Selective hydrogenation of 1,3-butadiene in presence of 1-butene under liquid phase conditions with NiPd/Al2O3 catalysts

The catalytic performance of Al2O3-supported monometallic and bimetallic catalysts in selective hydrogenation of 1,3-butadiene in the presence of 1-butene under liquid phase conditions was studied. Bimetallic catalysts were prepared by the coimpregnation method with the required amounts of the precursors salts [Ni(NO3)2·6H2O and Pd(NH3)4Cl2·H2O] over pellet-form γ-Al2O3 with a constant content of Pd (0.5 wt%) and varying Ni/Pd atomic ratio (0.25, 0.5, 0.75, and 1) obtaining egg-shell profiles of the active components. The catalysts were characterized by X-ray diffraction, temperature-programmed techniques, such as reduction in hydrogen and desorption of ammonia, N2 physisorption, and transmission electron microscopy. The catalytic test showed that the 1,3-butadiene was selectively hydrogenated when bimetallic catalysts were used. The addition of Ni to the Pd-based catalysts suppressed n-butane formation and increased recovery of 1-butene at medium conversion. Therefore, it was observed an improved catalytic performance of the bimetallic catalysts being highest in the case of the 1NiPd/Al2O3.


Introduction
Alkenes streams produced from cracking processes for use in the petrochemical industry contain small amounts of highly unsaturated hydrocarbons, which cause problems in downstream applications, e.g., because of the oligomerization of such impurities, as 1,3-butadiene (BD) and/or acetylene on the catalyst surface leading to deactivation and increased pressure drop across the catalytic bed. Selective hydrogenation is an effective and economic way of removing these impurities by transforming them into valuable alkenes [1]. This technology employs catalytic fixed beds with cocurrent flow of the liquid hydrocarbons and gaseous hydrogen, operating temperatures ranging from ambient temperature up to around 60-70°C and total pressure up to about 200 psi for maintaining the hydrocarbon stream in liquid phase while allowing the desired level of hydrogen partial pressure [2,3]. Al 2 O 3 -supported Pd or Pt catalysts are employed in selective hydrogenation of unsaturated carbon-carbon bonds, although Pd-based catalysts are more selective than their Pt-based counterparts in selective hydrogenation of dienes [2]. Nevertheless, either Pt-or Pd-monometallic catalysts give rise to undesirable side reactions, such as the isomerization or total hydrogenation. A way to improve the selectivity consists in including an additive in the process stream [4], e.g., piperidine [5], tributylphosphine [6], diethyl ether [6], carbon monoxide [6,7], butanethiol [6], dimethyl sulfate [6], tert-butyl chloride [6], isoprene [8], and others. However, catalytic performance is more strongly dependent on the transition metals and cometals used as catalysts. Consequently, aiming at minimizing these disadvantages has led to the development of Pd-based bimetallic catalysts, to further improve their selectivity and resistance to deactivation and/or poisoning. In this regard, some researchers have made important modifications to the Pd/Al 2 O 3 catalyst [1]. M-Pd/Al 2 O 3 bimetallic catalysts synthesized since 1980s belong to the so-called third generation. Some researchers have reported significant modifications by the addition of Cu [9], Co [10], Tl [11], Fe [12,13], Ag [12,[14][15][16][17], Au [12,14,17,18], Sn [19][20][21], and Pb [22]. A common characteristic of these bimetallic catalysts is that the second metal decorates the Pd-surface; thus, the decrease in formation of butane (BA) in the bimetallic catalysts could be assigned to diminished diene conversion. Several theoretical and experimental studies have shown that Ni-containing catalysts exhibit excellent activity and selectivity in this reaction [23][24][25][26][27]. Thus, it was reported that Pd overlayers on the surface of Ni or Ni-rich Pd 8 Ni 92 alloy showed much higher activity in the partial hydrogenation of BD than single crystal Pd [28][29][30]. In a recent theoretical and experimental study, it was also reported a high activity and selectivity of the bimetallic Ni-Pd structures in the selective hydrogenation of BD in the gas phase, both on single crystal model catalysts and on the real alumina supported ones [26]. Even though some work on this type of catalysts has been reported, most of these studies were carried out by quantum-mechanics simulations of the surface, while the reported experimental work usually shows gas-phase reactions, where mass transfer effects cannot be appreciated. To the best of our knowledge, no systematic study of the effect of the Ni-content in bimetallic NiPd/Al 2 O 3 catalysts on the selective hydrogenation of BD in the presence of 1-butene (BE) under liquid phase conditions has been reported. In this sense, the objective of this work is to carry out a preliminary study of the effect of the Ni/Pd atomic ratio in NiPd/A1 2 O 3 catalysts on some physicochemical properties of the bimetallic catalysts and on their behavior in this reaction of great importance for the petrochemical industry.

Experimental
For the design of a synthesis protocol, a review of previous studies on the deposition of metallic components on porous supports was necessary [31][32][33], considering that to inhibit the total hydrogenation toward BA, it is required that the active components must be in an egg-shell profile [34].
Photographic images of a fresh (Fig. 1a) and reduced ( Fig. 1b) catalyst showed that such a required profile was indeed obtained.

Catalysts preparation
Al 2 O 3 -supported catalysts were prepared by impregnation (monometallic catalysts) or coimpregnation (bimetallic catalysts, Ni/Pd atomic ratio from 0.25 to 1). In a typical experiment, the required amounts of the precursors salts [Ni(NO 3 ) 2 Á6H 2 O and/or Pd(NH 3 ) 4 Cl 2 ÁH 2 O] were dissolved in 10 mL of distilled H 2 O. Afterward, each solution was dripped over c-Al 2 O 3 (cylindrical pellets of 1/8 00 diameter, previously heat-treated at 120°C for 12 h). The obtained suspensions were treated under reduced pressure using rotary-evaporation (60°C/60 rpm) until the removal of most of the water was accomplished. The remaining moisture was subsequently eliminated at 60°C for 12 h in dry N 2 flow (100 mL/min) followed by calcination at 500°C for 4 h under a flow of synthetic air (100 mL/min). In all cases, temperature was increased at a linear rate of 3°C/min from room temperature to the final temperature. Monometallic and bimetallic catalysts are denoted by m/Al 2 O 3 or nNiPd/A1 2 O 3 , respectively, where m is the metal nature (0.3 wt% Ni and 0.5 wt% Pd) and n is the Ni/ Pd atomic ratio with a Pd-content 0.5 % wt%, e.g., 1NiPd/ Al 2 O 3 stands for a catalyst supported on Al 2 O 3 with 0.5 wt% Pd and an Ni/Pd atomic ratio of 1.

X-ray diffraction
XRD measurements were carried out between 20 and 80°/ 2h using a BRUKER-SIEMENS D5005 instrument with Cu-Ka radiation (k = 1.5456 Å ), Ni filter, and step rate of 0.02°/s. Phase identification was made using the JCPDS library [35]. Temperature-programmed reduction in hydrogen TPR-H 2 measurements were carried out in a stainless-steel reaction line coupled to a thermal conductivity detector (TCD). The samples were first placed in a U-shaped quartz reactor and heated up to 120°C in Ar flow (30 mL/min) for 2 h and then cooled down to room temperature. Next, the gas current was changed to a 5 % H 2 in Ar mixture, and the temperature was increased to 500°C, at a heating rate of 10°C/min. The water formed during the reduction treatment was collected in a molecular sieve trap at the reactor exit, and the temperature and H 2 consumption were registered with a TCD at 6 s intervals.
Temperature-programmed desorption of ammonia TPD-NH 3 was performed to estimate the acidity of the catalysts. For this experiment, the sample was placed in a U-shaped quartz reactor and dried at 120°C for 2 h in an Ar flow (30 mL/min). Then, the solid was cooled to 40°C, and pulses of NH 3 /Ar (0.3 vol% NH 3 ) were injected up to saturation. Finally, the thermal desorption was carried out from 40 to 500°C and registered every 6 s using a TCD.
Textural properties N 2 physisorption data were measured with a MICRO-MERITICS-ASAP 2010 automatic analyzer at liquid N 2 temperature. Prior to the experiments, the samples were degassed overnight under vacuum at 60°C. Specific surface areas were calculated by the Brunauer-Emmett-Teller method (S BET ), pore volume (V p ) was determined by N 2 adsorption at a relative pressure of 0.98, and pore size distributions were obtained from the desorption isotherms by means of the Barrett-Joyner-Halenda (BJH) method.

Transmission electron microscopy
Before of the analysis, samples were dispersed in an ethanol/water mixer and sonicated, then a drop was placed on a Cu grid covered with C/collodion. TEM images were obtained in an FEI microscope model TECNAY G2 SPINT BIO-TWIN using an accelerating voltage of 120 kV.

Reduction pretreatment
Prior to starting the activity tests, catalyst samples (*1 g) were reduced ex situ under H 2 (WHSV = 12,000 h -1 ) within a fixed-bed reactor (U-shaped PYREX Ò tube 200 mm in length and 17 mm in internal diameter). In all cases, the temperature was increased at a linear rate of 3°C/min from room temperature to a final temperature of 300°C, which was held for 2 h. After pretreatment, the samples were cooled to room temperature in H 2 and immersed in anhydrous heptane to prevent reoxidation during transport and transferring to the reactor.

Catalytic performance
The reactivity measurements were carried out in a recirculating system, including an external fixed-bed reactor (Fig. 2) under the following conditions: temperature, T = 40°C; total pressure, P = 200 psi; and weight hourly space velocity, WHSV = 120 h -1 . The fixed-bed reactor was a stainless-steel 316 tube with a length 110 mm in length and 17 mm internal diameter packed in three sections (L1 = glass microspheres/L2 = catalyst in pellet form/L3 = glass microspheres) separated with glass wool plugs. Afterward, this reactor was connected to the reaction system, and the surface was refreshed using a flow of 30 mL/min of H 2 for 30 min at the reaction temperature, to remove the protective solvent and expose the active sites. The stirred vessel was fed with BE, BD, and H 2 as reactants, hexane as solvent, and pentane as internal standard, with a volume composition of HX:PA:unsaturated compounds of 73:2:25 and a molar ratio BE:BD of 20:1. The reaction was started by pressurizing with H 2 and recirculating the liquid/gas mixture by means of a high pressure pump. Tests were run for at least 180 min.
The effluent stream from the reactor was analyzed with an online gas chromatograph (AGILENT TECHNOLO-GIES model GC 6890) equipped with a flame ionization detector and GS-GASPRO capillary column (60-m length and 0.32 mm internal diameter). The chromatograms were integrated by means of the ChemStation Plus software and converted into mass and mole percentages as recommended by Huang et al. [36]. Furthermore, product selectivities were calculated as mole of product or reactant divided by the total number of mole of feed. Results and discussion XRD patterns of the support and supported catalysts are shown in Fig. 3 Fig. 4. Reduction with hydrogen of the (un-)supported PdO catalysts is a thermodynamically favored reaction, and it has been found to occur at low temperature (\100°C) [37]. Pd/Al 2 O 3 catalyst presented a low-temperature reduction signal (Fig. 4b), which is ascribed to the reduction of PdO: Some authors have defined three regions for the reduction of the Al 2 O 3 -supported NiO catalysts [38]: (1) T r B 450°C, where support-NiO interaction is weak; (2) 450°C B T r B 750°C, corresponding to strong support-NiO interaction; and (3) T r C 750°C, where NiAl 2 O 4 reduction occurs. As the catalysts were calcined at 500°C and activated by H 2 reduction at 300°C, TPR-H 2 experiments were run only up to 500°C. In agreement with the cited references, Ni/Al 2 O 3 presented a broad signal between 300 and 500°C (Fig. 4g) that could be attributed to the reduction of weakly interacting NiO species: The position and shapes of both TPR-H 2 signals are modified when increasing Ni content in the Al 2 O 3supported NiPd catalysts. On the other hand, as previously reported by Galiasso and Ravigli-Nasca [39], a group of signals between 100 and 250°C (Fig. 4c-f) could be attributed to the formation of a Ni-Pd bimetallic phase. Given the low metal contents, no signals due to Pd, PdO, Ni, NiO, or Ni-Pd species were detected by us, and only the diffraction peaks of the support were observed (Fig. 3); signals of the Ni2p and Pd2d core-level spectra obtained by XPS are also weak and noisy (not shown). The obtained TPR-H 2 traces allow us to suggest that metallic species should be present, Pd, Ni-Pd alloys, or Ni modified Pd, which must exist after reductive activation. The modifying Ni species could be partially remain in unreduced state, as strongly interacting NiO-support and NiAl 2 O 4 phases have both been reported in samples reduced at low temperatures [38]. TPD-NH 3 profiles of the support and Al 2 O 3 -supported Ni, Pd, and NiPd catalysts are shown in Fig. 5, exhibiting the characteristic two peaks traces that allow to distinguish between weak acid sites plus physisorbed ammonia (low-temperature peak between 40 and 260°C) and strong acid sites (high-temperature peak between 320 and 500°C) [40]. The numbers of strong acid sites present, normalized by the specific area (acid site density), are listed in Table 1. Regarding the acidity measurements, it must be stressed that any effects of the supported metals on these parameters would be limited to the ''shell'' region of the catalyst pellets, i.e., the ''core'' region acidbase properties would not be affected. In the TPD-NH 3 results, it can be seen a growth of signal intensity with increasing Ni/Pd atomic ratio. These results show that the introduction of Ni to Pd/Al 2 O 3 catalysts generally increases the number of acid sites (Table 1).
A textural analysis was carried out using N 2 physisorption techniques, the adsorption-desorption isotherms/pore size distributions and textural properties of the samples are shown in Fig. 6 and Table 1, respectively. First, all samples presented isotherms that could be classified as Type-IV with hysteresis loop Type-H2 according to IUPAC (Fig. 6). This isotherm type is attributed to mesoporous materials, and a hysteresis loop of this kind is characteristic of spheroidal cavities, voids between spherical particles closely packed or ink-bottle-shaped pores. In addition, the hysteresis loop ends at high values of relative pressure (P/P 0 = 0.6-0.7), which suggests the presence of a high number of large pores (mesopores and/ or macropores). These results are corroborated by the pore size distributions (inset Fig. 6) and average pore size (Table 1) of about 7 ± 1 nm in the range of mesopores. Furthermore, in Table 1, it can be seen that S BET of the catalysts is decreased (by up to about 13 %) when compared with the support. This effect can be attributed to the partial blockage of the pores, which, likewise, produces a decrease in V p .
Conventional TEM images of the Ni/Al 2 O 3 , Pd/Al 2 O 3, and 1NiPd/Al 2 O 3 catalysts are shown in Fig. 7a-c, respectively. All catalysts exhibited nanoparticles with pseudospherical morphology. Furthermore, the three catalysts presented a similar particle size, where the majority ([90 % of total count) have diameters \10 nm.
The effect of the increase in the Ni/Pd atomic ratio using Al 2 O 3 -supported catalysts was studied by monitoring of the  reactants and products in the selective hydrogenation of BD in the presence of BE under liquid phase conditions (Fig. 8). As the target of these reactions is to reduce the amount of impurities, with minimal losses of BE and avoiding the formation of BA, diolefin conversion as a reaction time function (Fig. 9a) and the selectivities to BE (desired product, Fig. 9b) and BA (undesired product, Fig. 9c) as a diolefins conversion function were compared with the Al 2 O 3 -supported monometallic catalysts based on Ni and Pd. Interestingly, when NiPd bimetallic catalysts were used, BD was selectively consumed toward the formation of trans-2-butene (t-BE) and BE in the first minutes of reaction (Fig. 8b-e). Equally, it was observed that the addition of Ni does not change the c-BE/t-BE molar ratio but decreases their formation. On the other hand, Pd/Al 2 O 3 catalyst yielded a higher conversion of diolefin than Al 2 O 3supported NiPd bimetallic catalysts (Fig. 9a) with a significant loss of BE (Fig. 9b) and the formation of BA (Fig. 9c) at conversions of BD higher than 60 %. In contrary, Ni/Al 2 O 3 catalyst presented a very low conversion (Fig. 9a). These results indicate that Ni-free catalysts pro-  The changes observed with the addition of Ni to the Pdbased catalysts, i.e., a decrease in the conversion of BD (Fig. 9a), increase in the selectivity to BE at medium conversions (Fig. 9b), and a decrease in selectivity to BA (Fig. 9c), could be assigned to geometric effect caused by dilution of the Pd atoms due to Ni addition or formation of a Ni-Pd bimetallic alloy that would be formed during reduction at 300°C (Fig. 4), and these cause the decrease in the inherent activity of the Pd atoms when the Ni/Pd atomic ratio is incremented; even this beneficial effect has also been observed in an Al 2 O 3 -supported NiPd catalyst (0.91 % Pd-1.51 % Ni/c-Al 2 O 3, where Ni/Pd atomic ratio *3) [26]. This geometric effect causes by dilution together with partial poisoning of Ni-Pd ensembles by firmly held adspecies, as carbonaceous deposits, was previously reported for NiPd/ SiO 2 catalysts [24]. However, other authors have proposed another theories to explain these performances that most importantly indicate that: (1) a favorable geometrical arrangement of surface atoms could occur during annealing of Pd atom deposit on Ni(111), favoring the adsorption of the reactants during hydrogenation [29], (2) a strain relaxation effect of the Ni-Pd surfaces of a Pd monolayer on Ni(110) also could contribute to alkene hydrogenation [41,42], or (3) self-poisoning or competitive adsorption between BEs and oligomers when Pd 2 Ni 50 Nb 48 ribbon is used as catalysts has also been reported [43]. Equally, additional studies has been   [43], Pd/Al 2 O 3 /NiAl (100) [46], and NiPd/Al 2 O 3 [25,26]. However, to the best of our knowledge, there is not published research on the changes of Ni-content in Pd-catalysts supported on Al 2 O 3 and their behavior in similar reactions.

Conclusions
A series of Pd-based catalysts promoted with various amounts of Ni for the selective hydrogenation of 1,3-butadiene in the presence of 1-butene under liquid phase conditions were investigated in a recirculation system with external fixed-bed reactor at 40°C and a total pressure of 200 psi. The most important results showed that the increase in Ni/Pd atomic ratio suppressed n-butane formation at relatively long contact time, increasing the recovery of 1-butene at middle conversion of 1,3-butadiene. The 1NiPd/Al 2 O 3 catalyst (Pd = 0.5 wt% and Ni/Pd atomic ratio = 1 supported on c-Al 2 O 3 ) presented the best catalytic performance. A dilution effect caused by Ni-Pd alloy formation could explain the beneficial influence on the catalytic activity for this reaction of great importance in the petrochemical industry.