Organic Process Research & Development
Article
vacuum by hydrogen. The reaction mixture connected to the
balloon of hydrogen was vigorously stirred at room temperature
for 3 h, after which THF (50 mL) was added and the catalyst
recovered by filtration, rinsed with THF, and dried under
vacuum prior to reuse. The filtrate was concentrated to give a
crude product that was analyzed by ICP-OES for leaching of Pd
and Si. Conversion was assessed by TLC analysis (hexane/ethyl
acetate = 5:1) using potassium permanganate stain, by GC/MS
Hydrogenation of linolenic acid (cis C18:3, 3 (Figure 3))
involved consecutive saturation to cis C18:1 in 100%
conversion after 2 h and subsequent saturation of cis C18:1
to cis C18:0 after 3 more hours (entry 3).
Also the hydrogenation of eicosapentaenoic acid ethyl esther
(EPA, cis C20:5, 4 (Figure 4)) resulted in complete saturation
to cis C20:0 with 100% conversion after 3 h; similarly to the
hydrogenation of erucic acid (cis C22:1, 5 (Figure 5)), affording
complete saturation to cis C22:0 after 3 h. Again, isomerization
reactions of cis to trans isomers were not observed.
To explore the scope of the method, different vegetable oils
were then hydrogenated under the same optimized reaction
conditions of Table 1 in THF/MeOH (0.5 M) over 0.1 mol %
SiliaCat Pd(0). The results in Table 4 show that, in general,
complete substrate conversion of the vegetable oil to saturated
product was obtained after 3 h.
1
analysis, and by H NMR of isolated crude product.
3. RESULTS AND DISCUSSION
3.1. Solvent Effect. To study the influence of the solvent
on the catalytic process, a SiliaCat Pd(0) catalyst sample (0.05
mmol g−1 Pd load) was first tested in hydrogenation of oleic
acid (cis C18:1, 1 (Figure 1) in methanol, ethanol,
tetrahydrofuran, hexane, and ethyl acetate (Table 1).
3.3. Leaching and Catalyst Reusability. Leaching of Pd
from the catalyst was in each case tested by ICP-OES in the
reaction products. Solid Pd catalysts are notorious leachers12,
and we have recently investigated the leaching issue in
heterogeneous Pd catalysis.13 Here, catalysis was truly
heterogeneous, as shown by lack of reactivity of the reaction
filtrate obtained after 20 min of oleic acid conversion under the
optimized reaction conditions of Table 1. In general, the last
columns in Tables 1, 3, and 4 show that varying the reaction
conditions in the hydrogenation of oleic acid (Table 1) or the
FA (Table 3) or vegetable oil (Table 4) substrates resulted in
values of leached Pd in the isolated crude product that are <1
ppm.
The structure of the sol−gel nanocatalysts leads to such low
leaching of precious metal that the use of an ultralow amount of
catalyst (0.1 mol %) in an hydrogen balloon is enabled. Both
the latter factors indeed are known14 to ensure stabilisation of
the metal in its metallic or hydridic form, through higher
availability of H2 at the catalytically active Pd0 particles, thus
sustaining the catalytic activity of the entrapped palladium.
Accordingly, the reusability of the catalyst studied using
linolenic acid (C18:3) as substrate in five consecutive cycles
under the reaction conditions identified in Table 4 shows
(Table 5) full retention of the catalytic activity with minimal
leaching of Pd and Si. Complete substrate conversion in
saturated product (C18:0) with 96% selectivity was obtained
even after the fifth cycle.
The isolated yields in the entries 1 and 2 of Table 5 are lower
than the total conversion observed by GC/MS because during
isolation of the product a small amount of product is inevitably
lost. In entries 3 and 4 total conversion of the substrate is
observed, but a 1−3% of oleic acid nonselectively forms.
Similarly, in entry 5 the conversion of the substrate is 99% but
the yield by GC (MS) of final product is 95% (5% oleic acid),
so that the overall selectivity as defined in Table 5 is only 96%.
In conclusion, we have discovered that a modest 0.1 mol %
amount of SiliaCat Pd(0) selectively mediates the full
conversion of saturated vegetable oils at room temperature
under hydrogen balloon conditions. The catalyst is reusable
with negligible leaching of valued palladium. For comparison, at
the end of the 1990s, a state of the art 0.005% (wt %) Pd/Al2O3
catalyst for vegetable oils hydrogenation required a reaction
temperature of 102 °C in a continuous process,15 whereas a
more recent catalyst made of Pd nanoparticles entrapped in the
hexagonal porosity of SBA-15 mesoporous silica afforded
similar good results at 100 °C under 5 atm H2.16
Figure 1.
The best results were obtained in methanol (0.25 M
concentration with respect to the substrate) and in THF/
MeOH 5:1 by volume ratio (0.5 M concentration with respect
to the substrate) over 0.1 mol % catalyst (entries 3 and 11 in
Table 1). The hydrogenation reaction involved directly total
saturation of cis C18:1 to C18:0. No isomerization reaction of
cis C18:1 to trans C18:1 was observed.
3.2. Scope of the Method. To evaluate the catalyst
activity, the same substrate was hydrogenated with different
amounts of catalyst in the optimal solvent system (THF/
MeOH, 5:1 by volume ratio) and in MeOH 0.25 M under the
mild conditions of Table 1. Results in Table 2 show that, after
30 min of hydrogenation over 0.5−1 mol %, Pd proceeds with
full oleic acid conversion (entries 1, 3, and 4) whereas the
reaction time extends to 2−3 h over 0.1 mol % Pd depending
on the solvent (entries 2 and 6).
To further investigate the catalyst selectivity, hydrogenation
of different fatty acids containing different numbers of double
bonds was attempted. Reactions were conducted in methanol
0.25 M under the conditions of Table 1 over 0.1−0.2 mol %
SiliaCat Pd(0).
The results in Table 3 show that complete substrate
conversion into the corresponding saturated product was
obtained after 3−5 h. For instance (entry 1), in 3 h oleic
acid (cis C18:1) was completely converted to stearic acid (cis
C18:0) with no cis to trans isomerization reaction also in
methanol. Hydrogenation of linoleic acid (cis C18:2, 2 (Figure
2)) involved consecutive saturation steps: first, after 1 h, to cis
C18:1 in 90% yield, followed by subsequent saturation of cis
C18:1 to cis C18:0 after 3 more hours (entry 2 in Table 3).
1308
dx.doi.org/10.1021/op300115r | Org. Process Res. Dev. 2012, 16, 1307−1311