Synthesis and Characterization of Pd over Novel TiO 2 Mixtures: Insights on Metal–Support Interactions

: Palladium nanoparticles were supported on unusual mixtures of anatase, TiO 2 (II) and rutile titania phases by wet impregnation, obtaining catalysts with metal contents of ca. 0.25 wt % labeled Pd/Ti5, Pd/Ti45, and Pd/Ti120. Crystalline structures were confirmed by X-ray diffraction. Pd particle sizes in the range of 4–20 nm were observed by scanning-transmission electron Microscopy. External surface areas (S BET ) in the range 10–17 m 2 g − 1 were higher enough to achieve a good distribution of palladium over titanium oxide outer surface, as evidenced by energy-dispersive X-ray spectroscopy elemental profiles. Pd 0 /Pd δ + atomic ratio measured by X-ray photoelectron spectroscopy showed a decrease from Pd/Ti5 to Pd/Ti120, in line with the decrease in anatase phase present in the catalysts. This behavior suggested that palladium tended to form more TiPd x O structures in Pd/Ti5 whilst PdO x structures were more likely to be present on supports with greater amounts of TiO 2 (II) and rutile, due to the distinct metal–support interactions. An increase in reducibility and oxygen mobility from Pd/Ti5 to Pd/Ti120 was observed by temperature programmed measurements and associated to the different high-energy ball milled supports. Catalysts with improved properties reported herein could exhibit an excellent performance in oxidation reactions, e.g., glycerol selective oxidation. inducing low-temperature reducibility and an increase in the mobility of reactive oxygen species, as evidenced by TPM analyses, factors well-known to benefit oxidation reactions. Present catalysts are being tested in liquid-phase glycerol selective oxidation in order to study the influence of metal– support interactions on the catalytic performance.


Introduction
Palladium-based catalysts have been extensively used in oxidation reactions, where both dispersion and particle size of metal are considered the dominant factors for a good activity and selectivity [1,2]. However, the role of supports is not limited to be carriers for metal particles [3,4]. Several times, the interface between metal and support is a better active site than the metal itself and thus, an appropriate selection of a catalyst support could reduce the amount of expensive metals needed (e.g., noble metals) [3,5,6]. Moreover, metal-support interaction is a well-known phenomenon attributed to electronic and geometric effects that induce charge transfer between metal particles and support [5,7]. These interactions might not only enhance the adsorption capacity of active sites but also inhibit the sintering of metal particles during reaction [4]. A rational design is therefore crucial to obtain sustainable metal-supported catalyst with high activity, reusability, and low rate of deactivation [3,7].
TiO2 is a recognized support for metals, which shows different properties depending on the polymorphs present [3]. Since most reactions in heterogeneous catalysis take place on the surface of metal active phase or the interface between metal and support, the presence of various crystalline phases could enhance catalytic properties [3,8]. To the best of our knowledge, catalysts with modified TiO2 supports formed by defective structures and unusual polymorphic mixtures have not been used for oxidation reactions until now. Therefore, we propose the study of Pd catalysts supported on novel mixtures of anatase, TiO2 (II) and rutile (not commercially available) obtained by high-energy ball milling in our previous work [9] to evaluate the different metal-support interactions and their potential role in liquid-phase glycerol selective oxidation.

Catalysts Preparation
Wet impregnation method was used to deposit palladium on three different TiO2 supports obtained in our previous work [9]. Aqueous solutions of HCl with appropriate amounts of PdCl2 (99.99%, Sigma Aldrich, Buenos Aires, Argentina) were added to titania supports in order to achieve palladium loadings of 0.25 wt % on a metal basis. Recovered solids were dried in an oven under vacuum at 70 °C overnight. Then, samples were calcined in a muffle furnace under air at 500 °C for 4 h using a heating ramp of 10 °C min −1 to eliminate the rests of metal precursor. Final Pd supported catalysts were labeled Pd/Tix, where x corresponds to the milling time of previously synthesized supports.

Results and Discussion
XRD analysis was performed to evaluate the crystalline structure of supported catalysts. We observed ( Figure 1A) that samples were formed of different mixtures of titania polymorphs. Pd/Ti5 showed intense reflections associated to anatase phase (PDF 00-21-1272) but also two broad and lowintensity reflections at 27.54 ° and 31.53 ° attributed to (110) rutile reflection (PDF 00-21-1276) and (111) high-pressure TiO2 (II) reflection (PDF 00-72-0021), respectively. These observations confirmed the presence of an almost pure anatase phase in this catalyst, as expected for its support [9]. On the other hand, Pd/Ti45 exhibited an increase in both TiO2 (II) and rutile reflections but a decrease in anatase reflections. Pd/Ti120 also showed the reflection attributed to TiO2 (II) phase and the appearance of several high-intensity rutile reflections. An incipient signal of (101) PdO reflection (PDF 01-06-0515) at 34.04° was detected in Pd/Ti5 catalyst. This behavior suggested the presence of larger palladium crystallites deposited on this catalyst, as also confirmed in further TEM analysis. The absence of this signal in Pd/Ti45 and Pd/Ti120 catalysts could be associated to smaller particles finely dispersed on the supports favored by the greater number of defects and the interfaces between polymorphs present in these supports and under the detection limit of the equipment [9,10].
We also made a rough calculation of the average crystallite size of catalysts by means of Scherrer equation and the relative intensities of the most intense titania reflections of each crystalline phase, (101) anatase, (110) rutile, and (111) TiO2 (II), as suggested by Dulian et al. [11]. We observed ( Table  1) a reduction trend according to the extent of milling in the supports, which is in line with the particle refinement caused by milling process [9,12]. These results were also in line with the increase in SBET values (Table 1) observed in a further analysis. We estimated the catalysts composition using the most intense Bragg reflections of each titania phase and the equations reported by Duvarci et al. [13], obtaining the unusual mixtures of titania phases (not commercially available) present in the supports (Table 1).
STEM and EDS analyses were carried out to have an insight on the size of palladium species. We could observe Pd particles of 20 nm in Pd/Ti5 catalyst ( Figure 1B) and in the range 4-6 and 5-10 nm for Pd/Ti45 and Pd/Ti120 respectively (figures not shown). These observations sustained previous XRD results, where PdO reflection could only be seen in Pd/Ti5 catalyst due to the bigger size of palladium particles. We also made a rough estimation of palladium dispersion by means of the equations provided by Mahata et al. [14] and the average sizes of palladium particles, obtaining values of 6%, 22%, and 15% of metal dispersion for Pd/Ti5, Pd/Ti45, and Pd/Ti120 respectively. From textural analyses, we observed specific surface area (SBET) values (Table 1) in accordance with the milling time of each titania support. It is noteworthy that these surface areas were lower than typical values for metals supported on reducible oxides (at least 50 m 2 g −1 ). However, the small amounts of palladium used during preparation (0.25 wt %) and the enhanced titania supports allowed a good metal dispersion, as observed by STEM-EDS previous analysis. Pore diameters varied in the range 9-21 nm and total pore volumes located in the range 0.04-0.1 cm 3 g −1 (Table 1) as expected for the low porosity of titania [8]. It is important to highlight that pore diameters obtained are higher enough to avoid internal diffusion limitations of big reactant molecules, such as glycerol (average molecule size of 0.62 nm) [15]. Furthermore, the low porosities observed may indicate that SBET values corresponded almost entirely to external surface area [6,9]. This fact could induce the preferential location of palladium species on the external surface of supports, favoring the access of reactants to start the mechanism of reaction [6]. Morphology of catalysts was studied by SEM (micrographs not shown). Sample Pd/Ti5 showed particles with spherical shape and diameters between 120-220 nm and Pd/Ti45 catalyst presented grape-like clusters formed by a couple of particles, with average diameters between 100-200 nm. Observed agglomeration might be a consequence of electrostatic effects on fine particles (with increased surface energy) as reported by other authors [16]. Smaller sizes were found in case of Pd/Ti120 within a range between 80-180 nm. Moreover, the smaller particle sizes correlated well with the increase in SBET values (Table 1) from Pd/Ti5 to Pd/Ti120. Pd loadings measured by ICP-OES were found to be in the range 0.2-0.36 wt % compared to the theoretical 0.25 wt %. Differences were attributed to experimental errors during synthesis and/or digestion of the samples.
Surface oxidation state and metal-support interactions were studied by means of XPS (Table 2). At the Ti 2p level we found Ti 2p3/2 and Ti 2p1/2 doublets normally located at binding energies of 458.3 and 464 eV and attributed to Ti 4+ [9]. No signals associated to Ti 3+ were observed. XPS spectra of O 1s were deconvoluted into two signals (ca. 529.4 and 530.1 eV) associated to lattice oxygen (Olat) and oxygen adsorbed in oxygen vacancies (Oads) respectively [17]. The contents of oxygen vacancies Oads/(Olat + Oads) were in the same order as observed in the supports, with a slight variation due to palladium impregnation and further calcination [9]. These low values were previously explained by the migration of oxygen vacancies from the surface to the bulk in order to achieve polymorphic transformation during the synthesis of supports [9]. At the Pd 3d level we observed Pd 3d5/2 and Pd 3d3/2 doublets which were deconvoluted into four peaks. We selected Pd 3d5/2 signals for the interpretation of palladium oxidation state ( Table 2). It is well-known that Pd 3d5/2 level binding energy for Pd 0 is around 335.2 eV [2,6]. Signals with positive shifts respect to the Pd 0 signal are associated to cationic species (Pd δ+ with δ ≥ 2) and/or Pd/PdOx interfaces, while negative shifts are attributed to anionic species and structures such as TiPdxO, where palladium state could be more negative than zero [6]. Pd 0 /Pd δ+ atomic ratio (Table 2) showed a decrease from Pd/Ti5 to Pd/Ti120, in line with the decrease in anatase phase present in the catalysts. This behavior could be explained by the greater number of defects and oxygen mobility in the supports milled for longer time, that would lead to electron-deficient palladium species due to the transfer of electron density from palladium to titanium [1,2,4]. Pd/Ti at % were in all cases higher than the theoretical calculation for the bulk (0.19 at %), which confirmed the good dispersion of Pd on the external surface of supports [3].
H2-TPR measurements were performed to further study metal-support interactions and redox properties of catalysts. Signals observed at around 52 °C (Figure 2A) in all catalyst were associated to the formation of β-PdHx species on surface metallic palladium, that takes place under reduction at ambient temperature [6]. Negative signals observed at around 72 °C were attributed to the decomposition of formed hydrides, releasing the H2 [1]. The broad reduction signal with a maximum at 260 °C and the shoulder at 138 °C observed in this catalyst was associated to the reduction of PdO strongly interacting with the titania support [18]. In case of Pd/Ti45 and Pd/Ti120, well-defined signals at 141 °C and 160 °C (respectively) were attributed to the enhanced reduction of PdO species. This increase in reducibility, was related to the greater number of defects, lattice oxygen mobility, and amount of PdOx species present in the catalysts [1]. Strong signals found in the range 533-578 °C were ascribed to the reduction of surface TiO2 particles, as observed previously in the supports [9]. Pd/Ti120 also showed shoulders at 307 °C and 558 °C which could be associated to the reduction of Ti 4+ species located in defective structures of the polymorphs, due to the longer extent of milling of this catalyst support [9]. H2-uptakes ( Figure 2A) increased from Pd/Ti5 to Pd/Ti120. This behavior was expected due to the longer milling time of the supports, which generates structures with improved oxygen diffusion and reducibility [9].
We performed O2-TPD measurements ( Figure 2B) to obtain information regarding oxygen mobility in the catalysts. Desorption signals found below 100 °C were related to oxygen species weakly bonded to the surface, as reported by other authors [19]. Signals in the range 174-220 °C were attributed to oxygen species adsorbed on oxygen vacancies [19]. Furthermore, signals around 450 °C were ascribed to surface lattice oxygen while desorption peaks above 500 °C were associated to bulk lattice-oxygen release [19]. Well-defined signals observed in case of Pd/Ti45 and Pd/Ti120 catalysts, with more cationic species, were related to the desorption of more active-oxygen species from the surface of its more defective supports [9]. Weight-corrected areas (Figure 2 B) followed the same trend as H2 uptakes, increasing from Pd/Ti5 to Pd/Ti120. This fact could be ascribed to the greater amount of PdOx species present in these catalysts, in line with our previous XPS and H2-TPR analyses.

Conclusions
Pd-based catalysts reported in this work showed distinctive properties associated to the unusual mixtures of titania phases present in the supports, previously prepared by high-energy ball milling [9]. Metal nanoparticles were well dispersed on the outer surface of TiO2, as confirmed by XPS and STEM-EDS analyses. Pd/Ti5 tended to form more anionic Pd species in the form of TiPdxO structures, which seemed to be less exposed on the surface due to strong interaction with the support, whilst Pd/Ti45 and mainly Pd/Ti120 formed more cationic PdOx species located on the surface of catalysts, as observed by XPS measurements. In the case of glycerol oxidation, cationic species of palladium would benefit the adsorption of hydroxides and successive glycerol deprotonation, accelerating the mechanism of reaction. Achieved metal-support interactions could improve redox properties, inducing low-temperature reducibility and an increase in the mobility of reactive oxygen species, as evidenced by TPM analyses, factors well-known to benefit oxidation reactions. Present catalysts are being tested in liquid-phase glycerol selective oxidation in order to study the influence of metalsupport interactions on the catalytic performance.