Вы находитесь на странице: 1из 8

a) Hydrogenation of nitrobenzene to aniline

Please look up the common types of commercial catalyst for the reaction you have chosen. Describe
the catalysts in terms of their form and structure

Aniline (ANL) is an important raw material for the polyurethaneindustry, being
used mainly in the production of methylenediphenyl diisocyanate (MDI) [1,2].
Commercial ANL is predomi-nantly produced by the catalytic hydrogenation of
nitrobenzene(NB)Eq. (1), which can be performed in gaseous or in liquid phase
C6H5NO2 + 3H2 C6H5NH2 + 2H2O
One of the advantages of using liquid phase hydrogenation isto avoid the hot-spots (as a consequence
of the reaction exother-micity); moreover, when compared with the vapor-phase and fora given
reactor size, it usually shows a greater production capac-
ity and allows using reaction heat to produce steam. Typically, NBliquid-phase hydrogenation is
operated at 80250C under pres-sure with yields of 9899%. In vapor-phase, although yields of
99%or higher could be achieved, it is operated under pressure at slightlyhigher temperatures

ity and allows using reaction heat to produce steam. Typically, NBliquid-phase hydrogenation is
operated at 80250C under pres-sure with yields of 9899%. In vapor-phase, although yields of
99%or higher could be achieved, it is operated under pressure at slightlyhigher temperatures [1,5].
Usually, for thehydrogenation of nitro groups, the most used metals are palladium,platinum and
nickel, supported or not
Catalyst selection is very important for maximizingANL selectivity, keeping secondary products
formation low.

However, upon consult-ing the list of commercial catalysts available in the market, it ispossible to
verify that the most common is palladium supportedon Al2O3, with the variations restricted to their
shape and metalcontent [21].

Industrially, Ni catalysts are also very used due totheir low cost and high yields

4.1.1 CO Chemisorptions Studies of Pd/C Catalysts

The carbon monoxide chernisorption experiments on the palladium/carbon catalysts
were carried out and the metal dispersions determined as described in Section The three catalysts showed significantly different average dispersions with
catalyst, Pd/CNI showing the lowest average dispersion (8.9 %), whereas catalyst
Pd/CSXU gave the highest average dispersion with a value of 42.0 % (Table 4.1).
Calculation of the average metal particle size showed, as expected, that catalyst
Pd/CN I had the largest metal particles and that catalyst Pd/CSXU had the smallest. Effect of Metal Particle Size on Specific Activity

A higher rate of reaction in methanol than in IPA was confirmed upon calculation of
the specific activity, or turnover frequency, for each catalyst, although, a more
complex situation was revealed (Section, Figure 4.22). In methanol, it was
demonstrated that catalyst Pd/CNI, with the lowest dispersion, and catalyst
Pd/CSXU, with the highest dispersion, were equally efficient catalysts and Pd/CA1
had the lowest activity. However, when IPA was used as the reaction solvent, a
different pattern in turnover frequency is observed. Catalyst Pd/CNI, with the
largest average particle size was the most efficient catalyst in terms of turnover
frequency, followed by catalyst Pd/CSXU and Pd/CAL This suggested that the
reaction is structure sensitive as defined by Boudart in 1969 [178]. It is surprising
that a particle size effect should be seen, as from the work by Hardeveld and Hartog
[179], the surface should be changing very little between 9 mm and 12 mm.
However subsequent work by Bond in the early 1990s [ 180,18 1 ], suggested that it
was possible for metal crystallites to show some degree of structure sensitivity over a
small diameter range, as the models proposed by Hardeveld and Hartog assumed the
existence of perfectly regular and symmetrical metal particles across the support
surface, which is not feasible under normal reaction conditions. Therefore, a much
wider variation in the coordination number and number of neighbouring atoms to a
surface metal atom was possible in imperfect crystallites, the factors considered
ultimately responsible for changes in specific activity. Despite this, with particle
sizes greater than 7 nm, the surface properties would still not be expected to change
dramatically; meaning catalysts Pd/CNI and Pd/CAI would be expected to show
near identical behaviour.
Earlier studies into the structure sensitivity of the nitrobenzene hydrogenation
reaction have provided mixed results. Caturan and coworkers [182] studied the
reaction in an n-octane solvent over palladium catalysts with a metal crystallite
ranging between 1.3 - 10.6 nin and found the specific activity to decrease as the
metal dispersion increased. This is contradictory to our experimental results where
in both solvents: methanol and IPA, catalyst Pd/CSXU with an average crystallite
diameter of 2.6 nm, displayed greater activity than Pd/CAl with an average
crystallite diameter of 8.7 nm. However Neri et al [183], examined the
hydrogenation of 2,4-dinitrotoluene over Pd/C catalyst in ethanol. The palladium
particles ranged between 2.6 - 8.4 nm. and it was found that the specific activity
increased as the palladium dispersion increased which is exactly as observed in our
results over this particle size range.
The effect of particle size on the nitrobenzene hydrogenation reaction has also been
studied over a variety of supported platinum catalyst, and like with palladium, a
range of conflicting observations has been reported. The group of Macias Perez [54]
reported the hydrogenation in ethanol using Pt/C to be structure insensitive over an
average platinum particle diameter of 1-7 nm. While Zhao, Ikushima and Arai [ 184],
also using Pt/C and ethanol as the solvent, found the specific activity to decrease as
the Pt dispersion increased and the average platinum crystallite size decreased from
19.6 nm to 7.2 nm.
The dependenceo f the nitrobenzener eaction on the metal particle size appearst o be
a highly complicated area that can be strongly influenced by the specific reaction
conditions including the choice of solvent, active metal and support. Therefore it is
difficult to draw a conclusion from our experimental results. Hydrogenation
reactions in general, are classed as structure sensitive reactions [185] and this does
appear to be the case in our observations. However, why the catalytic activity of
catalysts Pd/CNI and Pd/CAI is notably different and the activity of Pd/CAI
markedly lower than the other two are hard to explain. Therefore, it must also be
consideredth at the support could be influencing the reactivity. The Effect of Carbon Support on Specific Activity
Consultation of the manufacturers data sheets [186-188] for the three carbon support
materials revealed that the surface pH of each carbon was different depending upon
the chemical activation process. The surface pH of CNI was reported as between 5.5
and 8.0, the surface pH of CAI as between 2.0 and 3.5 and the surface pH of SX
Ultra as neutral. A measurement of the pH of a methanol solution directly above the
catalyst surfaces gave values of pH 6, pH 5.5 and pH 4.6 for catalysts Pd/CSXU,
Pd/CNI and Pd/CA1 respectively [189]. Comparison of the pH values with the
turnover frequency graph displayed in Figure 4.22, revealed that in both methanol
and IPA the catalyst displaying the lowest activity was the catalyst with the lowest
surface pH; catalyst Pd/CA 1. The surface pH of the other two catalysts, Pd/CN I and
Pd/CSXU, was of a similar value and close to neutral and, in methanol the specific
activity of these catalysts was identical. There have been numerous studies into the
effect of the surface properties of carbonaceous materials upon the catalysis by the
supported metal and pH has been shown to be of importance [190]. It has been
postulated that a lower pH is linked to higher oxygen functionality in the carbon
support, which can affect the overall catalytic properties [191-194]. Recent studies
[195] have revealed that carbons with a more acidic character have increased
interactions with cationic species in basic aqueous solutions and should lead to an
increase in catalytic activity. A similar situation is described by Mikam, et al with
the adsorption of nitrogen containing species to a Pd-Cu/C catalyst in aqueous
solution [196]. However, with our own hydrogenations, the reaction was carried out
in methanol and IPA at neutral pH and so any surface oxygen moieties present on the
support surface, such as -CO and-C02will exist in a largely unprotonated form. An
increase in these groups would lead to a build up of electronegative oxygen across
the support material and would be likely to repel the nitro group of nitrobenzene as it
approachesth e surface to adsorb. Therefore, an acidic support material (Nont CAI)
could be expected to reduce the specific activity of the nitrobenzene hydrogenation
reaction by making the adsorption of nitrobenzene less favorable. The other two
supports, with a near neutral surface pH would allow adsorption of nitrobenzene onto
the support more readily and may increase the hydrogenation reaction by facilitating
the adsorption of the reactant onto the metal active sites via a spillover mechanism.
Title: Commercial Catalysts Screening for Liquid Phase
Nitrobenzene Hydrogenation
Couto, C. S., Madeira, L. M., Nunes, C. P., & Arajo, P. (2016). Commercial catalysts
screening for liquid phase nitrobenzene hydrogenation. Applied Catalysis A:
General, 522, 152-164. doi:10.1016/j.apcata.2016.04.0323.2.1 Characterisation of Pd/C
(Couto, Madeira, Nunes, & Arajo, 2016, p. xx)

catalysts pores/surface, blocking porosity. total pore volume

Particle size distribution determined by HRTEM for the group I samples revealed
that in catalyst I.2 Pd particles are homogeneously dispersed in the support with a
particle size distribution in the range of 0.5-5.5 nm (Figure 2). In catalyst I.3 average Pd
particle size is slightly higher, being the highest one for catalyst I.1, which distribution
is not so homogeneous.

Figure 2 Palladium particles size distribution in fresh catalysts of the group I

(determined by HRTEM).
High-performance gas-liquid-solid microreactor with polydopamine
functionalized surface coated by Pd nanocatalyst for nitrobenzene

By comparing the FESEM images of the pristine PTFE and

PDA/PTFE, it can be seen that a uniform PDA layer has been formed on the PTFE
surface. In addition, as shown in Fig. 3c, after the deposition of Pd, Pd nanoparticles
with the mean particle size of about 50 nm were observed on the PDA surface.
Moreover, Pd nanoparticles were uniformly distributed over the surface of PDA layer.
Based on the FESEM results, the particle size distribution was also obtained. As
shown in Fig. 4a, the sizes of the discrete Pd nanoparticles were uniform and more than half of the
nanoparticles lied in the range of about 50 nm. The thickness of the
catalyst layer was also characterized in this work. Fig. 4b shows the cross-sectional
image of the Pd-PDA layer coated inner surface of the PTFE capillary tube. The
observation suggested that the thickness of the Pd-PDA layer was only about 70-80
nm. Such a thin layer is enough for the light transmission, allowing us to visualize the
two-phase flow patterns in the microreactor.

8, 21, 25, 29, 51

Surface area (BET)
After impregnation of activated carbon with palladium salt, the BET surface area and pore volume were found to be
decreasing (Table 1). It was due to the impregnant, which occupies the meso and micro pores.

Micropore volume(DR/MT)
Bulk Density
-Pd/C catalyst have bulk densities 0.40 g/mL. Bulk density of Pd/C catalyst was found to be higher than the active
carbon and acid treated carbon because pores in Pd/C catalyst were occupied by palladium.(Granular activated

- Hardness of 95 respectively %.

Average crystalline size

Morphology / Particle Distribution (SEM analysis )

Benavidez et al., 2014, p. 2

The image contrast in this highangle annular dark field (HAADF) mode is atomic number depend-ent (Z1.7),
hence the bright regions represent Pd particles whilethe less bright regions represent the support
The Pd particles range in size from 0.3 to 2.8 nm on all of the three supports.
The Pd particles on carbon are likely spherical
The Pd particles ranged from 0.5 to 1.0 nm in diameter.
The Pd particle sizes hows a broad size distribution with a mean diameter of 10 nm,
Carbon -274 m2/g

Mean particle diam-eter 0.7 nm Pd/C

(Srivastava et al., 2012, p. 3)

- Pd/C catalyst indicated Pd particles are present in a crystal state.
- A SEM image of active carbon showed considerable small cavities, cracks and attached fine
particles over the surface, forming a system of complicated pore networks After impregnation with palladium and
its subsequent reduction leads to the dispersal of palladium on the surface and within the pores of Pd/C catalyst
Show particles roughly in the spherical shape and homogeneous distribution on the carbon support.
Mean nano particle size - 4.0nm

The Pd loadings in all the Pd/C samples were 2.3%,( ICP-AES)

Finally, carbon nanotube sample (named CNT) possesses multiple walls with a hollow inner core and
close both their ends.
It was observed that all the samples have nearly the same type of isotherm which is characteristic
of mesoporous materials with cylindrical shaped porosities. In general, in open-ended CNFs or CNTs, pores can
be divided mainly into inner hollow cavities of small diameter (narrowly distributed, mainly
310 nm), external walls, and aggregated pores (widely distributed, 20100 nm) formed by interaction of isolated
CNFs or CNTs, the latter being much more important for adsorption issues.
Surface areas of oxidized CNT samples are larger (Table 1, Fig. S1) than those of the original samples, suggesting
that the acid treatment could remove some cups of tubes, initially not opened, and
increasing in this way their surface area.

In general, the Pd NPs are distributed uniformly on the modified carbon materials except for some
larger Pd NPs that are randomly distributed on PSPA.
Moreover they proposed that Pd NPs (supported on oxidized CNT) were coated with a carbon shell,
which could prevent the growth of Pd

supported metal nanoparticle catalysts,activity is largely dependent on availability of surface sites to par-ticipate in a desired reaction.

Smaller particles, having higher surface to volumeratio, have higher dispersion and more metal surface sites, whichgenerally translates to
higher activity per mass of metal.

Advantage of carbon coverage over Al2O3 as support for Ni/C-Al2O3
catalyst in vapour phase hydrogenation of nitrobenzene to aniline
XRD pattern of 5NCA reduced
catalyst witnesses' poorly crystalline nature of metallic Ni species.
It is noteworthy to mention here that the interaction between Ni
and the support plays an important role in the dispersion of metallic
Ni crystallites. Carbon, being a noninteracting support cannot disperse
Ni species even though it possesses high surface area (~960 m2 g1).
This is because of the fact that most of the pores of activated carbon
are in micro pore range which are not accessible for the dispersion of
Ni particles
On the other hand, -Al2O3 with a surface area of
~220 m2 g1 is capable of dispersing Ni particles because of its
interacting nature with NiO particles. It is surprise to see that carbon
coverage of alumina results in fine dispersion of Ni particles, even
though the non-interaction nature of carbonwith NiO exists. The nature
of carbon in carbon covered alumina (CCA) is different from that present
in the activated carbon. In CCA, mesoporous nature of carbon and
in activated carbon, microporous nature of carbon is present. That is
the reason why the porous nature and the interaction behaviour of
the support with NiO are important parameters for the dispersion of
Ni. Vissers et al. [13] have stated that in spite of severe pyrolysis conditions
during the preparation of carbon covered alumina, it is not possible
to cover completely the alumina surface with carbon. The acidic
centres of Al2O3 are covered by carbon.Moreover the amount of carbon
in C-Al2O obtained from CHNS analysis is not very high (~10 wt%).