GLUCOSE HYDROGENATION TO SORBITOL
259
TABLE 1
After reaction, the content of Ni dissolved in the mixture
was analyzed by ICP. Only less than 1.0 ppm Ni was de-
termined in the reaction solution, indicating that the leach-
ing of the Ni-active sites during the glucose hydrogenation
could be neglected. The catalyst was used repetitively until
an abrupt decrease in the activity was observed. The num-
ber of repeat times (N) could reflect the lifetime of the
catalyst. No significant decrease in activity was observed in
the first 5 cycles of the hydrogenation, showing its excellent
durability in the present reaction. It was also found that the
deactivated catalyst could be easily regenerated by treating
it again in 6.0 M NaOH solution.
Structural Properties of the Different Catalysts
SBET Vpore
Sactive
Surface Ni atoms
per gram Ni
Sample
Composition (m2/g) (ml/g) (m2/g Ni)
Raney Ni-P
Regular Ni-P Ni88P12
Raney Ni Ni68Al32
Ni68Al25P7
87
1.2
106
0.050
0a
0.070
38
0.8
43
5.58 1020
1.20 1019
6.62 1020
a Too small to be detected.
ASAP 2010 (Micro-meritics). The active surface area was
measured by H2 chemisorption, from which the number
of surface active Ni atoms was also calculated assuming
H/Ni(s) = 1 and a surface area of 6.5 10
The catalytic behaviors of the as-prepared three Ni-
based catalysts are summarized in Table 2, from which
the following orders in the hydrogenation activities could
be obtained. (i) Hydrogen uptake rate per gram Ni(RHm ):
Raney Ni-P > Raney Ni regular Ni-P. (ii) Hydrogen u2p-
20
m2 per Ni
atom, based on an average of the areas for the (100), (110),
and (111) planes (12). These results are summarized in
Table 1. One can see that a considerable amount of Al was
still present in the Raney Ni-P catalyst after leaching for 6 h,
showing that the alkali leaching for Al is not complete un-
der the present condition. Similar to that of the Raney Ni,
pronounced pore volume was also determined in Raney
Ni-P, which accounted for its higher surface area (nearly
70 times) than that of regular Ni-P in which the pore vol-
ume was too small to be detected. In comparison with the
Raney Ni obtained at the same alkali leaching conditions,
the slightly lower BET surface area of the Raney Ni-P was
possibly due to the presence of the extra P species which
blocked some micropores in the catalyst. As can be seen
from Table 1, the pore volume of Raney Ni-P is also slightly
lower than that of Raney Ni catalyst.
take rate per meter square of active surface area (RHs ):
2
Raney Ni-P regular Ni-P > Raney Ni. (iii) TOF value:
=
Raney Ni-P regular Ni-P > Raney Ni.
=
According to our experimental results, the metallic Ni
species were proved to be the active sites for the glucose
hydrogenation since no significant activity was observed
when the catalyst was preoxidized in O2 flow and the high
activity could be recovered after it was reduced in H2 flow.
Therefore, the hydrogenation activity of the as-prepared
catalysts was mainly dependent on the the nature and the
surface dispersion of these Ni-active sites. Thus, the higher
hydrogenation rate per gram of Ni(RHm ) of Raney Ni-P than
2
of the regular Ni-P catalyst was mainly attributed to the
higher surface dispersion of Ni-active sites (i.e., the higher
active surface area or the larger number of the surface-
active Ni atoms). From Tables 1 and 2, one can see that
the ratio of surface Ni atoms per gram of Ni in the Raney
Ni-P and the regular Ni-P is 46.5, which is almost the same
Liquid phase hydrogenation of glucose was performed
at 393 K and 4.0 MPa in a stainless-steel autoclave, which
contained 1.0 g catalyst and 50 ml 50% (w/w) glucose aque-
ous solution. The stirring effect was carefully investigated
and a stirring rate of 1200 rpm was employed, which was
found to be sufficient to stir the reaction system so that the
higher stirring rate did not increase the reaction rate. The
initial hydrogenation activity was estimated by monitoring
the decrease of the pressure within the first 1.0-h period
which was then turned to be the H2 uptake rate per gram
Ni(RHm )or per meter square ofthe active surface area (RHs ).
as the ratio of their (RHm ) (47.7). However, since the ac-
2
tive surface area of Raney Ni-P is even lower than that of
Raney Ni catalyst, the different nature of Ni-active sites in
the Raney Ni-P from that in Raney Ni catalyst should be
considered in explaining the higher (RHm ) of that of Raney
2
Ni-P than that of Raney Ni. Since no significant electronic
interaction between metallic Ni and P in Ni-P amorphous
2
2
The reaction products were analyzed by a gas chromatog-
raphy equipped with a 25-m OV 101 capillary column and a
FID, in which the oven temperature was programmed at a
ramp of 4 K/min from 353 to 533 K. The catalyst was found
to be highly selective in the hydrogenation of glucose with
almost the exclusive formation of sorbitol. The sorbitol was
formed with selectivity up to 99.5% . The reaction conver-
sion was obtained by determining the remaining glucose in
the mixture after reaction for 6.0 h with Fehlings agent. In
order to exclude the effect of metal dispersion, the turnover
frequency (TOF) was also calculated according to the yield
TABLE 2
Hydrogenation Activities of Different Catalysts
RHm per weight
RHs per surface area Conv. TOF
2
2
1
Sample
(mmol/h g Ni)
(mmol/h m2 Ni)
(% )
(s )
Raney Ni-P
Regular Ni-P
Raney Ni
10.5
0.22
3.6
0.28
0.27
0.084
55.8
1.1
17.2
0.40
0.38
0.11
Note. Reaction conditions: 1.0 g catalyst, 50 ml 50% (w/w) glucose aque-
ous solution, T = 393 K, PH2 = 4.0 MPa, stirring rate = 1200 rpm, reaction
of sorbitol and the number of the surface-active Ni atoms. time = 6.0 h.