Tuning Catalytic Properties of Supported Bimetallic Pd / Ir Systems in the Hydrogenation of Cinnamaldehyde by Using the “ Water-in-Oil ” Microemulsion Method

Supported Pd/Ir bimetallic catalysts were synthesized by the “water-in-oil” microemulsion method at different precursor concentrations and characterized by XRD, XPS, SEM, TEM, and cyclic voltammetry. Depending on the preparation conditions, formation of bimetallic catalysts with different metal segregation and surface composition can be easily obtained, thus tuning the bimetallic structure of catalysts as well as their relative catalytic properties. Bimetallic Pd/Ir systems were efficiently tested in the hydrogenation of cinnamaldehyde showing a better performance than analogous monometallic catalysts.


Introduction
e synthetic procedure is surely one of the main key factors determining morphology and surface properties of heterogeneous catalysts [1].In the last decades, a lot of attention has been addressed to the preparation of supported bimetallic systems [2][3][4][5][6][7] that, at present, represent one of the most important categories of heterogeneous catalysts being involved in many types of reactions including oxidation, hydrogenation, dehydrogenation, hydrogenolysis, aqueous phase reforming (APR), and H-transfer reactions [8][9][10][11][12][13].In order to obtain bimetallic structures with a controlled surface composition, several synthetic methods have been proposed [11,[14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30].Among the many known, the microemulsion (ME) synthesis method seems to be one of the most promising synthetic procedures for the preparation of bimetallic particles due to the possibility of controlling the particle size and the composition (alloy, core-shell, and enriched core/enriched shell) [28,31,32].In the ME method, small amounts of an aqueous solution containing metal ions are added to the surfactant/oil phase system (micellar solution).Under these conditions, micelles are formed spontaneously and, by introducing a reducing agent, their nucleation is followed by the growth of the bimetallic particles [33].Nevertheless, the final bimetallic structure is conditioned by many factors such as the composition of microemulsions [34], the type of precursor [20,35], the difference of standard reduction potentials of metal ions (ΔE 0 , V) [28,36,37], and the concentration of precursor solutions [36,38,39].Many examples of bimetallic structures have been reported by using the microemulsion method including Pt/Ag [37], Pd/Ag [40], Au/Pt [38,41,42], Pd/Pt [43], Pd/Au [44,45], Pd/Ni [46], Pt/Cu [47], and Pt/ Co [48] systems.At present, there is still no insightful research available on Pd/Ir bimetallic structures obtained by the ME method.However, Pd/Ir systems are promising catalysts for hydrogenation [49], reduction [50] oxidation, and electro-oxidation [51][52][53] reactions.A synergy between Pd and Ir has been reported by Nakagawa et al. [54] in the silica-supported PdIr (1 : 1) alloy particles in the total hydrogenation of furfural.Accordingly, Zlotea et al. [55] reported that the catalytic performance of the CO oxidation could be enhanced when Ir is added to Pd.A significant promotion of Pd by Ir in nitrobenzene hydrogenation has been reported by Yang et al. [56] in mesoporous silica supporting Pd/Ir catalysts where the promotional effect of Ir has been related to its interaction with Pd and the consequent charge transfer from Ir to Pd.Some of the authors report that despite the large difference of reduction potentials (Ir � 1.19 V; Pd � 0.48 V), PdIr particles (Pd/Ir molar ratio � 10) of almost homogeneous alloy microstructure can be formed by using the ME method.Moreover, Ir has a higher surface energy and heat of sublimation (3.00 J/m 2 and 195 kJ/mol, respectively), and the values are higher than those of Pd (2.05 J/m 2 and 136 kJ/mol, respectively) [57,58].Both values lead to core-shell structure formation, with the core enriched in Ir and the shell enriched in Pd [17].
In the present work, supported (carbon and SiO 2 ) Pd/Ir bimetallic catalysts are synthesized by the ME method with a Pd 74 Ir 26 atomic ratio.eir morphological and structural characteristics were determined by XRD, XPS, and electron microscopy (SEM and TEM).Surface properties of PdIr particles, including Pd and Ir surface fractions, and the hydrogen sorption/desorption ability were characterized by means of cyclic voltammetry (CV) measurements.Catalytic tests were performed to gain more insights into Pd/Ir surface properties, taking into account, in particular, the role of Ir in promoting the modification of Pd reactivity.e hydrogenation of cinnamaldehyde was chosen as a model reaction, highlighting the properties of the catalyst surface, such as the distribution/type of the bimetallic structure as well as its role in the activity/selectivity pattern [59][60][61][62].

Catalysts Preparation.
Unless otherwise specified, all reagents were purchased from Sigma-Aldrich.Carbon (Vulcan XC72, CABOT) and silica (Davisil 634) were used as supports.
e monometallic Pd and Ir catalysts (2 wt.% metal loading) and the bimetallic Pd/Ir systems (2 wt.% of Pd metal and 1.3 wt.% of Ir metal loading) were prepared by the reverse "water-in-oil" microemulsion method as previously described [22,55].Polyoxyethylene (78) and octylphenylether (Triton-X 114) were used as surfactants and cyclohexane as the organic solvent, and the W � 5.5 was kept as the molar ratio of water/surfactant.e 0.02, 0.04, 0.08, 0.14, and 0.2 M aqueous solutions of IrCl 3 were prepared by using a commercial reagent.
e precursor solutions of Pd 2+ (PdCl 2 ) ions at different concentrations 0.02, 0.04, 0.08, 0.14, and 0.2 M were prepared using NaCl (dried at 120 °C) at a molar ratio of NaCl : PdCl 2 � 2. e synthesis of 0.5 g of 2% Pd/C and 0.5 g of 2% Pd/SiO 2 was carried out as follows: 0.94 cm 3 of a 0.2 M solution containing Pd 2+ ions and 16.8 cm 3 of a cyclohexane solution of the surfactant (0.62 M) were stirred until the microemulsion was transparent.en, NaBH 4 (in a 10 : 1 molar ratio of reducing agent/metal ions) was added and the stirring was continued for another hour.
e microemulsion was then divided into two parts, and 0.49 g of each support (carbon or SiO 2 ) was added under stirring for another hour.e deposition of metal particles on the supports was made by introducing the THF solvent.e ratio of the volume of added THF to the volume of the microemulsion was 3 : 1.Finally, the catalysts were filtered, washed with acetone, and dried.
All PdIr supported catalysts were synthesized using a total metal loading of 3.3 wt.% (Pd + Ir) and a Pd : Ir atomic ratio of 74 : 26 (Pd 74 Ir 26 ).During the preparation, a mixture of PdCl 4 2− and IrCl 3 solutions having different molar concentrations and cyclohexane as well as surfactant were mixed.
SEM studies were performed with a field emission scanning electron microscope (JEOL JSM-7500 F) equipped with the X-ray energy dispersive system (EDS).
TEM studies were performed on an FEI Tecnai G2 transmission electron microscope operating at 200 kV equipped with EDX analysis.
e cyclic voltammetry (CV) measurements were made by using a CH Instrument (Austin, TX, USA) Model CHI760D workstation with a scan rate of 50 mV using a 0.5 M solution of H 2 SO 4 .

Catalytic Tests.
Catalytic reactions were performed in a batch glass reactor at a temperature of 50 °C and a constant atmospheric pressure of hydrogen (20 cm 3 of a cinnamaldehyde solution in toluene having 0.05 mol/dm 3 of c 0 and 5 g/dm 3 of catalysts was used).e methodology has been described in detail previously [22,55].In all hydrogenation reactions, silica-supported catalysts were employed in order to exclude the fact that the electronic properties of the carbon support may influence the activity/selectivity of samples, so that the catalytic properties only derive from those of the system [63].
e catalyst and the solution were put in the reactor without addition of reagents in order to allow the in situ catalyst activation by hydrogen at a temperature of 50 °C before the hydrogenation test.en, the catalyst was mixed with the reagent solution and the experiment started.e progress of the reaction was measured from the consumption of hydrogen.Samples were withdrawn from the reactor via a tube at intervals of time and analyzed by GC.  200), (220), and (311) crystal planes, respectively, which indicate a typical cubic crystal (fcc) structure of Pd [64].e crystalline Ir also has an fcc structure, but no Ir diffraction peaks were observed in the 2% Ir/SiO 2 system. is may be caused by the low Ir loading, the amorphous distribution of the metal, or the presence of small crystallites of iridium.In addition, on analyzing the PdIr-0.2/SiO 2 catalyst, four diffraction peaks were detected at higher 2θ values compared to those of Pd/SiO 2 .is may be related to the presence of a crystalline structure with a lower d space relative to Ir species (d 111 � 2.217) compared with that of Pd (d 111 � 2.246) and the subsequent formation of a Pd-Ir alloy [65].

Results and Discussion
Accordingly, peaks of PdIr-0.02,PdIr-0.04,PdIr-0.08,PdIr-0.14, and PdIr-0.2 are presented in Figure 2, and it can be observed that the one peak located around 2θ � 40 °is slightly shifted toward higher 2θ values with respect to that of Pd/SiO 2 , indicating that some Ir atoms enter the Pd lattice, forming a PdIr alloy as well.Reduced lattice parameters in all PdIr crystallites (Table 1) compared to those of pure Pd confirm a PdIr alloy formation [66].Moreover, the shift differences in all PdIr catalysts may indicate a different surface and volume composition, related to the amount of Ir incorporated into Pd. is effect may depend on the concentration used in the precursor synthesis.In all XRD patterns, peak characteristics of IrO 2 , 2θ � 34.5 °and 54 °, were not observed [67].e average crystallite sizes of all prepared catalysts were estimated using the Scherrer equation [68] and are presented in Table 1.PdIr catalysts have a smaller particle size than the monometallic Pd.
SEM images indicate that, for PdIr bimetallic catalysts, particles are almost dispersed over the support (Figure 3).However, on increasing the concentration of precursors, more aggregates are visible, as confirmed by the TEM analysis.A similar effect was also observed by Liz-Marzán and Lado-Touriño [70] during the formation of Ag nanoparticles: when the AgNO 3 concentration rises, the mean diameter of silver nanoparticles decreases if AgNO 3 ≤ 0.2 M and increases if AgNO 3 > 0.2 M. At a higher concentration, silver nanoparticles begin to aggregate and form large particles.
is phenomenon has been explained by considering the formation of a large number of silver nanoparticles at a higher concentration of AgNO 3 .As a result, the collision frequency increases significantly, and this neutralizes the surfactant molecules' protective effect and results in a higher aggregation.
Particles size distribution, as determined by the TEM analysis, indicates a narrow range of 2-7 nm and the average particles size was calculated by counting over 100 particles for each catalyst (Figure 4).It can be observed that as the concentration of the precursor increases, a very slight particle-size increase is noticed: PdIr-0.02(4.4 nm), PdIr-0.08 (4.5 nm), and PdIr-0.2(4.8 nm), and it should be pointed out that the average size of the metal particles, calculated from the microscopic images, is consistent with the XRD results.Comparing these data with the particles size of Pd (6.0 nm) and Ir (4.2 nm) monometallic particles [69], it can be inferred that after addition of Ir, the size of bimetallic particles decreases and reaches the value close to that characteristic of Ir. is metal-type particle-size effect is the result of the main role of the nucleation process, being decisive for the particles' formation in the microemulsion medium.e nucleation process is determined by the metal properties, namely, the heat of metal vaporization ΔH VAP (or sublimation, ΔH S ) that determines the strength of the (M-M) metal-metal bond and the size of the critical nucleus, in accordance with previous results concerning interactions of Ir with Pt and Rh with Pd [71].e Pd/Ir atomic ratio, determined by EDS analysis (Table 1), confirms the presence of both Pd and Ir metals in the nanoparticles in agreement with the nominal composition.
e catalysts were also examined by X-ray photoelectron spectroscopy (XPS) experiments.e surface composition of Pd and Ir (in at.%) and the binding energies of palladium Pd 3d 5/2 and Ir 4f 7/2 peaks are included in Table 2.
e binding energy of Pd 3d 5/2 in all PdIr catalysts is slightly shifted to a higher energy value compared to the metallic state of Pd (335.3 eV) [44].PdIr-0.02shows the largest binding energy shift ∆E � 0.5 eV, while in other PdIr catalysts ∆E values are in a range of 0.1-0.4eV. e larger shift of the binding energy for the PdIr-0.02system can be ascribed to the strongest surface electronic modification of the Pd derived from its interaction with Ir [53].e presence of Ir on the surface of bimetallic particles is also observed in all catalysts.In the PdIr-0.02system, the atomic ratio, Pd to Ir (1.95/1.03), is 1.89, while in the PdIr-0.2, it is close to 1, suggesting a stronger surface enrichment in Ir atoms compared to that in the PdIr-0.02,obtained from the most diluted precursor.Nevertheless, the electron modification of the PdIr-0.02catalyst is more pronounced (BE shift � 0.5 eV). is effect may suggest a stronger dispersion of Ir on the surface, which results in a stronger modification of Pd in the presence of the more dispersed Ir.
e catalysts were also characterized by cyclic voltammetry (CV) in order to examine more accurately the structure of bimetallic particles.Voltammograms, recorded for all catalysts, are shown in Figure 5.   Peaks corresponding to adsorption and desorption of hydrogen are clearly detected in bimetallic systems together with a shift of the hydrogen desorption component (magni cation, area A of Figure 5) towards lower potentials for all studied catalysts compared to pure Pd samples.Chen et al. [52] attributed the e ect of the negative hydrogen desorption peak shift to the weakened adsorption energy of hydrogen.is is a consequence of the downshift of the Pd d-band centre related to the presence of Ir on both the surface and the bulk of the bimetallic structure.e shape of hydrogen desorption peaks is also changing, and this conrms the modi cation of the Pd electronic properties by adding Ir [51,[72][73][74].erefore, it can be assumed that, in each of the investigated catalysts, Ir is present on the surface of PdIr bimetallic particles, a ecting the Pd-Ir interaction and thus in uencing the degree of coverage by hydrogen in the surface.e presence of the second metal on the surface, as well as the electron interaction between both components, was also con rmed by XPS studies.
For the Pd catalysts, a reduction peak of PdO is observed at the potential of E 0.727 V and this is consistent with the reduction of surface-oxidized Pd(II) species [75,76].For all PdIr catalysts, only one reduction peak of oxidized forms of metals is visible, having a di erent potential than that of Pd/ C. Curves relative to PdIr catalysts do not show separate peak characteristic of separate reduction of oxidized forms of Pd and Ir, indicating the homogeneity of the alloy surface composition [77].
e surface composition of bimetallic particles (Pd and Ir surface fraction) was calculated from formula (1), including the linear correlation between the peak reduction potential and the surface composition (E P,Pd , E P,Ir are oxide reduction peak potentials for pure Pd (0.727 V) and Ir (0.505 V), and X Pd and X Ir are surface fractions of Pd and M) [77]: is method was, so far, used by many authors for the analysis of bimetallic lms, as well as for the determination of the surface composition of particles in PdPt/C, PdAu/C, PdIr/C, and PdPtAu/C catalysts [69,78,79].Surface fractions of Ir and Pd calculated from equation ( 1) are summarized in Table 2.
For all catalysts, the surface fraction of Ir is smaller (0.15-0.23) than that of the bulk composition (0.26 for the Pd 74 Ir 26 sample), indicating the palladium enrichment of the surface and the segregation of Ir in the core of particles.Nevertheless, the e ect of the surface increase of iridium on increasing the precursor concentration is evident.e lowest surface fraction is observed in the case of the PdIr-0.02catalyst, while for the catalyst synthesized from the precursor solution with the highest concentration (PdIr-0.2), the surface composition of Ir is almost the same as that of the bulk composition (0.23).Moreover, by analyzing the hydrogen desorption peak shapes (Figure 4, area A) of the catalysts, it can be concluded that for those systems, obtained from the precursor having the lowest concentration (PdIr-0.02), the shape signi cantly di ers than those in catalysts obtained from precursors with higher concentrations (PdIr-0.04,PdIr-0.08,PdIr-0.6 Journal of Chemistry that there is a different distribution (mixing in atomic-level alloy, nonatomic alloy, and atomic clusters) of Ir on the surface of bimetallic structures and that there are possible different interactions of Pd-Ir.Monte Carlo simulations [80,81] and experimental works [38,44,47,82] have also shown that the difference in the reduction potential of precursor ions, in the w/o microemulsion, may affect the type of bimetallic structures (core-shell, alloy, and cluster in cluster).In general, the higher is the difference in reduction potentials, the stronger is the segregation of metals and formation of core-shell structures is observed.It has been also reported that the resulting bimetallic arrangement reflects the combination of two potentially limiting factors: the reduction rate of each metal ion and the intermicellar exchange rate [83].Reduction potentials relative to 0.2 M PdCl 4 2− and IrCl 3 precursor concentrations have been previously determined by the linear voltammetry [69], exhibiting a significant difference (ΔE > 0.7 V).According to the literature, formation of core (Ir)-shell (Pd) structures is the most probable.Data relative to all bimetallic PdIr catalysts show that the structures are similar to those having an incomplete coreshell structure.
erefore, it can be concluded that the difference in reduction potentials is not the only decisive factor for the type of bimetallic structure formed following the w/o microemulsion route.Monte Carlo simulations reveal, in fact, that differences in nucleation and growth can also influence the resulting structures [84].Calculations show that in the case of heterogeneous nucleation, when the nucleation process is decisive, smaller heterogeneous critical nuclei compared to the homogeneous ones should prefer the final particles to form an alloy structure.ey also demonstrate that when the formation of heterogeneous nuclei is faster than the homogeneous nucleation, the degree of metal components alloying in the final particles is greater.It can therefore be suggested that, by changing the concentration of the precursor solutions, the nucleation processes may be "modified" and play a more decisive role in comparison to the processes resulting from the difference in reduction potentials.

Catalytic Hydrogenation of Cinnamaldehyde.
In order to verify the effect of both different iridium surface fraction and distribution, the catalytic hydrogenation of cinnamaldehyde was carried out.Simplified possible cinnamaldehyde hydrogenation pathways are presented in Scheme 1.
e catalytic test shows that hydrogenation of cinnamaldehyde leads to hydrocinnamaldehyde (HCAL), cinnamyl alcohol (COL), and the corresponding saturated alcohol (HCOL).β-Methyl styrene and phenyl propane have not been detected under the reaction conditions adopted.It is known that Pd is active and very selective towards C�C hydrogenation [85].Ir is much less active and generally selective towards C�O hydrogenation [86].In addition, monometallic Ir does not exhibit any observable activity relative to CAL hydrogenation under the studied reaction conditions.e activity relative to all catalysts is expressed by the initial rate of hydrogenation (R, related to the catalyst mass) and TOF (s −1 ) values calculated with the formula given by Long et al. [87].Results are reported in Table 1: the highest TOF value (Figure 6) belongs to the PdIr-0.02catalyst (0.138 s −1 ) that is higher than the one relative to the monometallic Pd catalyst (0.121 s −1 ).Under the adopted reaction condition, no catalytic activity of the monometallic Ir catalyst was observed.
Furthermore, the activity decreases on increasing the precursor concentration used in the synthesis, as inferred from the highest TOF value, detected for the sample PdIr-0.02(0.138 s −1 ) compared to the lowest one, calculated for the PdIr-0.2catalyst (0.024 s −1 ).All bimetallic catalysts, except PdIr-0.02,exhibit also a minor activity than the monometallic Pd catalyst.Yang et al. [56] observed a strong increase of the activity in the hydrogenation of nitrobenzene in the presence of Pd catalysts after the addition of small amounts of Ir (Pd/Ir molar ratio of 0.1). is was explained by the size decrease of bimetallic particles and the simultaneous increase of both active sites and Pd electron modifications in the presence of Ir.In our catalytic systems, it can be suggested that the effect of the particle size on the activity and selectivity can be excluded as all PdIr particles, regardless of the concentration, are in a similar, very narrow size range (4.4-4.9 nm) [56].Moreover, the effect of the surface metal composition on the selectivity of the CAL hydrogenation is reported in Figure 6 and shows the selectivity to the saturated alcohol (HCOL) obtained at 80% CAL conversion.
e monometallic Pd catalyst has a selectivity to HCOL of around 12%.Of all the bimetallic catalysts examined, the PdIr-0.02catalyst exhibits the highest selectivity towards the C�O hydrogenation (about 35%), which is nearly three times higher than the one calculated with the monometallic Pd.It can also be observed that as the concentration of the precursor used in the synthesis increases, the selectivity to HCOL decreases and is about 18% on using the PdIr-0.2.A remarkable PdIr synergy has been observed also by Nakagawa et al. [54] in the silica-supported Pd-Ir (1/1) alloy particles in the total hydrogenation of furfural.
e activity/selectivity enhancement of active sites formed in the PdIr alloys has been related to the presence of Ir that may promote the adsorption of C�O of furfural accompanied by a weaker adsorption of the furan ring.
In our catalytic systems, it can be suggested that differences in both activity and selectivity may result from a different promotional effect of Ir, which, in turns, depends on the surface architecture of the PdIr species.ere is, in fact, a clear tendency of the surface growth of the Ir fraction from 0.15 to 0.23 with the increasing concentration of the precursor solution (CV studies).Furthermore, in all PdIr catalysts, Ir is present on the surface; however, in the catalyst obtained from the lowest precursor concentration, the fraction of Ir is the lowest.Analogous results were also obtained with Monte Carlo simulations given by Tojo et al. [88], where the precursor concentration influence (0.01 M-0.40 M) on the bimetallic Au/Pt structure was studied, showing, theoretically, that the higher concentration of precursor with a faster reduction rate should afford a better core-shell structure.In our PdIr catalysts,

Journal of Chemistry
Ir is present on the surface and for the one obtained from the lowest precursor concentration, the fraction of Ir on the surface is the lowest.Moreover, it has also to be pointed out that, on using 0.02 M-0.2 M precursor concentrations, a complete core (Ir)-shell (Pd) structure is not obtained.erefore, Ir is the most dispersed on the surface of bimetallic particles synthesized from the 0.02 M precursor (PdIr-0.02),and this results in a strong Pd-electron modi cation (XPS studies).On the other hand, a strong decrease in activity, on increasing precursor concentrations, may both be derived from increasing the surface area of Ir or/and a weaker dispersion of Ir (formation of clusters on the surface) compared to Pd. is may result in a weaker electron modi cation of Pd (XPS studies) and smaller number of available Pd active centres.

Conclusions
Supported PdIr bimetallic catalysts were synthesized by the "water-in-oil" microemulsion method with di erent precursor concentrations.e concentration of the precursor solutions (0.02 M, 0.04 M, 0.08 M, 0.14 M, and 0.2 M) has a signi cant in uence on the type of PdIr bimetallic particles.XRD, TEM, and cyclic voltammetry (CV) studies con rm the formation of bimetallic structures containing Pd and Ir with di erent Ir fractions on the surface.For the PdIr-0.02catalyst obtained from the most diluted precursor solution (0.02 M), the highest activity in the hydrogenation of cinnamaldehyde and the highest selectivity in the hydrogenation of the C O bond were demonstrated.It can therefore be concluded that not only the di erence in reduction rates of both metals but also synthesis conditions may play an important role in the process of nucleation, growth, and formation of bimetallic particles.T. S. acknowledges nancial support received under the action Funds for Young Researchers organized by the Director of Jerzy Haber Institute of Catalysis and Surface Chemistry, Polish Academy of Sciences.

Table 1 :
Physicochemical properties of investigated catalysts: metal particle characterization (average size, the overall composition measured by EDS, and the local composition measured by STEM-EDX methods) and the initial rate of cinnamaldehyde hydrogenation.CatalystMetal particle size d (nm) Lattice parameter ( Å) Pd : Ir atomic ratio Initial rate R (mol• CAL 10 −5 •min −1 •g (cat)

Figure 5 :
Figure5: Cyclic voltammograms for mono Pd and bimetallic PdIr catalysts in N 2 -saturated 0.5 M H 2 SO 4 solution with a scan rate of 50 mV/s.e insets are magni ed views of the hydrogen desorption region (0.0-0.4 V) and the metal surface reduction region (0.5-0.9 V).

Figure 6 :
Figure 6: Selectivity to saturated alcohol (HCOL) obtained at 80% conversion of CAL with the TOF values for Pd and PdIr catalysts.

Table 2 :
XPS data and surface fraction of Pd and Ir components determined by CV measurements.