstocks.
In contrast to LA, IA has a high melting point of 1838C
and tends to polymerize rapidly at higher reaction temper-
atures. As a consequence, the use of a suitable solvent is
necessary to carry out the selective transformation of this
substrate. In an initial screening of various types of reaction
media, cyclic ethers were found to be suited best for the
reaction sequence. This opens
Conceptual process design methods were used to develop
a flow sheet for a possible continuous production process
based on the optimized reaction conditions (Figure 1).[23] The
simulation of the separation was performed by using the
the possibility to use biomass-
derived ethers and, indeed, the
application of 2-MTHF in the
reduction of IA in the presence
of p-TsOH and NH4PF6 resulted
in the exclusive formation of
3-MTHF (Table 2, entry 5).
Thus, a process for the produc-
tion of 3-MTHF can be readily
envisaged, where the reaction
medium is provided by the prod-
uct itself. Only minor changes
Figure 1. Flow sheet for a possible LA to 2-MTHF process. The flexible design can be readily adapted to
the production of 3-MTHF from IA as described in the text.
involving
a
feedback loop,
which recycles part of the
MTHF product back into the
RADFRAC model in the ASPEN PLUS flow sheet simu-
lator.[24] Missing property data were estimated using group
contribution methods and CosmoRS calculations.[25] In this
concept, hydrogen and levulinic acid are heated, compressed
to reaction conditions, and fed into a plug flow reactor. The
homogeneous catalyst system is introduced with the liquid
substrate in the starting phase of the process. Under
continuous conditions the reactor effluent is partially vapor-
ized and the remaining liquid stream, which contains the
catalyst and additives as well as the high boiling compounds
GVL and PDO, is recycled into the reactor. The potential
substrate stream, would be necessary to adjust the process of
Figure 1 for such a “self-breeding” production of 3-MTHF
from IA. Hence, the scope of the molecular pathways is
matched by the flexibility of the process engineering
approach.
In summary, the results of the present study demonstrate
that the two biogenic platform chemicals levulinic acid (LA)
and itaconic acid (IA) can be converted into a diverse set of
isomeric lactones, diols, and cyclic ethers with high to
excellent yields by using a multifunctional catalyst system
comprised of a ruthenium/phosphine complex in combination
Angew. Chem. Int. Ed. 2010, 49, 5510 –5514
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5513