Supported Bimetallic AuPd Nanoparticles as Catalyst for the Solvent-Free Selective Hydrogenation of Nitroarenes

Selective hydrogenation of nitrobenzene was carried out under solvent-free conditions using supported AuPd nanoparticles catalyst, prepared by modified impregnation method (MIm), as efficient catalyst. >99% yield of aniline (AN) was obtained after 15 hours at 90 °C, 3 bar H2 that can be used without any further purification or separation, therefore reducing cost and energy input. Supported AuPd nanoparticles catalyst, prepared by MIm, was found to be active and stable even after 4 recycle experiments whereas the same catalyst prepared by SIm deactivated during the recycle experiments. The most effective catalyst was tested for the chemoselective hydrogenation of 4-chloronitrobenzene (CNB) to 4-chloroaniline (CAN). The activation energy of CNB to CAN was found to be 25 kJ mol -1 , while that of CNB to AN was found to be 31 kJ mol -1 . Based on this, the yield of CAN was maximized (92%) by lowering the reaction temperature to 25 °C.


Introduction:
Aniline (AN) is an industrially very important intermediate for the production of dyes, agricultural chemicals, pharmaceuticals, polymers, etc. [1][2][3] About 85% of aniline is produced via the catalytic hydrogenation of nitrobenzene (NB) using gaseous H 2 under either liquid or gas phase conditions as it is environmentally benign and water is the only by-product in this reaction. [4] Moreover nitration of aromatic compounds is a very well established and optimised technology that makes nitroarenes a readily available feedstock for the production of bulk and/or fine chemicals. [5,6] Solvents, such as ethanol, [1,[7][8][9][10][11][12] methanol, [13] 2-propanol, [14] have been used to facilitate heat and mass transfer during liquid-phase reaction conditions. However, after the end of the reaction additional separation steps are always employed to separate the product, resulting in not only high cost and energy input but also higher E-factor of this process. [15] One way to circumvent this problem is to perform the reaction under solvent-free conditions, a green and atom-economical process. Sun and co-workers [4] developed an ultrafine Pt nanoparticles supported on multi-walled carbon nanotube catalyst for the "solvent-free" hydrogenation of nitrobenzene, however the target product AN was used as a solvent for this hydrogenation reaction. Recently Huang and co-workers [16] reported monometallic Pd nanoclusters supported on N-doped ordered mesoporous carbon catalyst for the solvent-free hydrogenation of NB. This monometallic catalyst has been reported to display good activity and selectivity for the reduction of various substituted nitro aromatics to their corresponding amines.
However, the synthesis of this catalyst is complicated, which makes it challenging for is produced by the reduction of chloro nitrobenzene, which in turn is produced by the nitration of chlorobenzene. However, the hydrogenation of chloronitrobenzene (CNB) to chloroaniline is a challenging task as the weak carbon-halogen bond is highly susceptible for cleavage, resulting in the formation of AN thus lowering the selectivity of CAN. [24,25] Fine tuning of Pd-, [26,27] Pt-, [25,28] and Ni-based [29] catalysts through nanostructure modification, selective poisoning and other strategies have been reported. However, for a given catalyst, optimization of reaction condition to increase the selectivity of CAN has not been reported yet. Based on these two themes, the aim of the present work is to develop an efficient and robust catalyst for the solvent-free hydrogenation of nitrobenzene and to investigate this system in detail through kinetic studies and activation energy calculation. The best catalyst will be utilized for the chemoselective hydrogenation of CNB to CAN and optimize the reaction condition to increase the selectivity of CAN.

Catalyst preparation
The catalysts were prepared by modified impregnation, conventional impregnation and sol-immobilization methods, as reported by our group previously elsewhere. [30,31] The total metal loading used here was 1 wt.%. For bimetallic AuPd catalyst, the weight percentages of Au and Pd were both 0.5 wt%.
Modified Impregnation (M Im ): Gold precursor solution was prepared by dissolving with deionized water (2 L) and dried in an oven at 120 °C overnight.
Scanning Transmission Electron Microscopic characterization of 1%AuPd/TiO 2 catalyst prepared by C Im , S Im and M Im was carried out using a JEOL JEM-2200FS aberration-corrected electron microscope at Lehigh University.

Catalytic testing
Solvent-free hydrogenation of nitrobenzene: nitrobenzene (78 mmol, ≥ 99.0%, Sigma Aldrich) and catalyst (0.1 g) were placed in a Colaver® glass reactor. The molar ratio of substrate/metal was 10,778. Before starting the reaction, the reactor was purged with N 2 first, followed by pressurizing with H 2 for five times. The reactor was then charged with 3 bar H 2 under isobaric condition and heated to 90 °C in an oil bath with a magnetic stirrer bar to start the reaction. After a selected reaction time, the reactor was cooled down to <5 o C using an ice bath, followed by depressurisation. The reaction mixture was then taken out, centrifuged to remove the solid catalyst and the activity for the first reuse. The catalyst in batch B1 was regenerated. The procedure was repeated until the 4 th reuse of the catalyst.
Kinetic study: For kinetic studies, ethanol (16 mL, ≥99.8%, Sigma Aldrich) was used as the solvent for the hydrogenation reaction. Nitrobenzene (10 mmol) and catalyst (12.5 mg) were used for this test. Reaction temperature was selected as 40, 50 and 60 °C. The conversion was kept below 20% to study the initial rate of the reaction.
Other reaction conditions and operation procedures were the same with the solvent-free reaction.
Chemoselective hydrogenation of 4-chloronitrobenzene: ethanol (16 mL) was used as the solvent for this hydrogenation reaction. 4-chloronitrobenzene (10 mmol, 99%, Sigma Aldrich) and catalyst (12.5 mg) were used. Reaction temperature was selected as 20, 40, 50 and 60 °C. Other reaction conditions and operation procedures were the same with the solvent-free reaction hydrogenation of nitrobenzene.

Results and discussion
The solvent-free hydrogenation of nitrobenzene was performed using 1%AuPd/ In an effort to find the best metal, support and synthesis method combination we prepared different catalysts and tested them for the solvent-free hydrogenation of NB at 90 o C. We prepared monometallic 1%Au/TiO 2 (M Im ) and 1%Pd/TiO 2 (M Im ) catalysts and their catalytic activities were compared with the bimetallic 1%AuPd/TiO 2 (M Im ) catalyst ( Synthesis strategy of supported metal nanoparticles play crucial role in determining the structural parameters such as particle size, composition and nanostructure and hence their catalytic properties like activity, selectivity and stability. [30] As shown in entries 6 and 7 in Table 2, the AuPd/TiO 2 catalysts prepared by S Im and C Im showed similar NB conversions (39 and 38% respectively) and AN selectivities, while both were less than those of the M Im catalyst (54%). From the data presented in Table 2, it is clear that 1%AuPd/TiO 2 (M Im ) is the best catalyst for the solvent-free hydrogenation of NB to AN. Time on line evolution of products was studied using 1% AuPd/TiO 2 (M Im ) catalyst at 90 o C, and the result was shown in Figure 1a. After 15 h reaction, the entire NB was converted and the GC yield of aniline was ca. 99% (GC did not show any other peak).
To further confirm the formation and purity of aniline, we compared the 1 H-NMR spectra of crude reaction mixture after 15 h of the reaction, without any purification, with that of commercial aniline (≥99.5% purity from Sigma Aldrich) as shown in One of the crucial requirement of a heterogeneous catalyst to be used for industrial applications is its stability and reusability. Keeping that in mind, we studied the reusability of 1%AuPd/TiO 2 (M Im ) and 1%AuPd/TiO 2 (S Im ) catalysts ( Figure 2). For the 1%AuPd/TiO 2 (S Im ) catalyst, the yield of aniline decreased from 38% to 17% after 4 cycles. However, for the 1%AuPd/TiO 2 (M Im ) catalyst, the aniline yield changed only slightly from 53% to 51%, indicating that M Im catalyst is more stable and reusable than S Im catalyst. A similar observation was previously reported by our group for the solvent-free selective oxidation of benzyl alcohol to benzaldehyde. [34] During that study, we observed that the PVA, used in the synthesis of 1%AuPd/TiO 2 (S Im ), blocked the active sites of the catalyst. We believe a similar phenomenon might be present hindering the reusability of 1%AuPd/TiO 2 (S Im ) catalyst. On the contrary, 1%AuPd/TiO 2 (M Im ) does not contain any stabilising ligands, hence this catalyst is more stable and reusable. parameter (result not presented) to eliminate the effect of mass transfer. The initial rate of the reaction was calculated at 40, 50 and 60 °C and the results are presented in Figure S1. The apparent activation energy (Ea) was calculated to be 37 kJ mol -1 from the dependence of ln(k) on 1/T with R 2 >0.99 ( Figure S1). This value is similar to the values reported recently (37 kJ mol -1 over Pt/γ-Al 2 O 3 by Peureux et al. [35] 28±5, 33±5 and 45±5 kJ mol -1 over Ru/FeO x by Easterday et al. [14] 35±1 kJ mol -1 over Pd/C by Turáková et al. [36]) for this reaction. This activation energy value also gives evidence that the reaction was carried out under kinetic regime, otherwise the value should be in the range of 5-15 kJ mol -1 if the diffusion played an important role. [36] Chemoselective hydrogenation of 4-chloronitrobenzene The most active and stable catalyst (1%AuPd/TiO 2 (M Im )) was tested for the  As shown in Table 3 entries 1-3, the 1%AuPd/TiO 2 (M Im ) catalyst gave the best yield (85%) of CAN among all the catalysts tested. Using this catalyst, we then optimized the reaction conditions to increase the CAN yield. As shown in Figure 3, we performed the reactions at different temperatures at the initial stages (CNB conversion < 20%) and calculated the formation rates of 4-chloroaniline and aniline. Using these data, we calculated the apparent activation energies for CNB to CAN and CNB to AN reactions independently using the Arrhenius plot ( Figure S2) was achieved (Table 3, entry 4). This is one of the highest yields of CAN reported in the literature.  Pd-rich shell structure (Figure 4 (d)). While the particles produced by S Im (Figure 4 (b)) and M Im (Figure 4 (f)) appear to be random alloys without a distinctive core-shell morphology. Some 1 nm or sub-nm clusters can also be occasionally found in the catalyst prepared by the C Im method (inlet of Figure 4(d)). Those are Pd-rich clusters that are probably strongly bonded to the oxide support and therefore did not merge to form bigger particles. This microscopy results suggest that the combination of smaller particles, absence of stabilizer ligands (PVA) and random alloy morphology could be the reason behind the observed higher catalytic activity and stability of 1%AuPd/TiO 2 (M Im ) catalyst for the solvent free hydrogenation of NB to AN and the chemoselective hydrogenation of CNB to CAN.

Conclusions
In summary, we have successfully developed a 1%AuPd/TiO 2 (M Im ) catalyst for the solvent-free hydrogenation of nitrobenzene to aniline and chemoselective hydrogenation of 4-chloronitrobenzene to 4-aniline. Some important conclusions that we report in this article are (1) Through systematic catalyst screening, 1%AuPd/TiO 2 catalyst prepared by modified impregnation method is the most active and stable catalyst for the solvent-free hydrogenation of nitrobenzene to aniline. After 15 hours reaction, 99% yield of aniline was obtained at 90 °C. The activation energy of nitrobenzene reduction to aniline was calculated to be 37 kJ mol -1 .
Using kinetic studies, we found that the activation energy of CNB transformation to CAN was calculated to be 25 kJ mol -1 , lower than that of CNB to AN (31 kJ mol -1 ). Thus by decreasing the reaction temperature from 60 °C to 25 °C, a 92% yield of CAN was achieved.