Precious-Metal-Decorated Chromium(IV) Oxide Nanowires as Efficient Catalysts for 2,4-Toluenediamine Synthesis

The catalytic hydrogenation of 2,4-dinitrotoluene (DNT) to 2,4-toluenediamine (TDA) is a key step in the production of polyurethanes; therefore, the development of efficient hydrogenation catalysts for industrial use is of paramount importance. In the present study, chromium(IV) oxide nanowires were decorated by palladium and platinum nanoparticles in a one-step, simple, and fast preparation method to yield highly efficient hydrogenation catalysts for immediate use. The nanoparticles were deposited onto the surface of CrO2 nanowires by using ultrasonic cavitation and ethanol as a reduction agent. Beneficially, the catalyst became catalytically active right at the end of the preparation and no further treatment was necessary. The activity of the Pd- and Pt-decorated CrO2 catalysts were compared in the hydrogenation of 2,4-dinitrotoluene (DNT). Both catalysts have shown high activity in the hydrogenation tests. The DNT conversion exceeded 98% in both cases, whereas the 2,4-toluenediamine (TDA) yields were 99.7 n/n% and 98.8 n/n%, with the Pd/CrO2 and Pt/CrO2, respectively, at 333 K and 20 bar H2 pressure. In the case of the Pt/CrO2 catalyst, 304.08 mol of TDA formed with 1 mol Pt after 1 h hydrogenation. Activation energies were also calculated to be approximately 24 kJ∙mol−1. Besides their immediate applicability, our catalysts were well dispersible in the reaction medium (methanolic solution of DNT). Moreover, because of their magnetic behavior, the catalysts were easy to handle and remove from the reaction media by using a magnetic field.


Introduction
Polyurethanes (PUs) are one of the most versatile class of polymers with specific mechanical, physical, biological, and chemical properties giving them great potential for use in different applications, such as flexible and rigid foams, adhesives, paints, coatings, and elastomers. There is an ever-growing demand for toluene diisocyanate (TDI), which is one of the isocyanate components for making PUs; by 2018, global TDI production capacity reached 3.355 million tons per year. The main commercial route for the manufacture of TDI starts with the nitration of toluene using nitric acid to produce dinitrotoluene followed by catalytic hydrogenation to toluene diamine.

Preparation and Characterization of the Pd-and Pt-Decorated Chromium(IV) Oxide Nanowires
The preparation of the palladium-decorated magnetic nanowires is highlighted in Figure 1. The detailed procedure for both Pd and Pt is found in the experimental section. The specific surface area is a key factor related to catalyst performance. Therefore, the specific surface area of the catalysts was determined based on the Brauner-Emmett-Teller (BET) method. There was no significant difference between the Pd/CrO2 and Pt/CrO2 as their corresponding values were found to be 31.54 m 2 g −1 and 30.08 m 2 g −1 , respectively. High-resolution transmission electron microscopy (HRTEM) was also applied in the characterization of the catalysts. The fibrous structure of the chromium dioxide nanowires can be seen on the TEM images, and it was found that their diameter varied between 6 and 60 nm, while their length was between 60 and 870 nm (Figure 2A-D). In the Pd/CrO2 system, the palladium nanoparticles were also located on the surface of the CrO2 nanowires, and the corresponding particle size was quite small, 6-20 nm ( Figure 2B). In contrast, in the case of the platinum-decorated sample, the particles formed larger (70-80 nm) aggregates ( Figure 2D).
On closer inspection, it can be observed that the 70-80 nm spherical Pt aggregates build up from several very small (only several nanometers) particles ( Figure S1 in the Supplementary Materials). The smaller platinum particles have a significantly larger specific surface with more catalytically active Pt atoms on the surface than in the case of palladium aggregates. These surface atoms are accessible to the reactant molecules (DNT and H2), which results in a more efficient transformation, as is described later. Deposition of the two precious metals was different, since the reduction of palladium(II)-ions from the corresponding nitrate salt by ethanol proceeded easily. However, the reduction of Pt(IV)ions from their stable complex (H2PtCl6) failed with methanol. Thus, the deposition of platinum particles was realized using another stronger reducing agent, namely, hydrazine solution. The different reduction methods may have resulted in different nucleation and crystallization rates, which could influence the particle size. Consequently, the different reduction methods may explain the aggregate formation of the platinum nanoparticles.
The presence of the elemental palladium was confirmed by XRD measurements. On the diffractogram of the Pd/CrO2 catalyst, reflections at 40.2° (111), 46.1° (200), and 66.9° (220) 2θ degrees (JCPDS card number 046-1043) could be seen, which can be associated with the presence of elemental palladium ( Figure 2E). The platinum catalyst was also identified and characterized, and reflection peaks were found at 39.8° (111), 45.9° (200), and 67.0° (220) (two theta degrees) (JCPDS card 04-0802), which indicates the presence of elemental platinum ( Figure 2E). Thus, the reduction of the Pd(II) and Pt(IV) ions was successful. Additional peaks were also visible on the diffractograms at 28.7°, 36.9°, 42.4°, 56.2°, 59.1°, and 71.7° (two theta degrees), which were identified as the reflections of the (110), (101), (111), (211), (220), and (301) crystalline phases of the catalyst support, magnetic chromium dioxide (JCPDS card number 84-1819). The specific surface area is a key factor related to catalyst performance. Therefore, the specific surface area of the catalysts was determined based on the Brauner-Emmett-Teller (BET) method. There was no significant difference between the Pd/CrO 2 and Pt/CrO 2 as their corresponding values were found to be 31.54 m 2 g −1 and 30.08 m 2 g −1 , respectively. High-resolution transmission electron microscopy (HRTEM) was also applied in the characterization of the catalysts. The fibrous structure of the chromium dioxide nanowires can be seen on the TEM images, and it was found that their diameter varied between 6 and 60 nm, while their length was between 60 and 870 nm (Figure 2A-D). In the Pd/CrO 2 system, the palladium nanoparticles were also located on the surface of the CrO 2 nanowires, and the corresponding particle size was quite small, 6-20 nm ( Figure 2B). In contrast, in the case of the platinum-decorated sample, the particles formed larger (70-80 nm) aggregates ( Figure 2D).

Comparison of the Catalytic Activity of the Synthesized Pd and Pt Decorated Chromium(IV) Oxide Nanowires
The catalytic activity of the magnetic Pd-and Pt-containing catalysts was tested in DNT hydrogenation at four different reaction temperatures, and the actual DNT concentration was continuously monitored. The precious metal-free CrO2 support was also tested at 333 K and 20 bar pressure. In the case of the palladium-and platinum-free CrO2, a 53.9 n/n% DNT conversion and a 20.1 n/n% TDA yield were reached after 240 min. The low DNT conversion and TDA yield indicate that Pd or Pt is necessary to achieve a catalytic system of high-performance.
By applying the Pd/CrO2 or Pt/CrO2 catalyst, the hydrogenation reaction under investigation happens according to a pseudo-first-order kinetics, which was verified by linear regression on the initial experimental points of the ln(cDNT) vs. reaction time plot (Figure 3). On closer inspection, it can be observed that the 70-80 nm spherical Pt aggregates build up from several very small (only several nanometers) particles ( Figure S1 in the Supplementary Materials). The smaller platinum particles have a significantly larger specific surface with more catalytically active Pt atoms on the surface than in the case of palladium aggregates. These surface atoms are accessible to the reactant molecules (DNT and H 2 ), which results in a more efficient transformation, as is described later. Deposition of the two precious metals was different, since the reduction of palladium(II)-ions from the corresponding nitrate salt by ethanol proceeded easily. However, the reduction of Pt(IV)ions from their stable complex (H 2 PtCl 6 ) failed with methanol. Thus, the deposition of platinum particles was realized using another stronger reducing agent, namely, hydrazine solution. The different reduction methods may have resulted in different nucleation and crystallization rates, which could influence the particle size. Consequently, the different reduction methods may explain the aggregate formation of the platinum nanoparticles.
The presence of the elemental palladium was confirmed by XRD measurements. On the diffractogram of the Pd/CrO 2 catalyst, reflections at 40.2 • (111), 46.1 • (200), and 66.9 • (220) 2θ degrees (JCPDS card number 046-1043) could be seen, which can be associated with the presence of elemental palladium ( Figure 2E). The platinum catalyst was also identified and characterized, and reflection peaks were found at 39.8 • (111), 45.9 • (200), and 67.0 • (220) (two theta degrees) (JCPDS card 04-0802), which indicates the presence of elemental platinum ( Figure 2E). Thus, the reduction of the Pd(II) and Pt(IV) ions was successful. Additional peaks were also visible on the diffractograms at 28.  The catalytic activity of the magnetic Pd-and Pt-containing catalysts was tested in DNT hydrogenation at four different reaction temperatures, and the actual DNT concentration was continuously monitored. The precious metal-free CrO 2 support was also tested at 333 K and 20 bar pressure. In the case of the palladium-and platinum-free CrO 2 , a 53.9 n/n% DNT conversion and a 20.1 n/n% TDA yield were reached after 240 min. The low DNT conversion and TDA yield indicate that Pd or Pt is necessary to achieve a catalytic system of high-performance.
By applying the Pd/CrO 2 or Pt/CrO 2 catalyst, the hydrogenation reaction under investigation happens according to a pseudo-first-order kinetics, which was verified by linear regression on the initial experimental points of the ln(c DNT ) vs. reaction time plot ( Figure 3).

Comparison of the Catalytic Activity of the Synthesized Pd and Pt Decorated Chromi Oxide Nanowires
The catalytic activity of the magnetic Pd-and Pt-containing catalysts was te DNT hydrogenation at four different reaction temperatures, and the actual DNT c tration was continuously monitored. The precious metal-free CrO2 support w tested at 333 K and 20 bar pressure. In the case of the palladium-and platinum-fre a 53.9 n/n% DNT conversion and a 20.1 n/n% TDA yield were reached after 240 m low DNT conversion and TDA yield indicate that Pd or Pt is necessary to achieve lytic system of high-performance.
By applying the Pd/CrO2 or Pt/CrO2 catalyst, the hydrogenation reaction un vestigation happens according to a pseudo-first-order kinetics, which was verified ear regression on the initial experimental points of the ln(cDNT) vs. reaction time pl ure 3). High DNT conversions were achieved with both catalysts. The platinum-con sample reacted faster, which can be seen by comparing the steep run of the conv High DNT conversions were achieved with both catalysts. The platinum-containing sample reacted faster, which can be seen by comparing the steep run of the conversion curves ( Figure 4A,B). The system reached full conversion with the Pt/CrO 2 catalyst after 40 min at 323 K and 20 bar hydrogen pressure. The Pd/CrO 2 sample was slightly less active than its counterpart, but, after 80 min of hydrogenation at 333 K and 20 bar, it was also able to reach full DNT conversion ( Figure 4B). curves ( Figure 4A,B). The system reached full conversion with the Pt/CrO2 catalyst after 40 min at 323 K and 20 bar hydrogen pressure. The Pd/CrO2 sample was slightly less active than its counterpart, but, after 80 min of hydrogenation at 333 K and 20 bar, it was also able to reach full DNT conversion ( Figure 4B). A non-linear regression method was applied to calculate the reaction rate constants (k) (see Equation (3) in the experimental section and Table 1) [35]. In the case of the Pt/CrO2, the k values were higher (e.g., 2.9 × 10 −3 s −1 at 333 K) compared to the Pd-containing catalyst (e.g., 1.5 × 10 −3 s −1 at 333 K) despite the lower quantity of the precious metal (1.55 × 10 −4 mol Pt in 1 g catalyst and 4.25 × 10 −4 mol Pd in 1 g catalyst, respectively).
The formation of semi-hydrogenated intermediates, 4-amino-2-nitrotoluene (4A2NT) and 2-amino-4-nitrotoluene (2A4NT), were confirmed during the hydrogenation ( Figure  5). In the case of the Pt/CrO2, the total amount of the intermediates was converted to TDA A non-linear regression method was applied to calculate the reaction rate constants (k) (see Equation (3) in the experimental section and Table 1) [35]. In the case of the Pt/CrO 2 , the k values were higher (e.g., 2.9 × 10 −3 s −1 at 333 K) compared to the Pd-containing catalyst (e.g., 1.5 × 10 −3 s −1 at 333 K) despite the lower quantity of the precious metal (1.55 × 10 −4 mol Pt in 1 g catalyst and 4.25 × 10 −4 mol Pd in 1 g catalyst, respectively).
The activation energies (Ea) were also calculated based on the Arrhenius plot by applying the rate constant (k) vs. temperature (T) diagrams ( Figure 4). The fitting was carried out using the original exponential Arrhenius expression (Equation (5)) to avoid the biased error propagation of the parameters ( Figure S2A,B). There was no significant difference between the Ea values of the reactions carried out with the tested catalysts (Table 1). However, these activation energies were lower compared to other experimental results in the literature (32.3 ± 7.5 and 38.6 ± 3.3 kJ/mol) [36,37].
The quantity of the adsorbed precious metals was significantly different in the case of the two catalysts. Thus, the catalytic activity of the two magnetic catalysts was compared by using the corresponding turnover numbers (TON) (Equation (4) in the experimental section), which were calculated after 1 h of hydrogenation. In both cases, the TON increased by increasing the temperature, and thus, the highest values were achieved at 333 K. In the case of the Pt/CrO 2 catalyst, 304.08 mol of TDA formed within 60 min at 333 K and 20 bar pressure ( Table 2). The palladium-containing catalyst was less active under the same reaction conditions because only 60.14 mol of TDA formed considering 1 mol of Pd. The formation of semi-hydrogenated intermediates, 4-amino-2-nitrotoluene (4A2NT) and 2-amino-4-nitrotoluene (2A4NT), were confirmed during the hydrogenation ( Figure 5). In the case of the Pt/CrO 2 , the total amount of the intermediates was converted to TDA at higher temperatures (323 K and 333 K). However, the semi-hydrogenated compounds have been detected even after four hours of catalytic hydrogenation with Pd/CrO 2 .
at higher temperatures (323 K and 333 K). However, the semi-hydrogenated compounds have been detected even after four hours of catalytic hydrogenation with Pd/CrO2.

Possible Mechanism for the Hydrogenation
The formation of TDA ( Figure 6, middle section) took place through six consecutive hydrogenation steps. In the first step, nitroso-nitrotoluene is formed, which is followed by a hydroxylamine-nitrotoluene formation. This leads to the formation of one of the abovementioned semi-hydrogenated compounds; depending on which nitro group is hydrogenated first, 2A4NT or 4A2NT will be formed. After the other nitro group is hydrogenated, the corresponding nitroso and hydroxylamine compounds will be produced. In the last step, both pathways will lead to the main product, TDA. These results are completely in line with the findings of Giovanni et al. [1], who also detected the presence of nitroso and hydroxylamine compounds during the hydrogenation of DNT over a Pd(C) catalyst. In their case, the two abovementioned compounds further reacted to give 2,2′dinitro-4,4′-azoxytoluene (DNAT). Our reaction mixture was subjected to Gas Chromatography-Mass Spectrometry (GC-MS) investigation to detect and identify any intermediate present. The possible structures were identified from the MS spectra using the NIST 08, NIST Mass Spectral Library. Much to our surprise, no DNAT or diazo compounds from the reaction of the corresponding amines and nitroxyl-compounds were detected, maybe because of their low stability under GC-MS conditions. Instead, several side-products were identified ( Figure 6, see 2-4) in trace amounts in addition to the two semi-

Possible Mechanism for the Hydrogenation
The formation of TDA ( Figure 6, middle section) took place through six consecutive hydrogenation steps. In the first step, nitroso-nitrotoluene is formed, which is followed by a hydroxylamine-nitrotoluene formation. This leads to the formation of one of the abovementioned semi-hydrogenated compounds; depending on which nitro group is hydrogenated first, 2A4NT or 4A2NT will be formed. After the other nitro group is hydrogenated, the corresponding nitroso and hydroxylamine compounds will be produced. In the last step, both pathways will lead to the main product, TDA. These results are completely in line with the findings of Giovanni et al. [1], who also detected the presence of nitroso and hydroxylamine compounds during the hydrogenation of DNT over a Pd(C) catalyst. In their case, the two abovementioned compounds further reacted to give 2,2 -dinitro-4,4azoxytoluene (DNAT). Our reaction mixture was subjected to Gas Chromatography-Mass Spectrometry (GC-MS) investigation to detect and identify any intermediate present. The possible structures were identified from the MS spectra using the NIST 08, NIST Mass Spectral Library. Much to our surprise, no DNAT or diazo compounds from the reaction of the corresponding amines and nitroxyl-compounds were detected, maybe because of their low stability under GC-MS conditions. Instead, several side-products were identified ( Figure 6, see 2-4) in trace amounts in addition to the two semi-hydrogenated intermediates (4A2NT and 2A4NT) during the reaction. The corresponding GC-MS spectra are presented in the Supplementary Materials as Figures S3-S5. In addition to the detected side-products, we assumed the formation of several other ones (1 and 5-7 in Figure 6). The reaction pathways were highlighted to explain the formation of these species. Nota condensation reaction, (E)-1-(2,4-dinitrophenyl)-N-(2-methyl-5-nitrophenyl)methanimine may form from 2,4-DNT by the loss of water ( Figure 6, see 1). Another condensed molecule, (Figure 6, see 2). Reaction between the 4-methylbenzene-1, 3 Figure 6, see 3). hydrogenated intermediates (4A2NT and 2A4NT) during the reaction. The corresponding GC-MS spectra are presented in the Supplementary Materials as Figures S3-S5. In addition to the detected side-products, we assumed the formation of several other ones (1 and 5-7 in Figure 6). The reaction pathways were highlighted to explain the formation of these species. Nota condensation reaction, (E)-1-(2,4-dinitrophenyl)-N-(2-methyl-5-nitrophenyl)methanimine may form from 2,4-DNT by the loss of water ( Figure 6, see 1). Another condensed molecule, 2-[(E)-[(2-methyl-5-nitrophenyl)imino]methyl]-5-nitrophenol was identified, which could be formed through side-reactions by water loss (Figure 6, see  2). Figure  6, see 3). Figure 6. Proposed mechanism of 2,4-dinitrotoluene hydrogenation to produce 2,4-toluenediamine (TDA) with Pd/CrO2 or Pt/CrO2 catalysts. The reaction pathways depict the formation of specific side-products (dashed frames) detected or assumed during the measurements. Figure 6, see 4). The Figure 6. Proposed mechanism of 2,4-dinitrotoluene hydrogenation to produce 2,4-toluenediamine (TDA) with Pd/CrO 2 or Pt/CrO 2 catalysts. The reaction pathways depict the formation of specific side-products (dashed frames) detected or assumed during the measurements. Figure 6, see 4). The formation of two stereoisomers, N-(2,5-dimethylphenyl)-2-methoxy-4-nitrobenzamide and N-(3methoxy-4-methylphenyl)-4-methyl-3-nitrobenzamide, could occur by the reaction of 1,4dimethyl-2-nitrobenzene and 2-methoxy-1-methyl-4-nitrobenzene ( Figure 6, see 5 and 6). Moreover, the (E)-1-(2-methoxy-4-methylphenyl)-N-(3-methoxy-4-methylphenyl)methanimine side-product could form through multiple steps ( Figure 6, see 7).

The 2-nitro-4-nitrosotoluene can react with 2-methoxy-4-methylphenol and reach 3methoxy-4-[(E)-[(4-methyl-3-nitrophenyl)imino]methyl]phenol (
As it is evident from Figure 6, compounds 1-4 and 7 contain imine bonds that can be formed via the reaction of the aromatic amines and the methyl group of the other toluene derivative. This reaction may seem very unlikely at first glance; however, Brandt et al. recently reported the direct alkylation of aliphatic amines via the activation of the sp 3 C-H bond of pure hydrocarbons, e.g., toluene using an amide-based nickel pincer catalyst [38]. The C-H bonds are very stable by nature, however, alkylation of nitrobenzene with toluene moieties has been reported using a heterogeneous cobalt catalyst in the presence of peroxide and hydrogen. [39] These findings indicate that, by altering the reaction conditions, our catalysts may be also used for the preparation of imines and/or amines by the direct reaction of C-H bonds and amines.

Turnover Number and Reusability
TDA yields were similar in the case of both catalysts, 98.8 n/n% (Pt/CrO 2 ) and 99.7 n/n% (Pd/CrO 2 ) at 333 K and 20 bar hydrogen pressure despite the fact that the quantities of the precious metals in the catalysts were different (Pt: 3.02 wt%, Pd: 4.52 wt%). At lower temperatures (303 K and 313 K), higher TDA yields (57.0 n/n% and 76.1 n/n%) were achieved by using the Pt/CrO 2 catalyst, despite its lower metal content. The catalysts were well dispersible in methanolic solution ( Figure 7B). Furthermore, because of the magnetic properties of the CrO 2 support, the separation of the catalysts from the reaction media was easy, efficient, and fast by using a magnetic field ( Figure 7C). The Pd/CrO 2 and Pt/CrO 2 particles formed stable dispersions with the reaction media in a methanolic solution of DNT and TDA ( Figure 7B). The stable catalyst dispersions could be separated using an external magnetic field since the magnetic catalyst particles were collectable by a neodymium magnet from the methanolic phase; therefore, the liquid became clear ( Figure 7C). formation of two stereoisomers, N-(2,5-dimethylphenyl)-2-methoxy-4-nitrobenzamide and N-(3-methoxy-4-methylphenyl)-4-methyl-3-nitrobenzamide, could occur by the reaction of 1,4-dimethyl-2-nitrobenzene and 2-methoxy-1-methyl-4-nitrobenzene ( Figure 6, see 5 and 6). Moreover, the (E)-1-(2-methoxy-4-methylphenyl)-N-(3-methoxy-4methylphenyl)methanimine side-product could form through multiple steps ( Figure 6,  see 7).
As it is evident from Figure 6, compounds 1-4 and 7 contain imine bonds that can be formed via the reaction of the aromatic amines and the methyl group of the other toluene derivative. This reaction may seem very unlikely at first glance; however, Brandt et al. recently reported the direct alkylation of aliphatic amines via the activation of the sp 3 C-H bond of pure hydrocarbons, e.g., toluene using an amide-based nickel pincer catalyst [38]. The C-H bonds are very stable by nature, however, alkylation of nitrobenzene with toluene moieties has been reported using a heterogeneous cobalt catalyst in the presence of peroxide and hydrogen. [39] These findings indicate that, by altering the reaction conditions, our catalysts may be also used for the preparation of imines and/or amines by the direct reaction of C-H bonds and amines.

Turnover Number and Reusability
TDA yields were similar in the case of both catalysts, 98.8 n/n% (Pt/CrO2) and 99.7 n/n% (Pd/CrO2) at 333 K and 20 bar hydrogen pressure despite the fact that the quantities of the precious metals in the catalysts were different (Pt: 3.02 wt%, Pd: 4.52 wt%). At lower temperatures (303 K and 313 K), higher TDA yields (57.0 n/n% and 76.1 n/n%) were achieved by using the Pt/CrO2 catalyst, despite its lower metal content. The catalysts were well dispersible in methanolic solution ( Figure 7B). Furthermore, because of the magnetic properties of the CrO2 support, the separation of the catalysts from the reaction media was easy, efficient, and fast by using a magnetic field ( Figure 7C). The Pd/CrO2 and Pt/CrO2 particles formed stable dispersions with the reaction media in a methanolic solution of DNT and TDA ( Figure 7B). The stable catalyst dispersions could be separated using an external magnetic field since the magnetic catalyst particles were collectable by a neodymium magnet from the methanolic phase; therefore, the liquid became clear ( Figure 7C). Four reusability tests of the catalysts were also carried out without regeneration. Before each reuse test, the catalysts were washed with methanol, and dried at 105 °C overnight; in this sense, the methanolic rinsing was a sufficient post-treatment of the catalysts Four reusability tests of the catalysts were also carried out without regeneration. Before each reuse test, the catalysts were washed with methanol, and dried at 105 • C overnight; in this sense, the methanolic rinsing was a sufficient post-treatment of the catalysts before the next use. The DNT conversion did not decrease after four cycles, and the maximum conversions were above 99 n/n % in the case of the Pt/CrO 2 catalyst ( Figure 8A). before the next use. The DNT conversion did not decrease after four cycles, and the maximum conversions were above 99 n/n % in the case of the Pt/CrO2 catalyst ( Figure 8A). However, after each cycle, a slight decrease in the slope was seen on the conversion curves, which indicate a small decrease in the reaction rate. A similar tendency was experienced in the case of the Pd/CrO2 catalyst, and the DNT conversion was also above 99 n/n% after 60 min. In summary, the catalysts remained active after four cycles without regeneration.

The Preparation of Pd/CrO2 and Pt/CrO2 Catalysts
Palladium nitrate dihydrate (0.25 g) was dissolved in 50 mL of patosolv and 2.00 g Magtrieve TM was added to the solution (Figure 1). The CrO2 was dispersed by using a Hielscher UIP1000hDT Ultrasound Homogenizer for 2 min (115 W, 19.43 kHz). Pd(IV) ions were reduced to elemental Pd nanoparticles. After the reduction, the prepared magnetic Pd/CrO2 catalyst was separated from the liquid phase by using a neodymium magnet, and it was washed by patosolv. The magnetic sample was dried at 105 °C overnight.
During the Pt/CrO2 catalyst preparation, 0.21 g of dihydrogen hexachloroplatinate(IV) dihydrate was dissolved in 50 mL distilled water, and 2.00 g Magtriev TM was dispersed in the solution by using the homogenizer. Hydrazine hydrate (1 mL) was applied as a reducing agent. The palladium and platinum nanoparticles were deposited onto the chromium(IV) oxide nanowires by adsorption, and the real Pd an Pt contents were checked by Inductively coupled plasma -optical emission spectrometry (ICP-OES) elemental analysis. The precious metal contents of the Pd/CrO2 and Pt/CrO2 catalyst were 4.52 wt% (4.25 × 10 −4 mol Pd in 1 g catalyst) and 3.02 wt% (1.55 × 10 −4 mol Pt in 1 g catalyst), respectively. However, after each cycle, a slight decrease in the slope was seen on the conversion curves, which indicate a small decrease in the reaction rate. A similar tendency was experienced in the case of the Pd/CrO 2 catalyst, and the DNT conversion was also above 99 n/n% after 60 min. In summary, the catalysts remained active after four cycles without regeneration.

The Preparation of Pd/CrO 2 and Pt/CrO 2 Catalysts
Palladium nitrate dihydrate (0.25 g) was dissolved in 50 mL of patosolv and 2.00 g Magtrieve TM was added to the solution (Figure 1). The CrO 2 was dispersed by using a Hielscher UIP1000hDT Ultrasound Homogenizer for 2 min (115 W, 19.43 kHz). Pd(IV) ions were reduced to elemental Pd nanoparticles. After the reduction, the prepared magnetic Pd/CrO 2 catalyst was separated from the liquid phase by using a neodymium magnet, and it was washed by patosolv. The magnetic sample was dried at 105 • C overnight.
During the Pt/CrO 2 catalyst preparation, 0.21 g of dihydrogen hexachloroplatinate(IV) dihydrate was dissolved in 50 mL distilled water, and 2.00 g Magtriev TM was dispersed in the solution by using the homogenizer. Hydrazine hydrate (1 mL) was applied as a reducing agent. The palladium and platinum nanoparticles were deposited onto the chromium(IV) oxide nanowires by adsorption, and the real Pd an Pt contents were checked by Inductively coupled plasma -optical emission spectrometry (ICP-OES) elemental analysis. The precious metal contents of the Pd/CrO 2 and Pt/CrO 2 catalyst were 4.52 wt% (4.25 × 10 −4 mol Pd in 1 g catalyst) and 3.02 wt% (1.55 × 10 −4 mol Pt in 1 g catalyst), respectively.

Characterization Technics
The morphology and particle size of the palladium-and platinum-decorated CrO 2 nanowires were examined by high-resolution transmission electron microscopy (HRTEM, FEI Technai G2 electron microscope, 200 kV). Aqueous suspensions of the samples were dosed onto the surface of copper grids (Ted Pella Inc., only carbon, 300 mesh). The particle size of the nanoparticles was measured on the HRTEM images based on the original scale bar by using the ImageJ software. X-ray diffraction (XRD) measurements were applied to identify the metal phases in the catalytic systems by using a Rigaku Miniflex II diffractometer with a Cu Kα radiation source (30 kV, 15 mA). Specific surface area measurements were also carried out by using a nitrogen adsorption-desorption method at 77 K with a Micromeritics ASAP 2020 sorptometer based on the Brauner-Emmett-Teller (BET) method. The quantities of platinum and palladium in the catalysts were measured by using a Varian 720-ES inductively coupled plasma-optical emission spectrometer (ICP-OES). The analysis was carried out by using the Merck Certipur ICP multi-element IV standard. The samples were solved in aqua regia (mixture of nitric acid and hydrochloric acid in a molar ratio of 1:3). The quantitative analysis of the samples was carried out by an Agilent 7890A gas chromatograph coupled with an Agilent 5975C Mass Selective detector. During the measurements, a Restek Rxi-1MS column was used (30 m × 0.25 mm × 0.25 mm). Three analytical standards, 2,4-toluenediamine, 2,4-dinitrotoluene, and 2-methyl-5-nitroaniline (Sigma Aldrich Ltd., Steinheim, Germany) were used in the analysis of the samples. The presence and structure of the intermediates were deduced from the data contained in the NIST 08, NIST Mass Spectral Library.

Catalytic Tests
The prepared chromium(IV)-oxide-supported catalysts were compared in the catalytic hydrogenation of 2,4-dinitrotoluene (DNT). The Büchi Uster Picoclave reactor with a 200 mL stainless steel vessel and heating jacket was used for the tests. Hydrogen pressure was regulated and kept at 20 bar, and the reaction was carried out at 303, 313, 323, and 333 K. Samples were collected after 5, 10, 15, 20, 30, 40, 60, 120, 180, and 240 min. The initial concentration of 2,4-dinitrotoluene was 40 mmol L −1 in methanol. Only 0.1 g of catalyst was added to 150 mL of DNT solution. In order to check the catalyst stability, reusability tests were also carried out four times without regeneration.
The concentrations of the main product (TDA), the detected by-products and intermediates were measured by using an Agilent 7890A gas chromatograph coupled with an Agilent 5975C Mass Selective detector. The TDA formation was followed by applying a set of analytical standards (2,4-dinitrotoluene, 2,4-diaminotolune, 2-methyl-5-nitroaniline, 2methyl-3-nitroaniline, 4-methyl-3-nitroaniline, and 4-methyl-2-nitroaniline, Sigma Aldrich Ltd., Steinheim, Germany). The efficiency of the catalysts was defined by calculating the conversion (X%) of DNT based on the following equation (Equation (1)): The TDA yield (Y%) in case of the catalysts was also calculated (Equation (2)): Y % = n f ormed TDA n theoritical TDA × 100 By assuming that the hydrogenation process is a first-order reaction, based on the initial and measured DNT concentrations (c 0 and c k , mol/dm 3 ), the reaction rate constant (k) was calculated at different temperatures by applying non-linear regression ( Figure 4) as follows (Equation (3)): Turnover numbers (TON) were also calculated after 60 min of hydrogenation to compare the activities of the catalysts based on the follow equation (Equation (4)): where n TDA is the number of moles of the formed TDA and n catalyst is the amount of the applied precious metal.
The activation energies (Ea) were calculated based on the Arrhenius equation by non-linear regression using (Equation (5)):

Conclusions
A fast, efficient, and easy catalyst preparation method was developed to produce magnetic chromium(IV) dioxide nanowires that supported palladium and platinum catalysts. The catalysts showed high catalytic activity in the hydrogenation of 2,4-dinitrotoluene. TDA yields were 99.7 n/n % (Pt/CrO 2 ) and 98.8 n/n% (Pd/CrO 2 ) at 333 K and 20 bar hydrogen pressure. By GC-MS investigations, trace amounts of side-products were also detected, some of which turned out to be imine derivatives. The imine bonds can be formed via the reaction of the aromatic amines and the methyl group of the other toluene derivative. Based on our experimental results, a possible mechanism for the formation of TDA and the detected side-products was proposed. These findings indicate that, by altering the reaction conditions, our catalysts may be used for the preparation of imines and/or amines by the direct reaction of C-H bonds and amines. Turnover numbers (TONs) were calculated after one hour of hydrogenation, and, in the case of the Pt/CrO 2 catalyst 304.8 mol of TDA was formed, whereas in the case of the Pd/CrO 2 catalyst, only 60.14 mol TDA was formed, considering 1 mol of precious metal at 333 K and 20 bar.
In summary, excellent magnetic Pd/CrO 2 and Pt/CrO 2 catalysts were achieved and successfully tested in a commercially important hydrogenation reaction. For commercial applications, one of the biggest advantages of these catalysts is their easy and fast preparation, which could reduce catalyst cost. They were well dispersible in the reaction media (methanolic solution of DNT), resulting in a remarkable catalytic activity. In addition, the CrO 2 magnetic support eliminates the inherent separation issue of nano-formulated catalyst particles. The prepared catalysts can be removed from the medium quickly and completely by using an external magnetic field, which may significantly reduce catalyst loss and, as a result, the production cost. The catalysts can be reused at least four times without regeneration, and GC-MS analysis of the reaction mixture suggests that, by altering the reaction conditions, carbon-nitrogen bonds can also be formed, making these nanoparticles multifunctional catalysts.