Therefore, the activity of Rh-In/C was due to the synergy
between Rh and In, as discussed later. In these tests, the reactor
was filled with 5-MPa H2 to maintain the metallic state of the
catalyst, although this reaction does not stoichiometrically
consume H2. Decreasing the H2 pressure decreased the selectiv-
ity toward amino alcohols and increased toward the by-products
(Figure S1, Supporting Information). A physical mixture of
Rh/C and In/C showed moderate activity, probably because of
the interaction between Rh metal and eluted In species (Entry 4).
Other additive metals were used for the amination of 1,2-PrD
with aqueous ammonia. Although addition of Ga, Sn, and Zn
afforded moderate activity, the promoting effect of these metals
was smaller than that of In (Entries 5-7). Catalysts with Ge or Bi
did not show the activity at all (Entries 8 and 9). To optimize
the In loading amount, the effect of In/Rh molar ratio on the
catalytic activity was investigated (Table S1). Even with small
amount of In (In/Rh = 0.1 and 0.2), the catalysts had about half
activity of the catalyst with In/Rh = 1. The catalytic activity
became higher with larger amount of In in the range In/Rh =
0.2-1. The catalyst with In/Rh = 2 had almost the same activity
as that with In/Rh = 1. The selectivity patterns were almost
independent of In amount when In/Rh > 0.2. We used the
catalyst with In/Rh = 1 in the following studies.
Recyclability of the Rh-In/C was examined (Figure S2,
Supporting Information). Larger amount (50 ¼ 200 mg) of
catalyst was used to decrease the percentage of catalyst loss
during the recovery process. The used catalyst was separated
from the reaction solution by centrifugation. Then, the recovered
wet catalyst was used for the next reaction without scale-down
of the system. The reaction rate was almost maintained even
after three-time uses.
We examined the dependence of the performance of Rh-In
catalysts on various carbon supports (Table S3, Supporting
Information). Carbon black (Vulcan XC-72) and three types of
activated carbons (FAC-10, Carboraffin, and Shirasagi M) were
tested as supporting materials. Rh-In supported on carbon black
showed much lower activity than those supported on activated
carbons (3.0% conv. versus 5.7-10.7%) and larger amount of
metal elution (Rh 9.5% versus 0.9-5.6%). It suggests that it is
difficult to stably keep the metal on the surface of carbon support
in the reaction solution containing large amount of ammonia.
The activity and amount of metal elution were changed with
different activated carbon supports. Rh-In/FAC-10 showed the
highest activity (10.7%) and small leaching amount of Rh and In
(0.93% and 1.2%, respectively).
Figure 1. Time course of amination of 1,2-PrD over Rh-In/C.
Reaction conditions: water 13.2 g, 1,2-PrD 0.76 g, NH3 aq 6.08 g (total
20 mL), Rh-In/C (Rh 5 wt %, In/Rh = 1) 50 mg, 453 K, H2 5 MPa (at
r.t.), 4-160 h.
Table 2. Catalytic performances in the amination of various C3
alcohols with ammonia over Rh-In/Ca
Conv.
/%
Entry
1
Reactant
Product (selectivity/%)
1-Propanol
3.1
Propylamine (83),
dipropylamine (17)
Isopropylamine (>98)
2-Amino-1-propanol (42),
1-amino-2-propanol (47),
dimethylpiperazines (10),
others (<1)
1-Propanol (26),
propylamine (64), others (9)
1-Propanol (5),
2
3
2-Propanol
6.0
10.7
1,2-PrDb
4
5
1,3-PrDb
1.9
1.7
Glycerol
2-propanol (10),
2-amino-1-propanol (15),
1-amino-2-propanol (50),
others (20)
aReaction conditions: substrate 10 mmol, NH3 0.1 mol, water
balance (total 20 mL), Rh-In/C (Rh 5 wt %, In/Rh = 1) 50 mg,
b
453 K, H2 5 MPa (at r.t.), 24 h. PrD: propanediol.
cause of the steric hindrance of isopropyl group. Rh-In/C
catalyzed the amination of 1,2-PrD with the highest conversion
(Entry 3). 1,3-Propanediol and glycerol, both of which have an
OH group at the C3 position, showed low reactivity (Entries 4
and 5). In addition, most products had fewer functional groups
than the substrate (propylamine and 1-propanol from 1,3-
propanediol; propanols and aminopropanols from glycerol).
These results indicated that hydrogenolysis of alcohols with OH
groups at C1 and C3 positions proceeded over Rh-In/C. The
reaction solution was basic; hence, the hydrogenolysis reaction
might proceed through dehydrogenation, dehydration, and then
hydrogenation, which mechanism has been proposed for the
hydrogenolysis of glycerol to 1,2-PrD under basic conditions.24
We characterized the Rh-In/C catalyst with XRD, TEM,
and XPS. Figure 2 shows the XRD pattern of Rh-In/C reduced
with H2. The pattern showed the peaks at 2ª = 39.8, 72.6, 57.8,
and 86.0°, as well as the peak of carbon (ca. 20°). The peak
positions were different from those due to fcc Rh metal
(2ª = 41.08° for (111) reflection). These peaks suggested the
The time course of the amination of 1,2-PrD with aqueous
ammonia catalyzed by Rh-In/C (In/Rh = 1) is shown in
Figure 1. The initial total selectivity of amino alcohols was
94%. At longer reaction time, the selectivity to dimethylpiper-
azines increased and that to aminopropanols decreased. The yield
of total amino alcohols was 26% at 38% conversion at 160 h.
We applied the amination system to various C3 alcohols to
see the effects of the substrate structure (Table 2). The mono-ols,
1-propanol and 2-propanol, showed moderate reactivity (Entries
1 and 2). In the amination of 1-propanol, both the primary and
secondary amines were produced (Entry 1). 2-Propanol showed
higher conversion than 1-propanol (Entry 2). This is probably
because the amination of 2-propanol proceeds through the
ketone intermediate, which is more stable than the aldehyde one.
Secondary diisopropylamine was not produced, probably be-
© 2014 The Chemical Society of Japan | 823
Takanashi, Tsukasa
