Si
Chen
a,
Xiaohui
Huang
bc,
Dieter
Schild
d,
Di
Wang
be,
Christian
Kübel
bce and
Silke
Behrens
*a
aInstitute of Catalysis Research and Technology (IKFT), Karlsruhe Institute of Technology (KIT), Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany. E-mail: Silke.Behrens@kit.eu
bInstitute of Nanotechnology (INT), Karlsruhe Institute of Technology (KIT), Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany
cJoint Research Laboratory Nanomaterials, Technische Universität Darmstadt, Jovanka-Bontschits-Straße 2, 64287, Darmstadt, Germany
dInstitute for Nuclear Waste Disposal (INE), Karlsruhe Institute of Technology (KIT), Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany
eKarlsruhe Nano Micro Facility, Karlsruhe Institute of Technology, Hermann-von-Helmholtz-Platz 1, 76344, Eggenstein-Leopoldshafen, Germany
First published on 23rd November 2022
Intermetallic nanoparticles (NPs) are highly interesting materials in catalysis due to their geometrically ordered structures and altered electronic properties, but the synthesis of defined intermetallic NPs remains a challenge. Here, we report a novel and facile approach for the synthesis of intermetallic Pd–In NPs in ionic liquids (ILs) at moderate temperatures. Depending on the molar ratio of the metal precursors and the reaction temperature, single-phase Pd3In, PdIn and Pd3In7 NPs were obtained, which was confirmed, e.g. by powder X-ray diffraction, electron microscopy, and optical emission spectroscopy with inductively coupled plasma. The Pd–In NPs stabilized in ILs were used as catalysts in the liquid-phase semi-hydrogenation of diphenylacetylene (DPA). Highly ordered PdIn NPs with a CsCl type structure revealed both high activity and selectivity to cis-stilbene even at full DPA conversion. Intermetallic compounds such as PdIn can be used to isolate contiguous Pd atoms with another base metal into single Pd sites, thereby increasing the catalytic selectivity of Pd while stabilizing the individual sites in the intermetallic structures. This work may provide new pathways for the synthesis of single-phase intermetallic NPs and future insights into a more rational design of bimetallic catalysts with specific catalytic properties.
It is well-known that bimetallic nanoparticles (NPs), which can be present as alloys,9,10 core–shell,8 or (segregated) hetero structures, generally exhibit unique physico–chemical properties and enhanced catalytic performance compared to their monometallic counterparts due to synergistic effects between the two metals.11–13 Many efforts have been devoted to synthesize and optimize Pd catalysts by bimetallic Pd–M alloys (e.g. with M = Cu,14 Ag,15–17 Au,18,19 Zn,20 Ga,21 In,22–24 Sn,25,26 or Bi)27 or even ternary Pd-based phases (such as Pd–Bi–Se28 or Pd–Ga–Sn),29,30 improving catalytic selectivity for the selective hydrogenation of alkynes. In addition, non-precious, binary alloys (Ni/Fe),31 intermetallic compounds (Al13Fe4),32 transition metal-free Zintl phases (BaGa2), or non-precious, ternary phases (Cu–Ni–Fe)33 have been reported as interesting catalytic materials for alkyne semi-hydrogenation in the gas or liquid phase. In general, alloys with random-type structures are prone to changes in response to different reaction conditions, including reconstructions, segregation and oxidative/reductive evolution, thus resulting in reduced catalytic activity and stability.34 In this direction, nanoparticles of intermetallic compounds have attracted extensive attention, because they combine specific crystal structures with long-range atomic ordering and altered electronic structures caused by chemical bonding.35–39 For example, Armbrüster et al. reported that unsupported intermetallic PdGa and Pd2Ga NPs show high selectivity (77% and 60%) and long-term stability for the semi-hydrogenation of acetylene.40 As an electronic analog of Ga in group 13 of the periodic table, alloying In with Pd-based catalysts was reported as a promising alternative to the Pd–Ga systems. Similar to Pd–Ga systems, Pd–In catalysts have isolated Pd sites on the catalyst surface, but have higher stability and oxidation resistance (oxidative potentials ∼−0.3 V (In) and ∼−0.5 V (Ga)). Even at the molecular level, heterobimetallic Pd–In complexes represent interesting homogeneous catalysts in the hydrogenation of alkynes.41 For example, Pd(μ-O2C(CH3))4In(O2C(CH3)) was demonstrated to homogeneously catalyze the liquid-phase hydrogenation of phenylacetylene and styrene in acetic acid, dimethylformamide, and ethyl acetate. Pd metal was not formed until complete hydrogenation to the alkene, with phenylacetylene and styrene acting as stabilizing π ligands for the active form of the complex. The defined 1:
1 stoichiometry of the Pd–In complex was further exploited in incipient wetness impregnation for generation of PdIn particles on various supports.22,42 The impregnated complexes were treated in several steps by calcination at high temperatures (550 °C), followed by H2 reduction upon 400 °C to obtain particles of the desired intermetallic PdIn phase with high selectivity in gas- and liquid-phase semi-hydrogenation of alkynes.22,43,44 The stoichiometry was relevant with respect to alkyne semi-hydrogenation. For gas-phase acetylene semi-hydrogenation, Feng et al. demonstrated that intermetallic PdIn NPs with isolated, single-atom Pd sites exhibited higher selectivity than Pd3In NPs, which was further supported by density functional theory (DFT) calculations.45 To the best of our knowledge, however, so far there have been no studies on defined particles of other Pd–In phase compositions in the liquid-phase semi-hydrogenation of diphenylacetylene.
In general, the Pd–In system can form several intermetallic compounds with different Pd/In stoichiometry, including Pd3In, Pd2In, PdIn, Pd3In7, Pd2In3, and Pd5In3, and it is challenging to achieve single-phase particles.46 In addition, thermal annealing (typically over 500 °C) of dry NP powders is typically required to obtain compounds with the ordered intermetallic structure, while such high temperatures may easily cause NP aggregation/sintering with an increase of particle size.47,48 In this context, we make use of the unique physico–chemical properties of room temperature ionic liquids (ILs) as tunable and neoteric reaction medium (with low-volatility, thermal stability, high polarity, high conductivity and solubility) for the synthesis of intermetallic Pd–In NPs at moderate temperatures.49–51 A series of Pd–In bimetallic NPs with different structures and well-defined phase compositions was obtained, which were characterized by powder X-ray diffraction (XRD), scanning and transmission electron microscopy with energy-dispersive X-ray analysis (SEM-EDX, TEM, high-angle, annular dark-field scanning transmission electron microscopy (HAADF-STEM)), optical emission spectroscopy with inductively coupled plasma (ICP-OES), and X-ray photoelectron spectroscopy (XPS). The single-phase PdIn, Pd3In and Pd3In7 NP sols in the ionic liquids were employed as catalysts in the liquid-phase semi-hydrogenation of diphenyl acetylene (DPA) to investigate the influence of catalyst composition and structure on the catalytic performance. In particular, intermetallic PdIn NPs with a molar Pd/In ratio of 1:
1 revealed a high activity and excellent selectivity towards stilbene.
Conversion (X), selectivity (S) and yield (Y) were calculated according to the following equations:
YCST = SCSTXDPA![]() ![]() |
XRD analysis was employed to analyze the structure and phase composition of the Pd–In NPs. The XRD patterns of Pd–In bimetallic NPs with different compositions and crystal structures are shown in Fig. 1 (for corresponding calculated XRD patterns see Fig. S1‡). The XRD pattern of the Pd reference reveals reflections at 40.1°, 46.5°, 67.9°, 81.8° and 86.3° (2θ) which are characteristic for the face-centered cubic (fcc) Pd phase. The reflections of the Pd3In NPs (molar Pd/In ratio 3:
1) were at 39.2°, 45°, 66°, 79.2°, and 83.5° (2θ) and thus, shifted towards lower diffraction angels as compared to their monometallic Pd NP counterparts. This is in good agreement with Pd3In L10 phase with AuCu3 type structure (ICDD 98-024-7193) instead of tetragonal ZrAl3 and TiAl3 type structures mainly reported at high temperature.56 However, it was not possible to determine whether the cubic Pd3In NPs reveal an ordered structure as superlattice reflections were not observed due to the closely similar X-ray scattering factors of Pd and In.57 The molar Pd/In ratio as determined by ICP-OES and SEM-EDX analysis is 3
:
1 and corresponds to the molar precursor ratio used for NP synthesis (Table 1 and Fig. S6‡). For PdIn NPs, we observed a different set of reflections at 39.4° (110), 56.8° (200), 71.3° (211), and 84.5° (220) (2θ) which were consistent with the ordered intermetallic PdIn B2 phase with CsCl type structure (ICDD 98-005-9473). The molar Pd/In ratio as determined by ICP-OES and SEM-EDX is 1
:
1 (Table 1). Rietveld analysis further supported the single-phase composition of the intermetallic PdIn NPs (Fig. S2‡). For the Pd–In NPs (molar ratio 3
:
7), the XRD diagrams exhibited sharper reflections matching the reflections of the intermetallic Pd3In7 reference (ICDD 98-040-8314). Metallic In or In2O3 was not observed by XRD and Rietveld analysis (Table 1 and Fig. S4‡). For all NPs, the crystallite sizes were calculated according to the Scherrer equation. For Pd3In, PdIn and Pd NPs, the reflections were rather broad and of low intensity due to the small NP size. The crystallite sizes were 4 nm (Pd NPs), 5 nm (Pd3In NPs), and 9 nm (PdIn NPs) with a slight increase with increasing In content. The crystallite sizes were in a similar range as the mean particle sizes determined by TEM images (Fig. 2). However, the crystallite size was calculated to be 60 nm for intermetallic Pd3In7 NPs and thus, much larger than the ones for Pd, PdIn and Pd3In NPs.
![]() | ||
Fig. 2 TEM images with particle size distribution of (a) Pd NPs, (b) Pd3In NPs, (c) PdIn NPs, and (d) Pd3In7 NPs. |
Entry | Catalyst | Molar Pd![]() ![]() |
Mean NP sizeb (nm) | Crystallite sizec (nm) | Bragg angle (hkl) | |
---|---|---|---|---|---|---|
Precursor | NPsa | |||||
a Determined by ICP-OES and SEM-EDX. b Determined by statistical measurement from a number n of NPs (n > 100) based on TEM images. c The crystallite size was calculated according to the Scherrer equation. | ||||||
1 | Pd | 1![]() ![]() |
1![]() ![]() |
3.0 ± 0.5 | 4 | 40.1° (111) |
2 | Pd3In | 3![]() ![]() |
3![]() ![]() |
2.9 ± 0.9 | 5 | 39.2° (111) |
3 | PdIn | 1![]() ![]() |
1![]() ![]() |
4.8 ± 1.8 | 9 | 39.4° (110) |
4 | Pd3In7 | 3![]() ![]() |
1![]() ![]() |
8.5 ± 1.8 | 60 | 40.7° (330) |
The Pd–In NPs were isolated from the IL by precipitation and their morphology was examined by transmission electron microscopy (TEM). The mean particle size and size distribution of the Pd–In NPs were determined by statistical measurement from a number n of NPs (n > 100) based on the TEM images (Table 1 and Fig. 2). The monometallic Pd (Fig. 2a) and Pd3In NPs (Fig. 2b) had a nearly spherical shape and a small size with an average diameter about 3 nm, which was in good agreement with crystallite sizes calculated by the Scherrer equation (Table 1). Similarly, the intermetallic PdIn NPs (Fig. 2c) also revealed a spherical shape with a mean diameter of 4.8 ± 1.8 nm, which was slightly larger compared to the Pd and Pd3In NPs. It should be noted that a few particles with larger diameters of around 10 nm were also present. The Pd3In7 NPs exhibited the largest diameter (8.5 ± 1.8 nm) with a contribution of relatively large NPs as observed in Fig. 2d. This is in good agreement with the much larger crystallite size calculated based on the Scherrer equation by XRD analysis (Table 1). It should be noted that the Scherrer formula does not consider the particle size distribution. If a few large and small NPs are present in the powder sample, the diffraction profile represents a convolution of narrow and broad reflections, whereby the narrow reflections of larger NPs dominate the profile. Overall, the particle sizes increased with increasing amount of In used for the synthesis of bimetallic Pd–In NPs.
Fig. 3a displays a high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image of intermetallic PdIn NPs. Fig. 3b shows a high-resolution HRTEM of a single PdIn NP with a lattice spacing of 0.23 nm, which can be assigned to the (110) planes of the intermetallic B2 PdIn phase (space group: Pmm).45 To further investigate the distribution of Pd and In, EDS elemental maps and EDS line profiles of PdIn nanoparticles were obtained (Fig. 3c–f and Fig. S10‡), which indicated an overall homogeneous distribution of Pd and In within each particle without phase segregation. These results clearly support the successful synthesis of intermetallic PdIn NPs. In addition, Pd3In and Pd3In7 NPs were analyzed by HAADF-STEM imaging, EDS elemental mapping and EDS line scan profiles (Fig. S8, S9, S11 and S12‡). They also showed a uniform Pd and In distribution over the particle core, but some surface enrichment of the more dominant element (i.e. Pd for Pd3In and In for Pd3In7) was observed in this case. For the Pd3In7 NPs, a thin amorphous In oxide coating was further observed besides the crystalline Pd3In7 core by HAADF-STEM imaging (Fig. S9‡), which may be assigned to the partial oxidation of the In atoms on the surface upon exposure to air during sample preparation. For Pd3In NPs, some minor (statistical) variation in the Pd/In ratio for few NPs can also not be completely excluded (Fig. S8d‡).
Additionally, X-ray photoelectron spectroscopy (XPS) was performed on PdIn, Pd3In and Pd3In7 NPs to further investigate the electronic interaction between Pd and In. The XPS spectra of the Pd 3d and In 3d region are shown in Fig. 4. In the Pd 3d region (Fig. 4a), the Pd 3d5/2 peaks were located at 335.9, 336.1 and 336.4 eV, respectively for Pd3In, PdIn and Pd3In7 NPs, gradually shifted to higher binding energies (BE) compared to Pd(0) (335.1 eV). Upon addition of In, the electronic states of Pd are gradually filled for the In-rich phases, resulting in a partial negative charge on the Pd atoms. This was previously attributed to a simple charge transfer from In to Pd atoms according to Pauling's electronegativity values (1.78 (In), 2.20 (Pd)) and in agreement with the formation of intermetallic Pd–In NPs.58 Similar results have also been reported for intermetallic Pd–Ga59 and Pd–Sn60 NPs. However, it should be noted that such a simple charge transfer model, based on (minor) electronegativity differences, cannot account, for example, for the strong covalent bonding character in the 1:
1 Pd1In1 intermetallic compound influencing the electronic VB structure of the bulk phase. For all Pd–In NPs, the In 3d5/2 elemental line at 443.9 eV dominates and corresponds to indium in the metallic state as reported for intermetallic Pd–In compounds.61,62 Shoulders at higher binding energy of In 3d5/2 (444.9 eV) are characteristic for In3+ and decreased for In-poor intermetallic NPs (Table S1‡). In addition, the valence band (VB) spectra revealed the density of states at the Fermi energy (Ef) caused by Pd 4d states (Fig. S13‡). With increasing In content the maximum of the VB shifted to higher BE between 2 and 4 eV below Ef attributed to localized Pd 4d states.63 It is well known that the position of d-band center with respect to the Fermi level has a significant effect on the catalytic adsorption characteristics.64
The catalytic properties of the as-prepared PdIn, Pd3In and Pd3In7 NPs dispersed in the IL along with the monometallic Pd NPs were tested in the liquid-phase, semi-hydrogenation of DPA. Fig. 5 shows the possible reaction pathways, intermediates and products in the hydrogenation of DPA. DPA hydrogenation involves the semi-hydrogenation to (cis)-stilbene (CST) or (trans)-stilbene (TST) and their hydrogen-mediated isomerization. The catalytic hydrogenation of inner alkynes reveals an intrinsic stereoselectivity to (Z)-alkenes, because of their syn addition style and the catalytic hydrogenation of alkynes to (E)-alkenes, in principle, barely occurs.27 Recently, a tandem catalytic system was reported comprising Pd3Pb/SiO2 or Pd3Bi/SiO2 and RhSb/SiO2 for alkyne semi-hydrogenation and for alkene isomerization, respectively, which allowed also the one-pot TST synthesis from DPA.27,65 Generally, DPA is hydrogenated to stilbene and subsequently to undesired DPE. CST and TST derivates are interesting for use as dyes, liquid crystals, optical brighteners, OLEDs, or in the production of food additives. However, the selective hydrogenation of substituted carbon–carbon triple bonds can be challenging because the heats of adsorption of the reactants and intermediates are often similar. DPA hydrogenation is also associated with irreversible over-hydrogenation to 1,2-diphenylethane (DPE), which reduces the yield of cis- and trans-alkenes.
The catalytic performance of the Pd–In NPs with different compositions and crystal structures is compared to the Pd reference NPs in Fig. 6 and summarized in Table 2. For the monometallic Pd reference NPs, the Smax towards the desired CST and DPA conversion were 30% and 100% after 10 min of reaction, respectively, with DPE as the only product. Alloying Pd with In increased both CST selectivity and yield for all Pd–In NP catalysts. In general, the addition of In influences the surface geometric and electronic structures as well as the PdHx hydride formation of Pd catalysts.66 The Pd–In based catalysts revealed a significantly enhanced Smax (CST) in the range of 83 to 85% which was maintained even for higher DPA conversions (Fig. 6d). For example, intermetallic PdIn NPs showed 100% DPA conversion after 15 min which was comparable to the Pd NPs, while the CST selectivity significantly increased to 83%. Thus, the maximum CST yield increased from 30% for the Pd reference NPs to 83% for the PdIn NPs (Fig. 6c, Fig. S14‡ and Table 2). The Pd3In NP catalyst also revealed an enhanced maximum CST yield (70%) which was reached after 30 min of reaction. The volcano-type curve was comparable to earlier findings of Shaun K. Johnston et al., where TiO2-supported Pd3In NPs revealed a higher selectivity than their Pd counterpart but they were also less selective at full conversion (Fig. S15‡).67 Interestingly, the CST selectivity of PdIn NPs only slightly decreased over reaction time while the selectivity considerably decreased to 29.3% over Pd3In NP catalyst as a function of the reaction time (Fig. 6b): In case of intermetallic PdIn NP catalysts, the over hydrogenation to DPE seemed to be inhibited. In this case, a CST and TST yield of 75% and 5%, respectively, were reached after 15 min of reaction and remained nearly constant over 3 h of reaction (Table S2‡). For the intermetallic Pd3In7 (In-rich) catalyst, also a high maximum CST selectivity of 85% was observed but at the same time the DPA conversion decreased to 45%, affording an overall low Ymax(CST) of 40%.
Entry | Catalysts | Y(CST) [%] | S(CST) [%] | ||
---|---|---|---|---|---|
Y max | Y DPA50 | S max | S DPA50 | ||
1 | Pd | 30 (5 min) | 4 | 30 (5 min) | 17 |
2 | PdIn | 83 (15 min) | 43 | 86 (5 min) | 82 |
3 | Pd3In | 70 (30 min) | 38 | 83 (5 min) | 82 |
4 | Pd3In7 | 47 (20 min) | N/A | 85(10 min) | N/A |
The rather low activity over Pd3In7 NPs could be a result of the much higher In content of the NPs and/or the In surface segregation and the consequently lower number of active Pd centers on the NP surface (Fig. S16‡). In addition, the Pd3In7 NPs were also larger in size than the other Pd–In NPs and the Pd reference catalyst. The DPA conversion did not exceed 45% over time of reaction, probably due to catalyst deactivation. However, it should be noted that the catalyst structure did not change after reaction. The XRD patterns of the PdIn, Pd3In and Pd3In7 NPs after catalytic reaction are displayed in Fig. S17–19.‡Fig. 6d shows the CST selectivity as a function of DPA conversion. It is apparent that all Pd–In catalysts reveal a higher CST selectivity compared to the Pd reference over the total course of the hydrogenation reaction. Especially, the intermetallic PdIn NPs showed a much higher CST selectivity compared to Pd3In and Pd3In7 at high DPA conversion (>95%), yet remained around 80%. Overall, the intermetallic PdIn NPs synthesized in ILs combined high activity and selectivity and exhibited the best catalytic properties of the catalysts studied in this work, as maintaining excellent cis-stilbene selectivity at full conversion of DPA is a major challenge. The catalytic performance of other catalysts in the semi-hydrogenation of DPA is summarized in Table S3.‡ It should be noted that the catalytic activity and selectivity are not only influenced by the reaction conditions but also by the presence of different support materials.
The enhanced catalytic performance of Pd-based, bimetallic catalysts is usually ascribed to the isolation of active Pd sites (geometric effect) and/or the modified electronic structure (electronic effect). Based on DFT calculations, Pd single atoms were shown to be surrounded by In atoms in the most exposed (110) plane of PdIn, resulting in complete isolation of Pd atoms.45 Besides inhibition of Pd hydride formation and electronic effects, isolated Pd sites offered by the intermetallic structure of PdIn NPs could have also contributed to the enhanced catalytic performance observed in this study.68 For Pd3In NP catalysts, Pd surface enrichment and the high Pd content could be one reason for the lower CST selectivity at 100% DPA conversion, e.g., favoring the consecutive hydrogenation to DPE or affecting adsorption energies.
According to a stepwise hydrogenation mechanism, the route and selectivity of the reaction are determined by two factors, a kinetic (i.e. the reaction rate ratio for alkyne vs. alkene hydrogenation) and a thermodynamic factor (i.e. the alkyne and alkene adsorption energies). The overall decrease of the reaction rates and, in particular, the decrease in CST hydrogenation rates, was previously reported for alloying Pd with In in supported Pd/In catalysts.68 However, the Pd modification with In did not reveal a direct correlation between the change in the kinetic parameters of the reaction and the selectivity suggesting the importance of a thermodynamic factor. Negative adsorption energies for acetylene (−107 kJ mol−1) and ethylene (−49 kJ mol−1) adsorption at monometallic Pd(111) surfaces indicate relatively slow alkene desorption from Pd, causing overhydrogenation to the alkane.27 In general, the adsorption energy of alkenes is typically reduced in the presence of secondary metal atoms adjacent to Pd atoms, accelerating also alkene desorption.27 Recently, also high CST selectivities were reported for unsupported nanoporous Pd (Table S3‡) where the addition of a base further enhanced the CST selectivity.69 Heterolytic H–H bond cleavage was suggested to contribute to the enhanced catalytic performance in this case. We have previously demonstrated the synthesis of intermetallic PdSn NPs with orthorhombic structure (Pnma) in ILs using a similar approach. These PdSn NPs exhibited a comparably high CST selectivity in the semi-hydrogenation of DPA under these conditions while DPA conversion and CST yield were ∼30% and ∼24%, respectively, and thus much lower than for the PdIn NP catalysts.25
Footnotes |
† Dedicated to Prof. Dr Dieter Fenske on the occasion of his 80th birthday. |
‡ Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2nr03674f |
This journal is © The Royal Society of Chemistry 2022 |