D. S. Bhosale et al. / Tetrahedron: Asymmetry xxx (2015) xxx–xxx
3
the characteristic, very strong absorption band of asymmetric
valence vibrations of –COOÀ group at 1510 cmÀ1, which indicates
that the acid is present in the form of its salt.26 In this case, it is
the ammonium salt, which is supported by the presence of the
absorption band of asymmetric valence vibrations of –NH+3 at
3404 cmÀ1 and the pair of bands at 3329 cmÀ1 and 3200 cmÀ1 cor-
responding to the valence vibrations of group –NH2.26 This is the
reason as to why the spectra of Fe3O4ÁSiO2–(COOH)2 (Fig. 1B) and
Fe3O4ÁSiO2–(COO)2Cu (Fig. 1C) are almost identical, the presumed
exchange of –NH+3 ions for Cu2+ ions being almost indistinguishable
by means of IR. In the spectrum of the catalyst Fe3O4ÁSiO2–(COO)2-
CuL (Fig. 1D), there appeared a distinct band with wave number ca
1600 cmÀ1, which corresponds to the C@C valence vibration of 2-
functionalized pyridine.20,26 The ligand coordination is also sup-
ported by the distinct strenghtening of the bands about the wave
number of 2965 cmÀ1 belonging to C–H vibrations of aliphatic
bonds26 of the ligand molecule.
The textural characteristics of Fe3O4ÁSiO2–NH2 and Fe3O4ÁSiO2–
(COO)2CuL (surface area, micro-pore volume and total pore
volume) based on the determination of the nitrogen adsorption/
desorption isotherms are summarized in Table 1. In particular,
the changes in SBET illustrate a pronounced reduction of the surface
area, which was accessible for nitrogen molecules due to the
organic phase introduction. The particles Fe3O4ÁSiO2–(COO)2CuL
with the anchored catalyst exhibited a decrease in SBET by 46%
related to particles Fe3O4ÁSiO2–NH2. The change in pore volume
is even more dramatic; volumes decrease from 0.165 cm3/g
(Fe3O4ÁSiO2–NH2) to 0.018 cm3/g for the sample of catalyst
(Fe3O4ÁSiO2–(COO)2CuL). The volume of micro-pores is negligible
in both samples as is evidenced by values less than 0.002 cm3/g.
Interaction parameter C in BET equation is slightly decreasing with
increasing content of organic phase, which indicates a decreasing
wettability of surface by nitrogen.
Fe3O4ÁSiO2–(COO)2CuL was also tested in the synthesis of (R)-1-
(2,2-dimethyl-4H-benzo[d-1,3]dioxin-6-yl)-2-nitroethanol, which
is an intermediate in the synthesis of (R)-Salmeterol, a medical
drug used as a long-acting b2-adrenoreceptor agonist (trade name
SereventÒ).27 In this case, the obtained yield was 72% with 91% ee,
which was analogous to the results obtained with the homoge-
neous catalyst L/Cu(OAc)2 under comparable conditions28
(Scheme 2).
All of the reactions studied were performed by stirring the
dispersed system containing the corresponding aldehyde, nitro-
methane, ethanol and catalyst Fe3O4ÁSiO2–(COO)2CuL. After the
reaction, the catalyst was simply separated by means of an exter-
nal magnet and reused in the next reaction cycle (Fig. 3). From this
point of view, Sheldon’s filtration test was performed,29 which
involved the separation of the heterogeneous catalyst part way
through a reaction, followed by continuation of the reaction in
the absence of the immobilized catalyst. The result was negative,
with the reaction stopping completely after filtration.
The possibility of recycling the catalyst Fe3O4ÁSiO2–(COO)2CuL
was tested for the reaction of pivalaldehyde with nitromethane
(Fig. 4). Figure 4 shows that after 10 reaction cycles no significant
changes in the enantioselectivity occur (ꢀ94% ee). However, the
conversion exhibited a slow decrease from 88% down to 78%. The
slight lowering of the catalyst activity in the reaction of pivaloy-
laldehyde with nitromethane is documented in Figure 5, where
the conversion-time dependence is presented for the fresh catalyst
and for the ten times reused catalyst.
This decrease in catalyst activity cannot be explained by the loss
in the number of reaction centers, because neither the content of
the nitrogen nor the content of copper was decreased in the ten
times reused catalyst (3.66% N; 4.31% Cu). In the case of washing
out of the ligand itself (if a case of lower stability of the complex
would be presumed), a decrease in enantioselectivity would have
to be observed. Therefore, the decrease in activity can be explained
either by partial inhibition or a gradual loss of the catalyst due to
handling.21
The catalyst particles of Fe3O4ÁSiO2–(COO)2CuL were repre-
sented by means of scanning electron microscopy (SEM) (Fig. 2).
Figure 2 shows that the particles have spherical shapes and
agglomerate over the course of drying.
A
comparison of
The average hydrodynamic size of catalyst particles was deter-
mined (DLS) in the reaction mixture of nitromethane, 2-methoxy-
benzaldehyde and ethanol at 10 °C and at the concentrations of
Fe3O4ÁSiO2–(COO)2CuL from 0.12 mg/mL to 12 mg/mL. The increas-
ing concentration of catalyst was connected with the increase in
the hydrodynamic size of particles: 115 34 nm (0.12 mg/mL),
450 39 nm (0.5 mg/mL), 593 27 nm (1.2 mg/mL), 774 46 nm
(5 mg/mL), and 834 92 nm (12 mg/mL). At catalyst concentra-
tions of below 0.12 mg/mL, the hydrodynamic size of the catalyst
particles did not decrease any further. These results show that a
reversible formation of the aggregates takes place in the reaction
medium, and the distribution of their dynamic size values is deter-
mined by the character of medium and by concentration of the
catalyst.
Figure 6 (Table 3) presents the conversion-time dependences
for the reaction of nitromethane with 2-methoxybenzaldehyde in
the presence of various amounts of the catalyst Fe3O4ÁSiO2–
(COO)2CuL, namely in the range of 0.12–12 mg/mL (0.05–5 mol
%). The obtained enantiomeric excess values were in the presence
of different amounts of the catalyst and were almost comparable
(81–87% ee). Within the above-mentioned concentration range,
the catalyst is well dispersed in the reaction medium and repre-
sents more or less aggregated nanoparticles of hydrodynamic size
values from ꢀ115 nm to ꢀ830 nm, depending on its concentration
(according to DLS, Table 3). It was also proved that the intensity of
stirring of reaction mixture does not affect the observed reaction
rate. The time dependence in Figure 6 shows that between the cat-
alyst concentrations 0.5–1.2 mg/mL, the effect is only negligible.
However, the increase in catalyst concentration above 1.2 mg/mL
results in a deceleration of the reaction. Such concentration
Figure 2A and B does not show any significant differences between
the morphologies of freshly prepared catalysts and after operation
of ten reaction cycles.
The synthesized catalyst Fe3O4ÁSiO2–(COO)2CuL was further
tested for the Henry reaction of various aldehydes with nitro-
methane in ethanol at 10 °C (Table 2).
Table 2 presents the values of conversion and enantiomeric
excess for the reactions of individual aldehydes in comparison
with the results previously published,21 where the catalyst was
ligand L combined with cupric acetate. The application of catalyst
Fe3O4ÁSiO2–(COO)2CuL gave the corresponding 2-nitroethanols
with an (R)-configuration in excess. Table 2 shows that the applica-
tion of a heterogeneous catalyst generally led to negligible or only
a mild decrease in the enantioselectivity. Only in the series of
reactions of 4-substituted benzaldehydes, did we observe a more
significant decrease in the enantioselectivity (Table 2, entries
3–5;
D
up to À24%). On the other hand, in the case of pentanal,
we observed an increase in enantioselectivity as compared with
the homogeneous catalyst (Table 2, entry 7;
D +7% ee). The catalyst
Table 1
BET surface area (SBET in m2/g), volume of micro-pores (V in cm3/g) and total volume
l
of pores (Vtot in cm3/g determined at relative pressure 0.98) of the prepared particles
Sample
BET isotherm
V
(cm3/g)
Vtot (cm3/g)
l
SBET (m2/g)
C
Fe3O4ÁSiO2–NH2
22
10
66
52
0.0009
0.0001
0.165
0.018
Fe3O4ÁSiO2–(COO)2CuL