1972 J. Phys. Chem. B, Vol. 102, No. 11, 1998
Arai et al.
different ways. The activity of the latter was higher compared
with the former. The size of platinum particles may be
important to explain this result as well, for which the present
result is informative. In addition, we should consider the
oxidation state of platinum species and/or some metal-support
interactions. The reduction of platinum precursors was difficult
to complete and the adsorption of hydrogen was very weak in
the case of alumina support as indicated by TPR and TPD
measurements. So, it is likely that there are incompletely
δ+
reduced Pt species and/or there occurs the formation of an
alloy between platinum and aluminum. However, the roles of
these species other than metallic platinum are not clear at
present.
Cabello et al. have recently studied gas-phase hydrogenation
of acetonitrile on nickel-based catalysts prepared from hydro-
10
talcite-like precursors. They have reported that the hydroge-
nation of the nitrile to the imine takes place on nickel sites and
the rate of this process determines the overall rate of reaction.
Figure 8. Logarithmic plot of TOF
Figure 5, using the same symbol for all the catalysts different in the
amount of platinum loaded.
0
against D for the data given in
6
According to reaction scheme proposed by Verhaak et al.,
Cabello et al. assume that the subsequent hydrogenation occurs
on acid sites and nickel sites and the selectivity depends on the
nature of those sites. The state of reduction of nickel was shown
to be modified by addition of magnesium species. However,
this does not affect the adsorption of acetonitrile and hydrogen
as examined by the heats of adsorption, little influencing the
rate of reaction.
tion can leach out the impurities, and they can end up on the
surface of metal particles, modifying their catalytic activities.
Further work is needed to clarify the effects of the oxidation
state of platinum and the supports.
27
Similar correlations between the specific activity and the
degree of metal dispersion as observed in the present work are
In the present hydrogenation, triethylamine is the main
product irrespective of D. As mentioned above, the electronic
state of platinum particles depends on D, and it affects the
reactivity of the CN group of adsorbed acetonitrile. However,
this does not influence the selectivity, suggesting that the
hydrogenation of the nitrile to the imine is the rate-determining
step, as in the case of Cabello et al.,10 and the following
hydrogenation and condensation producing the tertiary amine
are faster steps. Since the silica gel used is an inert support, it
is less likely to play a role in the formation of triethylamine,
and it should also occur on the surface of platinum particles.
The role of support in the hydrogenation of nitriles has been
given different opinions, and it is still a matter of argument.
Volf and Pasek stated that the activity of supported nickel
catalysts was influenced by the nature of support but the
selectivity was not.1 This agrees with our previous observations
with nickel catalysts.8 In contrast, Verhaak et al. indicated the
significance of the acid-base properties of supports in deter-
11
reported in the literature. An example is the work of Boitiaux
et al., who studied the liquid-phase hydrogenation of unsaturated
hydrocarbons including 1-butyne, 1,3-butadiene, and isoprene
28
over supported palladium catalysts. It was observed that the
turnover frequency decreased with an increase in the palladium
dispersion. They explained that the low activities of very small
particles were due to the formation of a strong complex of these
hydrocarbons.
Structure Sensitivity. We should like to note a few aspects
concerning the structure-sensitivity of supported metal catalysts.
The results of Figure 5 are plotted in the logarithmic form in
Figure 8, and one can see a good linear correlation. Similar
correlations are reported for several catalytic reactions in the
11
literature. Farin and Avnir gave a phenomenological expres-
sion for those correlations and proposed a parameter to measure
29
the structure sensitivity. Recently we have proposed to use
the slope itself of such a linear correlation as a measure of the
30
degree of structure sensitivity. In the present case, we can
6
mining the activity and selectivity of supported nickel catalysts.
say that the reaction is negatively structure-sensitive and the
degree of structure sensitivity is 0.53 ( 0.13 (95% confidence
limits), which corresponds to a Farin and Avnir reaction
dimension of 2.53. In addition, we have noted that TOF at D
) 1 is more useful to examine the influence of support than
TOF at D < 1 because the metal is in the form of two-
dimensional islands at D ) 1 and the influence of support should
be the strongest in this extreme case. Those concepts have been
applied to a few catalytic reactions in a recent work and are
being used to examine the effects of various supports on the
gas-phase hydrogenation of nitriles as well. Further, it is
In a previous work, we found that diethylamine was produced
with a selectivity higher than that of triethylamine by gas-phase
9
hydrogenation of acetonitrile over supported platinum catalysts.
This is different from the present result that triethylamine is
selectively formed. The previous catalysts were prepared from
different metal precursor and support materials, platinum
tetraamine dichloride and Silbead-N silica gel including Al2O3
in 2 wt % (Mizusawa Industrial Chemicals, Ltd.), through ion-
exchange adsorption. The present results demonstrate that the
high selectivity of triethylamine does not change with the size
of metallic platinum particles. The previous catalysts should
demonstrated in a few cases that the structure sensitivity depends
δ+
31-33
have other Pt species due to the incompleteness of reduction
on reaction conditions.
We should also pay attention to
and/or the formation of an alloy with aluminum, as mentioned
above. This may influence the selectivity as well as the activity.
In addition, the acid/base properties of previous catalysts should
be different, a probable reason being high pH value on the
loading of platinum precursors, and this would also influence
the selectivity. As pointed out by Nonneman et al., impurities
present in the supports would be also important; wet impregna-
this respect in our cases by using different reaction conditions
in future work.
In closing, we briefly deal with the deactivation during the
course of hydrogenation. As described above, a possible
explanation is that the main product, triethylamine, accumulates
on the surface of catalysts, covering the active sites to cause
the deactivation. The triethylamine is basic and its adsorption