Selective Hydrogenation of Acetylene over Pd-Mn / Al 2 O 3 Catalysts

: Novel bimetallic Pd-Mn / Al 2 O 3 catalysts are designed by the decomposition of cyclopentadienylmanganese tricarbonyl (cymantrene) on reduced Pd / Al 2 O 3 in an H 2 atmosphere. The peculiarities of cymantrene decomposition on palladium and, thus, the formation of bimetallic Pd-Mn catalysts are studied. The catalysts are characterized by N 2 adsorption, H 2 pulse chemisorption, temperature-programmed desorption of hydrogen (TPD-H 2 ), transmission electron microscopy (TEM), energy-dispersive X-ray spectroscopy (EDX), X-ray di ﬀ raction (XRD), and di ﬀ use reﬂectance infrared Fourier transform spectroscopy (DRIFTS). The modiﬁed catalysts show the changed hydrogen chemisorption properties and the absence of weakly bonded hydrogen. Using an organomanganese precursor provides an uniform Mn distribution on the catalyst surface. Tested in hydrogenation of acetylene, the catalysts show both higher activity and selectivity to ethylene (20% higher) compared to the non-modiﬁed Pd / Al 2 O 3 catalyst. The inﬂuence of the addition of Mn and temperature treatment on catalyst performance is studied. The optimal Mn content and treatment temperature are found. It is established that modiﬁcation with Mn changes the route of acetylene hydrogenation from a consecutive scheme for Pd / Al 2 O 3 to parallel one for the Pd-Mn samples. The reaction rate shows zero overall order by reagents for all tested catalysts. validation, A.G., V.S.; V.V., D.M.; resources, V.V., A.G.; data curation, D.M., V.S., E.S., M.K.; writing—original draft preparation, D.M., V.S.; writing—review and editing, D.M., V.S., A.G.; visualization, D.M., V.L.; supervision, V.V., V.S.; project administration, funding A.G.,


Introduction
Ethylene is one of the commonly used monomers in the petrochemical industry worldwide and is produced by the steam cracking of hydrocarbons. Ethylene cuts typically comprise 0.5%-2% of acetylene, which is a poison for the polymerization catalysts and should be removed by selective hydrogenation to ethylene [1]. A number of active metals (Pd, Ni, Au) modified with a wide range of elements (Ag, Cu, Si, Ga, Sn, Pb, In, S, Fe) and supported on various carriers (Al 2 O 3 , SiO 2 , TiO 2 , ZnO) were investigated .
Monometallic Pd catalysts show a high activity but low selectivity to ethylene, so Pd is typically promoted with other metals. In industry, Pd-Ag/Al 2 O 3 catalysts are widely used and much research is devoted to Pd-Ag compositions supported on alumina or silica. It is supposed that the promotion is based on an increased electronic density of the Pd d-band resulting in a decrease in ethylene [3] or hydrogen adsorption with further spill over [4]. In addition, it is suggested that the promotion is caused by not only an electronic but also a geometric effect [5], or just geometric [8]. Pd-Ag catalysts expose not only a higher selectivity to ethylene, but also a lower yield of C 6+ hydrocarbons (green oil) compared with the Pd/Al 2 O 3 sample. Moreover, the Mn-modified samples have shown higher activity. It is found that the addition of Mn suppresses the hydrogen chemisorption on Pd catalysts.

Cymantrene Decomposition on Pd/Al 2 O 3
To investigate the formation of Pd-Mn/Al 2 O 3 catalysts, the decomposition of cymantrene on Pd/Al 2 O 3 was performed in a temperature-programmed regime in an H 2 flow with a mass spectrometry analysis of effluent gas. Figure 1 shows the mass-spectra of cymantrene decomposition products in the range of 40-400 • C.

Cymantrene Decomposition on Pd/Al2O3
To investigate the formation of Pd-Mn/Al2O3 catalysts, the decomposition of cymantrene on Pd/Al2O3 was performed in a temperature-programmed regime in an H2 flow with a mass spectrometry analysis of effluent gas. Figure 1 shows the mass-spectra of cymantrene decomposition products in the range of 40-400 °C. As shown, at the initial step (temperatures of 80-150 °C) there are peaks with m/z 28 (carbon monoxide), 42 (cyclopentane), and 67 (cyclopentene) [45]. In the temperature range of 270-400 °C, one can observe two peaks with m/z 16 (methane) and 28 (carbon monoxide). The peaks corresponding to cyclopentadiene (m/z 65, 66) are absent. More details about the mass spectra interpretation are shown in the Supplementary Materials.
We may conclude, therefore, that the cyclopentadienyl ligand of cymantrene is removed after hydrogenation, mainly as cyclopentane at 80-150 °C. As for carbon monoxide, it is strongly bonded with metals, and may be removed as methane at temperatures above 270 °C [46]. However, as evidenced by the mass-spectrometric analysis of the effluent gas, to complete a CO removal a treatment in an H2 flow at 400 °C for 10 min is necessary. Table 1 summarizes the properties of the prepared catalysts. The designation of the samples shows the atomic Mn/Pd ratio and the treatment temperature, which is the final temperature of the cymantrene decomposition. As Table 1 shows, the Brunauer-Emmett-Teller (BET) surface area of the samples is the same within the margin of error, which indicates that the addition of Mn has no significant effect on the surface area of the catalysts. However, the samples show quite a different behavior in H2 chemisorption.  As shown, at the initial step (temperatures of 80-150 • C) there are peaks with m/z 28 (carbon monoxide), 42 (cyclopentane), and 67 (cyclopentene) [45]. In the temperature range of 270-400 • C, one can observe two peaks with m/z 16 (methane) and 28 (carbon monoxide). The peaks corresponding to cyclopentadiene (m/z 65, 66) are absent. More details about the mass spectra interpretation are shown in the Supplementary Materials.

Catalysts Characterization
We may conclude, therefore, that the cyclopentadienyl ligand of cymantrene is removed after hydrogenation, mainly as cyclopentane at 80-150 • C. As for carbon monoxide, it is strongly bonded with metals, and may be removed as methane at temperatures above 270 • C [46]. However, as evidenced by the mass-spectrometric analysis of the effluent gas, to complete a CO removal a treatment in an H 2 flow at 400 • C for 10 min is necessary. Table 1 summarizes the properties of the prepared catalysts. The designation of the samples shows the atomic Mn/Pd ratio and the treatment temperature, which is the final temperature of the cymantrene decomposition. As Table 1 shows, the Brunauer-Emmett-Teller (BET) surface area of the samples is the same within the margin of error, which indicates that the addition of Mn has no significant effect on the surface area of the catalysts. However, the samples show quite a different behavior in H 2 chemisorption. The non-promoted Pd/Al 2 O 3 catalyst uptakes a significant amount of H 2 (1.2 µmol/g), but any addition of Mn decreases the H 2 adsorption. For example, an addition of 0.029% Mn (PdMn-1-250 and PdMn-1-330) decreases the H 2 adsorption to 1.1 µmol/g (for the sample treated at 330 • C) and to 0.18 µmol/g (for the sample treated at 250 • C). This trend is enhanced by a further addition of Mn: both PdMn-2-300 and PdMn-2-350 uptake significantly less H 2 (0.03 and 0.04 µmol/g). It should be noted that there are two possible reasons for the decreasing H 2 adsorption: the shielding of Pd with Mn atoms and the blocking of H 2 adsorption sites by residual CO ligands. Moreover, the selectivity of Pd-Mn catalysts to ethylene is correlated with their H 2 adsorption, as shown in Table 1.

Catalysts Characterization
As depicted in Figure 2, the non-modified Pd-250 desorbs H 2 in the range of 80-250 • C, indicating a desorption of weakly bonded hydrogen at low temperatures and strongly bonded hydrogen (or Pd hydride decomposition) at a temperature ramp. The non-promoted Pd/Al2O3 catalyst uptakes a significant amount of H2 (1.2 µmol/g), but any addition of Mn decreases the H2 adsorption. For example, an addition of 0.029% Mn (PdMn-1-250 and PdMn-1-330) decreases the H2 adsorption to 1.1 µmol/g (for the sample treated at 330 °C) and to 0.18 µmol/g (for the sample treated at 250 °C). This trend is enhanced by a further addition of Mn: both PdMn-2-300 and PdMn-2-350 uptake significantly less H2 (0.03 and 0.04 µmol/g). It should be noted that there are two possible reasons for the decreasing H2 adsorption: the shielding of Pd with Mn atoms and the blocking of H2 adsorption sites by residual CO ligands. Moreover, the selectivity of Pd-Mn catalysts to ethylene is correlated with their H2 adsorption, as shown in Table 1.
As depicted in Figure 2, the non-modified Pd-250 desorbs H2 in the range of 80-250 °C, indicating a desorption of weakly bonded hydrogen at low temperatures and strongly bonded hydrogen (or Pd hydride decomposition) at a temperature ramp. The samples with a low Mn content (PdMn-1-250 and PdMn-1-330) do not desorb H2 at temperatures below 230 °C, which indicates the presence of strongly bonded hydrogen (or Pd hydride). The samples (PdMn-2-300 and PdMn-2-350) with a high Mn content demonstrate only an insignificant H2 desorption at a temperature of 250 °C, and these findings correlate with the chemisorption data (Table 1). Decreasing strongly chemisorbed hydrogen is recommended for acetylene selective hydrogenation as reported in [4,25,47].
As depicted on the TEM images of the PdMn-2-300, Pd nanoparticles (NP) with a lattice spacing of about 0.228 nm are found, which are indexed as the (111) plane for cubic palladium doped with clusters from single Mn atoms ( Figure 3a) and Mn crystallites ( Figure 3b) [48,49]. Depending on the lattice spacing, the fringes on the TEM images could be assigned to Pd nanoparticles or Mn crystallites and in some cases to manganese oxides with a lattice spacing of about 0.47-0.49 nm [49,50]. Due to the overlapping of Mn crystallites on Pd NPs, it is difficult to measure accurately the palladium nanoparticles' size and their distributions [51], but TEM images show Pd NPs in the range of 5-10 nm with a mean particle size of about 6.7 ± 0.2 nm (Figure 3c), in agreement with the literature data [48].  The samples with a low Mn content (PdMn-1-250 and PdMn-1-330) do not desorb H 2 at temperatures below 230 • C, which indicates the presence of strongly bonded hydrogen (or Pd hydride). The samples (PdMn-2-300 and PdMn-2-350) with a high Mn content demonstrate only an insignificant H 2 desorption at a temperature of 250 • C, and these findings correlate with the chemisorption data (Table 1). Decreasing strongly chemisorbed hydrogen is recommended for acetylene selective hydrogenation as reported in [4,25,47].
As depicted on the TEM images of the PdMn-2-300, Pd nanoparticles (NP) with a lattice spacing of about 0.228 nm are found, which are indexed as the (111) plane for cubic palladium doped with clusters from single Mn atoms ( Figure 3a) and Mn crystallites (Figure 3b) [48,49]. Depending on the lattice spacing, the fringes on the TEM images could be assigned to Pd nanoparticles or Mn crystallites and in some cases to manganese oxides with a lattice spacing of about 0.47-0.49 nm [49,50]. Due to the overlapping of Mn crystallites on Pd NPs, it is difficult to measure accurately the palladium nanoparticles' size and their distributions [51], but TEM images show Pd NPs in the range of 5-10 nm with a mean particle size of about 6.7 ± 0.2 nm (Figure 3c), in agreement with the literature data [48].  XRD found no reflections, which could be related to Pd and Mn due to the low metal content, as Figure 4 shows. XRD found no reflections, which could be related to Pd and Mn due to the low metal content, as Figure 4 shows. Additional information about the chemisorption properties of the catalysts is obtained using DRIFT spectroscopy of PdMn-1-330. Figure 5 shows two spectra of the catalyst samples. For the first measurement, one sample is just treated in a vacuum for 2 h. For the second measurement, another sample is preliminarily treated with H2 at 250 °C (30 min), acetylene at 20 °C (10 min), and H2 at 20 °C (10 min) with a final purge with Ar at 20 °C (10 min) and treated in a vacuum for 2 h. As the spectra show, there are two regions: 2500-3800 cm −1 , which corresponds to the vibrational spectra of O-H and C-H bonds, and 700-2400 cm −1 , ascribed to the vibrational spectra of Al2O3, adsorbed water, carbonyls, and others [52]. After the pretreatment, seven additional bands  XRD found no reflections, which could be related to Pd and Mn due to the low metal content, as Figure 4 shows.  As the spectra show, there are two regions: 2500-3800 cm −1 , which corresponds to the vibrational spectra of O-H and C-H bonds, and 700-2400 cm −1 , ascribed to the vibrational spectra of Al2O3, adsorbed water, carbonyls, and others [52]. After the pretreatment, seven additional bands As the spectra show, there are two regions: 2500-3800 cm −1 , which corresponds to the vibrational spectra of O-H and C-H bonds, and 700-2400 cm −1 , ascribed to the vibrational spectra of Al 2 O 3 , adsorbed water, carbonyls, and others [52]. After the pretreatment, seven additional bands are observed: 2962, 2932, 2872, 1483, 1252, 1220, and 1002 cm −1 . The bands 2962, 2932, and 2872 cm −1 may be ascribed to C-H stretching in the C 2 H 6 molecule [53] and the bands 1220 and 1252 cm −1 to vibrations of C-C bonds in the C 2 H 2 molecule [54]. The band 1002 cm −1 may be assigned to C=C bending in the C 2 H 4 [53]. It should be stressed that all bands above are observed only after the treatment of PdMn-1-330 with C 2 H 2 . After vacuum treatment of the sample (2 h, 200 • C), the intensity of the spectra in the region of 2962-2872 cm −1 decreases slightly, which points out the strong chemisorption of the species above.
The DRIFT spectra of adsorbed CO are considered in the Supplementary Materials. As shown by the DRIFT, the CO adsorption over the Pd-Mn catalysts was weak and negligible (at most 0.015 units Kubelka-Munk). After the vacuum treatment at room temperature, all peaks in the range of 2195-1871 cm −1 disappeared. So, one may conclude that there is an absence of strong CO chemisorption on the catalysts.   [53] and the bands 1220 and 1252 cm −1 to vibrations of C-C bonds in the C2H2 molecule [54]. The band 1002 cm −1 may be assigned to C=C bending in the C2H4 [53]. It should be stressed that all bands above are observed only after the treatment of PdMn-1-330 with C2H2. After vacuum treatment of the sample (2 h, 200 °C), the intensity of the spectra in the region of 2962-2872 cm −1 decreases slightly, which points out the strong chemisorption of the species above. The DRIFT spectra of adsorbed CO are considered in the Supplementary Materials. As shown by the DRIFT, the CO adsorption over the Pd-Mn catalysts was weak and negligible (at most 0.015 units Kubelka-Munk). After the vacuum treatment at room temperature, all peaks in the range of 2195-1871 cm −1 disappeared. So, one may conclude that there is an absence of strong CO chemisorption on the catalysts.   Based on these experimental data and taking into account the selectivity obtained (Figure 7), we may consider a mass ratio of Mn/Pd~1 (atomic ratio Mn/Pd~2) and a treatment temperature of 300 • C for 30 min as optimal, corresponding to the PdMn-2-300 sample.

Catalytic Tests
Catalysts 2020, 10, x FOR PEER REVIEW 8 of 14 Based on these experimental data and taking into account the selectivity obtained (Figure 7), we may consider a mass ratio of Mn/Pd ~ 1 (atomic ratio Mn/Pd ~ 2) and a treatment temperature of 300 °C for 30 min as optimal, corresponding to the PdMn-2-300 sample.  Figure 6b shows the conversion vs. contact time for PdMn-2-300 at 30, 40, and 50 °C. At all temperatures, the X(t) lines are straight, so the zero order by reagents is kept. Figure 7 shows the selectivity to ethylene on the conversion for all samples at 40 °C. For Pd-250, the selectivity is the lowest and the curve has a maximum, which is typical in the case of a consecutive scheme of acetylene hydrogenation to ethylene and ethane: in accordance with [55]. However, Mn-containing samples maintain a selectivity at a level of 80%-92% up to acetylene conversions of more than 70%, which implies the parallel scheme of acetylene hydrogenation to ethylene and ethane: as previously observed on Pd-Fe/Al2O3 catalysts [22]. The PdMn-1-250 sample shows a selectivity of ~ 91% at a conversion of below 70%. Increasing the treating temperature to 330 °C (PdMn-1-330), and thus eliminating the strongly chemisorbed CO ligands, decreases the selectivity to ~80% at the same conversions. Both samples with high Mn contents show a better selectivity at high conversions. The most selective is PdMn-2-300, of which the selectivity is ~92% in the conversion range of below 70%. The  Figure 6b shows the conversion vs. contact time for PdMn-2-300 at 30, 40, and 50 • C. At all temperatures, the X(t) lines are straight, so the zero order by reagents is kept. Figure 7 shows the selectivity to ethylene on the conversion for all samples at 40 • C. For Pd-250, the selectivity is the lowest and the curve has a maximum, which is typical in the case of a consecutive scheme of acetylene hydrogenation to ethylene and ethane: in accordance with [55]. However, Mn-containing samples maintain a selectivity at a level of 80%-92% up to acetylene conversions of more than 70%, which implies the parallel scheme of acetylene hydrogenation to ethylene and ethane: as previously observed on Pd-Fe/Al 2 O 3 catalysts [22]. The PdMn-1-250 sample shows a selectivity of~91% at a conversion of below 70%. Increasing the treating temperature to 330 • C (PdMn-1-330), and thus eliminating the strongly chemisorbed CO ligands, decreases the selectivity to~80% at the same conversions.
Both samples with high Mn contents show a better selectivity at high conversions. The most selective is PdMn-2-300, of which the selectivity is~92% in the conversion range of below 70%. The result is in accordance with the data published [56], where a catalyst having an Mn/Pd atomic ratio of two provides the highest selectivity to 1,3-butadiene in vinylacetylene hydrogenation. Figure 7b shows the influence of the reaction temperature on the S(X) curve for PdMn-2-300 at 40 and 50 • C. It is obvious that the selectivity is the same (~90%) up to a conversion of~70% irrespective of the temperature, which indicates that the activation energies of both reaction routes in scheme two are very close.
Our catalyst advantages are illustrated in Table 2, comprising the characteristics of the best Pd-containing catalysts in acetylene hydrogenation.  [24]: acetylene concentration and conversion, GHSV, Pd content and Pd density.
As Table 2 shows, the catalysts developed are of the same order of activity (in terms of turnover frequency (TOF)) but exceed the known one in ethylene yield (the product of X and S, calculated by acetylene (defficiency) conversion without hydrogen (excess)) and molar activity under mild conditions.
One may conclude, therefore, that modification with manganese improves both the activity (in terms of mol/g Pd /h) and selectivity of palladium catalysts in acetylene hydrogenation.
For qualitative evaluation of the catalyst stability on a laboratory scale, the selectivity to the C 4 compound (namely, 1,3-butadiene as an initial dimerization product of the acetylenic species [9]) is used, which is a fundamental indicator of the palladium-containing catalyst stability in the selective hydrogenation of acetylene [8]. The absence of C 4 hydrocarbons is confirmed by GC and MS-analyses for all set experiments performed. More details about the analyses and the stability of the catalysts are shown in the Supplementary Materials. Based on the evidence above, we consider the Pd-Mn/Al 2 O 3 catalysts to be stable for at least 5 h.

Catalysts Preparation
The initial catalyst 0.068% Pd/Al 2 O 3 was prepared by a wet impregnation of γ-Al 2 O 3 (preliminary calcined for 3 h at 600 • C) with an aqueous ammonia solution of PdCl 2 at pH = 12 (24 h). After a vacuum evaporation of the solvent, the catalyst was dried out at 70 • C for 12 h. Then, the catalyst was reduced with H 2 (20 mL/min) at 250 • C for 1 h.
Pd-Mn/Al 2 O 3 samples were prepared by a wet impregnation of the reduced Pd/Al 2 O 3 sample with a cymantrene solution in n-hexane. After the vacuum evaporation of the solvent, the samples were treated in an H 2 flow (20 mL/min) at 250-350 • C for 1 h. The effluent gas was analyzed with a quadrupole mass spectrometer QMS-200 (Stanford Research Systems, Sunnyvale, CA, USA).

Catalyst Characterization
The BET surface area was measured using Gemini VII (Micromeritics Instrument Corp.; Norcross, GA, USA). The samples were degassed at 150 • C for 3 h. The specific surface area was calculated using the BET model for adsorption data in the range of relative pressures P/P 0 = 0.05-0.30.
The metal content of the samples was measured by atomic absorption spectrometry (Perkin-Elmer-AAS, Waltham, MA, USA).
Pulse chemisorption of H 2 and TPD-H 2 was performed by AutoChem 2950HP (Micromeritics Instrument Corp.; Norcross, GA, USA). The samples were preliminarily reduced with H 2 at 250 • C for 1 h, purged with Ar for 30 min and cooled down to 35 • C. The pulse chemisorption was performed with a mixture 10% H 2 + Ar (balance), with a pulse volume of 0.5 mL, in an Ar flow (40 mL/min). TPD-H 2 was performed in an Ar flow (40 mL/min) at a heating rate of 30 K/min to 250 • C.
Transmission electron microscopy (TEM) analysis was carried out using a JEOL JEM-2100 microscope (Jeol Ltd.; Tokyo, Japan) with a 200 kV electron beam energy-dispersive X-ray analyzer (EDX). The mapping of the elements was carried out by scanning transmission electron microscopy (STEM). The samples were milled in an Eppendorf with a glass rod and ultrasonically suspended in isopropanol.
Diffuse reflectance infrared Fourier transform spectroscopy was done using a NICOLET Protégé 460 (Nicolet, Madison, WI, USA) in the range of 6000-400 cm −1 at a resolution of 4 cm −1 . For each sample, 500 spectra were recorded to get a good signal-noise ratio. CaF 2 was used as a standard. The spectra were processed with OMNIC software.

Catalystic Tests
Acetylene hydrogenation was performed in a quartz reactor at atmospheric pressure using AutoChem 2950HP (Micromeritics Instrument Corp.; Norcross, GA, USA). At a given temperature, the flow rate of the reaction mixture was changed to get various values of conversion and selectivity. The contact time was in the range of 0.26-1.81 s −1 , the reaction temperature was in the range of 30-50 ºC.
A mixture of 1.94%H 2 + 1.05%C 2 H 2 + 5.01%C 2 H 4 + Ar (balance) was used as a modeling feed preliminarily prepared in a cylinder. The effluent gas was analyzed online using a quadrupole mass spectrometer QMS-200 (Stanford Research Systems, Sunnyvale, CA, USA) and off-line using FID and TCD detectors in a GC experimental laboratory chromatograph (Gubkin University-Chromos, on the basis of GC-1000 model, Moscow-Dzerjinsk, Russia) using a packed column with HyeSep N. At given operating conditions (temperature, flow rate), the effluent gas was analyzed three times and the final concentration was calculated as the mean value of the three analyses. The carbon balance was closed within 4%.
The acetylene conversion was calculated by the equation: and ethylene selectivity by:

Conclusions
A number of Pd-Mn/Al 2 O 3 catalysts were designed by the decomposition of cymantrene on reduced Pd/Al 2 O 3 in an H 2 atmosphere. The formation of bimetallic catalysts was studied by mass spectrometry analysis of the decomposition products. It was found that the decomposition of cymantrene takes place with hydrogenation of cyclopentadienyl ligands to cyclopentene and cyclopentane, and CO ligands are partially removed by conversion to methane. The catalysts are characterized using N 2 adsorption, H 2 pulse chemisorption, TPD-H 2 , TEM, EDX, XRD, and DRIFT spectroscopy. Using the organic precursor-cymantrene provides a high and uniform distribution of Mn over Pd. The addition of manganese changes the H 2 chemisorption and desorption properties of the catalyst: the Pd-Mn/Al 2 O 3 samples have shown either a strong chemisorption of H 2 or an insignificant H 2 chemisorption. At the same time, unsaturated C 2 hydrocarbons are strongly chemisorbed on Pd-Mn/Al 2 O 3 samples and cannot be removed even under vacuum treatment at elevated temperature. Catalytic tests of the novel Pd-Mn/Al 2 O 3 catalysts in hydrogenation of acetylene have shown a higher activity and selectivity thereof to ethylene (up to 20% higher) compared to the non-promoted Pd/Al 2 O 3 catalyst. The optimal Mn/Pd ratio and treatment temperature are found. The overall reaction order by reagents is zero for all catalysts, but modification with Mn changes the reaction route from a consecutive pathway for Pd/Al 2 O 3 to a parallel one for Pd-Mn/Al 2 O 3 catalysts.