64
M. Ferreira et al. / Catalysis Communications 63 (2015) 62–65
β-CD was also without notable effect on the melting point (compare en-
tries 2, 5, 9 and 10). So, the modification of the size or the chemical mod-
ification of CDs had no major impact on the melting points. By contrast,
the chemical modification of CD seemed to be a crucial factor for the vis-
cosity since the values decreased when CD was hydroxypropylated
polar ligand is well soluble in all the LMMs used (up to 3% w/w). The re-
action was performed at 90 °C and stopped after 1 h (Table 2).
The value of conversion varied from 41% for the mixture DMU/
CRYSME-β-CD (30/70) to 99% for the mixture DMU/RAME-β-CD (70/
30). The potential effect of cavity size was investigated by performing
six experiments with the three natives CD (α-CD, β-CD and γ-CD; en-
tries 1, 2 and 3) and the three hydroxypropylated CD (HP-α-CD, HP-
β-CD and HP-γ-CD; entries 4, 5 and 6). For the native series, the conver-
sions depended on the cavity size since the best results were obtained in
the case of β-CD and γ-CD whereas in the case of HP series, the three
CDs allowed reaching similar conversions (between 75% and 79%). So,
the values of conversion did not follow the size of the cavity. The effect
of chemical modification was also explored in the case of the β-CD se-
ries. The conversions were increased in the case of LMMs based on
DMU and hydroxypropylated CDs compared to those based on DMU
and native CDs (compare entries 1, 2 and 3 with 4, 5 and 6, respective-
ly). The methylation of β-CD (CRYSME-β-CD and RAME-β-CD) also had
a beneficial effect and it can be noticed that a higher methylation degree
led to a better conversion (compare entries 2 with 7 and 8). The highest
catalytic activity was obtained in the case of DMU/RAME-β-CD (70/30)
(Table 1 — compare entries 1–3 with 4–6) [12]. This tendency was con-
firmed in the case of the methylated β-CDs. Indeed, the values of viscos-
ity were lower for the mixtures DMU/CRYSME-β-CD and DMU/RAME-
β-CD compared to DMU/β-CD (compare entries 9 and 10 with 2).
Concerning the size of CDs, the number of glucopyranose units did not
significantly modify the viscosities of the LMMs inside a series (for na-
tive series, compare entries 1–3 and for hydroxypropyl series, compare
entries 4–6). The influence of the CD amount was also explored by vary-
ing the ratios DMU/β-CD, DMU/CRYSME-β-CD and DMU/RAME-β-CD
from 80/20 to 20/80. For the ratios 80/20 and 20/80, no melt was ob-
served (entries 11, 14, 17, 18 and 21). A melt was observed for the ratios
DMU/CD comprised between 70/30 and 30/70 except for the mixture
DMU/β-CD (30/70). This exception could be due to the well-known
strong intramolecular hydrogen bonds observed for β-CD decreasing
the formation of hydrogen bonds with DMU [13]. The melting points de-
creased with the increasing CD amount (for β-CD, entries 2, 12; for
CRYSME-β-CD, entries 9, 15, 16; for RAME-β-CD, entries 10, 19, 20).
For example in the case of DMU/RAME-β-CD, the melting point was
equal to 90 °C, 71 °C and 50 °C for the mixtures 70/30, 50/50 and 30/
−
1
and was equal to 1980 h . The influence of the proportion of CD was
also studied in the case of LMMs based on DMU and β-CD, CRYSME-β-
CD or RAME-β-CD at DMU/CD ratios varying from 70/30 to 30/70. Unex-
pectedly, when the CD amount in the LMMs increased, the conversion
decreased. Indeed, one would think that an increase in CD amount
would lead to an increase in conversion. Indeed in aqueous biphasic
hydroformylation reaction, CDs are known to play the role of mass
transfer agents by forming inclusion complexes with hydrophobic sub-
strates at the interface [14–16]. As comparison, LMM based on DMU and
α-D-methylglucopyranoside (70/30) was used and a conversion of 66%
was reached. Although α-D-methylglucopyranose did not possess a cav-
ity, this conversion was equal to those obtained with β-CD or γ-CD,
demonstrating that the formation of inclusion complexes was not nec-
essary to obtain high catalytic activities.
7
0, respectively. The viscosities firstly decreased from the DMU/CD
ratio 70/30 to 50/50 and secondly increased from the ratio 50/50 to
0/70 (entries 9, 15, 16 and 10, 19, 20). As comparison, both measure-
3
ments were performed with a mixture of 70% of DMU and 30% of α-D-
methylglucopyranoside (α-D-MeGluPyr – CD building block – entry
2
2). The melting point was in the same order of magnitude than the
three natives CD whereas the viscosity was lower (compare entries 1–
3
and 22).
To evaluate the potentiality of these new solvents, the rhodium-
catalyzed hydroformylation reaction was selected as a model reaction
and 1-decene was chosen as a non-miscible substrate in order to obtain
a biphasic system. Tris(m-sulfonatophenyl)phosphine trisodium salt
In order to rationalize these results, the 1-decene conversion was
plotted versus the viscosity (Fig. 2). The broader trend showed a de-
crease in the conversion when the viscosity increased with nevertheless
(
TPPTS) was used as ligand to maintain rhodium in the LMMs. This
Table 2
Rhodium-catalyzed hydroformylation of 1-decene in the various LMMs .
a
Entry
Solvent
% weight of CD in the solvent
Conversion of 1-decene (%)b
Selectivity in aldehydes (%)b
l/bb,c
1
2
3
4
5
6
7
8
DMU/α-CD
DMU/β-CD
DMU/γ-CD
DMU/HP-α-CD
DMU/HP-β-CD
30
30
30
30
30
30
30
30
30
30
50
50
70
50
70
30
45
66
68
78
75
79
76
99
95
98
53
51
41
89
63
66
96
94
95
96
92
94
94
90
90
90
93
94
91
94
94
94
2.3
2.3
2.4
2.3
2.3
2.3
2.3
2.2
2.2
2.2
2.2
2.1
2.0
2.2
2.2
2.4
DMU/HP-γ-CD
DMU/CRYSME-β-CD
DMU/RAME-β-CD
DMU/RAME-β-CD
DMU/RAME-β-CD
DMU/β-CD
DMU/CRYSME-β-CD
DMU/CRYSME-β-CD
DMU/RAME-β-CD
DMU/RAME-β-CD
DMU/α-D-MeGluPyr
d
9
1
1
1
1
1
1
1
0e
1
2
3
4
5
f
6
a
Experimental conditions: [Rh(acac)(CO)
2
] = 21 μmol (1 eq.), TPPTS = 105 μmol (5 eq.), 1-decene = 42 mmol (2000 eq.), solvent (DMU/CD) = 6 g, 90 °C, 50 bar CO/H
2
(1/1),
1
500 rpm, reaction time = 1 h.
b
Conversion, selectivity and linear to branched aldehydes ratio (l/b) determined by 1H NMR and GC.
Branched aldehyde (b) is racemic.
c
d
e
f
Catalyst recycle test performed using the catalytic phase recovered from run 8.
Catalyst recycle test performed using the catalytic phase recovered from run 9.
α-D-methylglucopyranoside (α-D-MeGluPyr) was used as CD building block.