- Email: [email protected]
SBA-15-supported Pd catalysts: The effect of pretreatment conditions on particle size and its application to benzyl alcohol oxidation
Journal of Catalysis 350 (2017) 21–29
Contents lists available at ScienceDirect
Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat
SBA-15-supported Pd catalysts: The effect of pretreatment conditions on particle size and its application to benzyl alcohol oxidation Chun-Hsia Liu a, Chih-Yu Lin b, Jeng-Lung Chen a, Kueih-Tzu Lu a, Jyh-Fu Lee a, Jin-Ming Chen a,⇑ a b
National Synchrotron Radiation Research Center, Hsinchu 30076, Taiwan Department of Chemistry, National Tsing Hua University, Hsinchu 30013, Taiwan
a r t i c l e
i n f o
Article history: Received 11 November 2016 Revised 30 December 2016 Accepted 24 January 2017
Keywords: Palladium catalyst Mesoporous material Alcohol oxidation X-ray absorption spectra
a b s t r a c t SBA-15-supported Pd catalysts were prepared by impregnation, thermal pretreatment, and H2 reduction. The pretreatment conditions have an important influence on the resulting size (from 1.3 to 10.3 nm) of the Pd particles; the catalyst prepared under vacuum conditions possessed the smallest particles (size 1.3 nm). X-ray absorption spectra were used to analyze the transformation of the Pd precursor in each step of the preparation. These catalysts were applied to the oxidation of benzyl alcohol by molecular oxygen; the catalyst with the smallest Pd particles exhibited the highest turnover frequency (9684 h 1). The catalyst was reusable three times without loss of activity (with conversion >96%). X-ray absorption spectra were recorded ex situ to monitor the oxidation and the coordination structure of Pd at varied temperatures of reaction; when the reaction temperature was above 100 °C, the Pd species were in the metallic state. This result indicates that metallic Pd might be the active phase in the oxidation of benzyl alcohol by molecular oxygen. Ó 2017 Elsevier Inc. All rights reserved.
1. Introduction The oxidation of alcohols to aldehydes or ketones is an important reaction in the production of fine and specialty chemicals [1,2]. In contrast to a traditional method using stoichiometric oxidation with mineral oxidizing agents, a recently developed method using a noble metal (e.g., Ru, Pt, Pd, or Au) as a catalyst and molecular oxygen as the oxidant constitutes a sustainable and green process [1,2]. Among those catalysts, Pd-containing catalysts appear to be promising because both high activity and high selectivity can be obtained concurrently [3–11]. As Pd is recognized to be a scarce resource, the synthesis of a Pd catalyst with high dispersion and high efficiency is an attractive topic in catalytic research [12–15]. Among various methods of synthesizing a supported Pd catalyst are impregnation [7,12,16,17], a ligand-assisted method [14–16,18,19], and an ion-exchange method [4,12,17,20]. With the latter two methods and on improving the interaction of the Pd precursor and the support through grafting a hydrophilic or hydrophobic functional group (e.g., ANH, ASOH, ACH@CH2, or ACH3) or adjusting the surface charge of the support, catalysts consisting of Pd nanoparticles (size <2 nm) were synthesized [4,12,14–20], but the impregnation ⇑ Corresponding author. E-mail address: [email protected] (J.-M. Chen). http://dx.doi.org/10.1016/j.jcat.2017.01.019 0021-9517/Ó 2017 Elsevier Inc. All rights reserved.
method yielded Pd catalysts that generally showed a broad distribution of size of Pd particles larger than 2 nm [7,12,16,17,19]. A synthesis of a Pd catalyst that possesses a narrow distribution of particle size <2 nm with a simple impregnation method is attractive [16]. When we synthesized SBA-15-supported copper catalysts by a vacuum-thermal method, we found that the vacuum condition was vital for preparing highly dispersed copper catalysts; the transformation of the copper species during the heat treatment was altered by the vacuum condition [21]. The atmosphere during the thermal decomposition is reported to be an important factor in tuning the resulting metal or metal-oxide size in supported catalysts [12,22–24]. Sietsma et al. found that gaseous dinitrogen monoxide prevented the rapid decomposition of Ni3(NO3)2(OH)4 that is observed in air, resulting in smaller NiO nanoparticles than those prepared in air [22]. Zou et al. discovered that, when the chemical structure of the adsorbed palladium complex is varied during the pretreatment (i.e., with H2, O2, or N2), the resulting silica-supported Pd catalysts possessed varied particle sizes [23]. Yuranov et al. showed that reducing the adsorbed palladium species before oxidation gave the resulting SBA-15-supported PdO catalyst smaller PdO nanoparticles than without that reduction [24]. To the best of our knowledge, there is no report of supported Pd catalysts from a vacuum-thermal preparation.
22
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
Here we report the preparation of SBA-15-supported Pd catalysts with an impregnation method and subsequent varied pretreatment conditions and H2 reduction, and tests of oxidation of benzyl alcohol with molecular oxygen. SBA-15 was selected as a support because it has great chemical stability, a large surface area for metal dispersion, and porosity of nanometer size to facilitate molecular diffusion [18,21,25]. The pretreatment condition was found to have an important influence on the resulting Pd particle size (from 1.3 to 10.3 nm); among those catalysts the catalyst pretreated under vacuum conditions possessed the smallest particles (1.3 nm). X-ray absorption spectroscopy (XAS) served for analysis of the transformation of the Pd precursor in each preparation step. These catalysts were applied to the oxidation of benzyl alcohol with molecular oxygen, which showed that the catalyst with the smallest Pd particles exhibited the greatest activity. The catalyst was reusable three times without loss of activity (with conversion >96%). XAS measurements were recorded ex situ to monitor the oxidation and the coordination structure of Pd at varied reaction temperatures.
2. Experiments
drate (ZrOCl28H2O) in hydrochloric acid (HCl). The mixture was stirred at 35 °C for 24 h. After sulfuric acid (H2SO4) was added to the mixture in a calculated proportion to remove the triblock copolymer [26], the mixture was aged at 90 °C for 24 h. The molar proportions of the final mixture were 1 TEOS:0.017 P123:5.9HCl:193 H2O:2.9 H2SO4:0.05ZrOCl2. The solid product was washed with propanone, filtered, and dried in air. The product was heated further at 350 °C to produce the SBA-15 support. To impregnate Pd nanoparticles, we loaded the SBA-15 support with a solution of dichlorobis(acetonitrile) palladium (PdCl2(MeCN)2) in propanone (nominally 1 mass% Pd by an incipient wetness method) and dried it. The sample was subsequently heated under varied gases (hydrogen, air, nitrogen), or vacuum, at 350 °C for 2 h with a ramp rate of 1 °C min 1. The resulting samples are referred to as Sample X, in which X denotes the conditions of thermal treatment (with X = H2, air, N2, or Vac for treatment under hydrogen, air, nitrogen, or vacuum (ca. 0.1 Torr), respectively). Each sample X was further reduced under pure hydrogen at 350 °C for 2 h. After cooling to 23 °C, the resulting samples were stored under atmospheric conditions. The resulting samples are referred to as Sample X–H2, in which X denotes the conditions of thermal treatment (with X = Air, N2, and Vac for pretreatment in air, nitrogen or vacuum, respectively).
2.1. Preparation of the catalyst The synthesis of mesoporous SBA-15 silica involved the addition of tetraethoxysilane (TEOS) to a solution of triblock copolymer (Pluronic P123, EO20PO70EO20) and zirconium oxychloride octahy-
2.2. Characterization of the catalyst X-ray diffraction (XRD) patterns were recorded on a diffractometer (Mac Science 18MPX) using Cu Ka radiation. N2 physisorption isotherms were measured (QuantachromeAutosorb-1-MP) at 77 K. The isotherms were analyzed with nonlocal densityfunctional theory (NLDFT) to evaluate the pore sizes of the samples using the kernel of NLDFT equilibrium capillary condensation isotherms of nitrogen at 77 K on silica (adsorption branch, assuming cylindrical pore geometry). The BET surface areas were calculated from the adsorption branches in the relative pressure range 0.05–0.30; the total pore volumes were evaluated at relative pressure 0.95. Inductively coupled plasma mass spectral (ICP-MS) data were recorded (Perkin–Elmer SCIEX-ELAN 5000). Transmission electron microscopy (TEM) images were taken from ultramicrotomed samples (thickness 80–100 nm) with an electron microscope (JEOL JEM-ARM 200 FTH) equipped with an energydispersive X-ray (EDX) spectrometer and a high-angle annular dark-field (HAADF) detector. X-ray photoelectron spectra (XPS) were recorded on an Ulvac-PHI PHI Quantera device. 2.3. XAS
Fig. 1. Small-angle and wide-angle XRD patterns of SBA-15 and SBA-15-supported Pd materials.
Pd K-edge XAS were recorded at beamline 01C1 at Taiwan Light Source (storage-ring energy 1.5 GeV, ring current 360 mA) in the National Synchrotron Radiation Research Center. In a typical experiment, the Pd catalyst (about 120 mg) was pressed to form self-supporting wafers and mounted on a sample holder for measurements. Spectra were recorded near 23 °C in a transmission
Table 1 Physical properties of materials.a.
a b c
Sample
a (nm)
D (nm)
W (nm)
S (m2 g
SBA-15 H2 Air-H2 N2-H2 Vac-H2
12.1 12.1 12.1 12.1 12.1
8.1 8.1 8.1 8.1 8.1
4.0 4.0 4.0 4.0 4.0
1023 1001 1022 1008 1024
1
)
Vt (cm3 g 1.38 1.31 1.31 1.31 1.31
a: unit cell parameter; D: mesopore diameter; W: wall thickness; S: BET surface area; Vt: total pore volume. DPd: Pd particle size calculated from XRD pattern. DPd: average particle size observed by TEM.
1
)
DPdb (nm)
DPdc (nm)
— 5.6 4.5 10.2 —
— 6.6 5.2 10.3 1.3
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
23
mode with two gas-filled ionization chambers in series to measure the intensities of the incident beam and the transmitted beam. A Pd foil that served as a reference for energy calibration was measured simultaneously with a third ionization chamber. X-ray absorption near-edge structure (XANES) spectra and extended X-ray absorption fine-structure (EXAFS) spectra were all recorded near 23 °C. For the measurement of the XAS of the Pd catalyst under working conditions, spectra were recorded ex situ under ambient conditions and in a fluorescent mode. In a typical experiment, a mixture of benzyl alcohol (0.87 mL, 8.4 mmol), p-xylene (2.0 mL), and Sample Vac–H2 (9 mg, 0.01 mol.% of Pd relative to benzyl alcohol) was stirred (800 rpm) in flowing O2 (20 sccm) at a reaction temperature in a range from 25 to 140 °C for 0.5 h. After the reaction, the flow of the oxygen was stopped and the solution was cooled near 23 °C under ambient conditions for the measurement of XAS. For the in situ Pd K-edge XAS measurement of the Pd catalyst in pure H2, a custom-made stainless steel cell was used. Two holes were made in the cell, one on the front and the other on the back. After the solid samples were placed inside the cell, the holes were closed with Kapton film to avoid exposure of the sample to the ambient atmosphere. Before the XAS measurement, the samples were reduced at 350 °C in pure H2 and then cooled to room temperature for the measurement. 2.4. Testing the catalyst In a typical reaction, a mixture of benzyl alcohol (0.87 mL, 8.4 mmol), p-xylene (2.0 mL), and the SBA-15-supported Pd catalyst (9 mg, 0.01 mol.% of Pd relative to benzyl alcohol) was stirred (at 800 rpm) under flowing O2 (20 sccm) at 120 °C for 3 h. After the reaction, the solid catalyst was filtered off while the liquid phase (with decane added as internal standard) was analyzed on a gas chromatograph (Shimadzu GC-2014 equipped with a FID detector and a SGE BP-5 column). To test reuse, the catalysts were collected with centrifugation, washed with p-xylene, dried, and then reused in further runs. 3. Results and discussion 3.1. Characterization of SBA-15-supported Pd materials Mesoporous SBA-15 with short mesochannels was prepared as a support for the deposition of palladium. The synthesis involved the addition of Zr(IV) salt to the synthesis solution, which was removed on the addition of sulfuric acid to the synthesis solution before the aging step [26]. The Zr/Si ratio of the resulting SBA-15 was 0.002 (i.e., 0.4 mass% ZrO2) determined with ICP-MS; the remaining Zr species were inaccessible to the acidic solution and would be inside the silica framework rather than on the mesopore surface. The resulting SBA-15 support exhibited a highly ordered p6mm mesostructure with unit cell parameter a = 12.1 nm and uniform mesopores of width 8.1 nm (see Fig. 1 for the XRD pattern, Fig. S1 for the N2 physisorption isotherm, and Table 1). The SBA-15 support was impregnated with PdCl2(MeCN)2 in propanone solution and heated at 350 °C under varied atmospheres. ICP-MS analysis indicated that all Pd-containing materials had Pd about 0.95 mass%, near the nominal value 1 mass% Pd. The small-angle XRD data in Fig. 1 confirmed the retention of the ordered p6mm mesostructure after the Pd deposition and thermal treatment. The wide-angle XRD data revealed that the Pd species in Sample H2 consisted of Pd nanocrystals (5.6 nm, JCPDS file 870638). An XRD signal contributed from PdO (JCPDS file 75-584) was observed in the XRD pattern of Sample Air, which indicated the formation of PdO. Signals contributed from Pd and PdO were
Fig. 2. Bright-field TEM images of (a) sample H2, (b) sample Air–H2, (c) sample N2– H2, and (d) sample Vac–H2.
observed in the XRD pattern of Sample N2, indicating the formation of metallic Pd and PdO simultaneously after heating under an inert gas. In the XRD pattern of Sample Vac, no signal was contributed from a Pd species after the thermal treatment, indicating the formation of amorphous Pd species or highly dispersed nanoparticles.
24
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
Fig. 3. Dark-field TEM images of sample Vac–H2 viewed along (a, b) and perpendicular to (c, d) the axis of the hexagonally arranged mesopores. Scale bars indicate 20 nm.
After the subsequent reduction at 350 °C, the signal contributed by PdO in the XRD patterns of Samples Air and N2 vanished and signals from metallic Pd were observed. The Pd particle sizes calculated with the Scherrer equation were 4.5 nm for Sample Air–H2 and 10.2 nm for Sample N2–H2. For the XRD pattern of Sample Vac–H2, no additional peaks were formed after the reduction in H2. In N2 physisorption isotherms of the Sample X–H2 series (see Fig. S1), they exhibited sharp steps and hysteresis loops of type H1 and isotherms similar to that of the SBA-15 support. The minor change of the isotherms occurred because of the small content of Pd. The reduced samples were characterized with a TEM to observe the Pd nanoparticles; the results appear in Figs. 2 and S2. In four samples, hexagonally arranged mesopores of channel type were clearly observed; the average pore diameter was near 8 nm, in agreement with the value derived from the N2 physisorption data (see Table 1). For the sample with direct reduction (i.e., Sample H2) the Pd particles were large (7 nm) and aggregation of the particles was observed. For Sample Air–H2, the Pd particle size decreased to 5.0 nm. For Sample N2–H2, the Pd particles were larger than for Sample H2 and were inhomogeneously dispersed in the silica support. The distribution of particle sizes was broad, from 4 to 20 nm. For Sample Vac–H2, no discrete Pd nanoparticles were observed in the silica support, although the EDS data confirmed the presence of Pd. We further characterized the Sample Vac–H2 with TEM-HAADF viewed along and perpendicular to the axis of the hexagonal mesopores; the results appear in Figs. 3 and S3. Bright spots, in the range 0.6–2.1 nm with average size 1.3 nm, were homogeneously distributed in the SBA-15 support. Most spots were incorporated into the wall of SBA-15. The particles in Sample Vac–H2 had a narrow distribution of particle size, average size 1.3 nm (cf. Figs. 3 and S2). Overall, the Pd nanoparticle size decreased in the sequence Sample N2–H2 (10.3 nm) > Sample H2 (6.6 nm) > Sample Air–H2 (5.2 nm) > Sample Vac–H2 (1.3 nm). It should be noticed that for the synthesis of the catalysts with small Pd size (<2 nm) and a narrow distribution of particle size, the exis-
tence of a functional group or adjustment of the surface charge of the support is generally required to improve the interaction between the Pd precursor and the support [4,12,14,16–18]. It is interesting that the impregnation method and pretreatment under vacuum achieved a narrow distribution of particle size with Pd size <2 nm. We recorded XPS to characterize the oxidation state of Pd; Fig. S4 shows XPS of the reduced samples in the Pd3d region. For Sample H2 and Sample N2–H2, the signals of binding energy were observed at 340.8 and 335.5 eV, corresponding to Pd3d3/2 and Pd3d5/2, respectively. Pd in both samples was in metallic form after reduction with hydrogen at 350 °C. For Samples Air–H2 and Vac– H2, the signals of binding energy were observed at 342.1, 340.8, 336.8, and 335.5 eV, which are assigned to PdO3d3/2, Pd3d3/2, PdO3d5/2, and Pd3d5/2, respectively: both metallic Pd and PdO were hence present in the samples. Under such reduction, the Pd species are expected to be totally reduced, so that the formation of PdO is expected to be associated with the reoxidation of the surface of the metallic Pd nanoparticles when the Pd materials were stored under air [4,19]. The result indicated that the Pd nanoparticles formed in Samples Air–H2 and Sample Vac–H2 were smaller than those in Samples H2 and N2–H2, consistent with the TEM analysis. The inverse correlation between Pd particle size and oxidation state over SBA-15 in as-prepared catalysts was reported [4]. The molar ratios of surface PdO in total palladium were 18% for Sample Air– H2 and 28 % for Sample Vac–H2. 3.2. XAS of SBA-15-supported Pd materials According to our experimental results, the pretreatment conditions of PdCl2(MeCN)2 were important for the formation of small and highly dispersed Pd nanoparticles in SBA-15. We further characterized with XAS the SBA-15-supported Pd catalysts before and after pretreatment and subsequent hydrogen reduction to evaluate the transformation of the oxidation state and the coordination
25
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
Fig. 4. (a) Pd K-edge XANES spectra and (b) Fourier transform profiles of Pd K-edge k3-weighted EXAFS data (-) and fitted results () of SBA-15-supported Pd samples before and after thermal pretreatments. Data ranges for Fourier transformation (without phase correction) and fitting are Dk = 3.2–13.6 Å 1 and DR = 1.0–3.8 Å, respectively.
structure of palladium at each preparation step. Fig. 4 shows Pd Kedge XANES spectra and the corresponding Fourier transform (FT) profiles of k3-weighted EXAFS data of the impregnated sample, Sample Air, Sample N2, and Sample Vac with the reference spectra of PdCl2 (MeCN)2, PdCl2, Pd foil, and PdO for comparison. The structural parameters from the EXAFS fit of samples are listed in Table 2.
After the SBA-15 support was impregnated with the solution of PdCl2(MeCN)2 in propanone and dried under atmospheric conditions, the XANES spectrum of the impregnated sample (labeled imp.) showed an absorption-edge profile different from that of PdCl2(MeCN)2, indicating a change of the coordination structure of the impregnated Pd(II) complex. The intensities of the signals at 24,369 and 24,386 eV in the spectrum decreased; the spectrum exhibited an absorption-edge profile similar to that of PdCl2 rather than of PdCl2(MeCN)2 (see Fig. 4a). In its FT profile and fitted value (see Fig. 4b and Table 2), the intensity of the signal attributed to PdAN coordination (R = 2.00 Å, NPdAN = 1.1) decreased while the intensity of the signal attributed to PdACl coordination (R = 2.30 Å, NPdACl = 2.9) increased, relative to the signals attributed to PdAN coordination (R = 1.95 Å, NPdAN = 2.0) and PdACl coordination (R = 2.31 Å, NPdACl = 2.0) in the spectrum of PdCl2(MeCN)2. These results indicated that, when the PdCl2(MeCN)2 complex was impregnated in the pores of the SBA-15 support, one MeCN ligand in PdCl2(MeCN)2 left and the Pd center became further coordinated with a chlorine atom from a nearby Pd complex. The elimination of MeCN from PdCl2(MeCN)2 might be due to the weakness of the PdAN bond in that a p-electron associated with the C„N triple bond was withdrawn by the Lewis acid site on the surface of SBA-15 [27]. The impregnated sample was further subjected to thermal treatments under air, N2, or vacuum. After the thermal treatment in air, the XANES spectrum of Sample Air changed. The first signal at 24,371 eV became sharper and shifted to decreased energy. In its FT profile and fitted value (see Fig. 4b and Table 2), the intensity of the signal attributed to PdACl coordination (R = 2.30 Å, NPdACl = 0.5) decreased and two signals appeared, attributed to the PdAO shell (R = 2.03 Å, NPdAO = 3.0) and the PdAPd shell (R = 3.03 Å, NPdAPd = 1.1). The fitted result indicated that, after the thermal treatment in air at 350 °C, the Pd species in Sample Air had mostly a local structure of PdO type, but some PdACl bonds remained (see Fig. 4b and Table 2). After the thermal treatment under N2, the XANES spectrum of Sample N2 showed an absorption edge profile similar to that of the Pd foil, indicating the reduction of Pd(II). The FT profile exhibited a sharp signal attributed to PdAPd coordination (R = 2.74 Å, NPdAPd = 7.3) and a broad signal attributed to PdAO coordination (R = 2.00 Å, NPdAO = 0.7) (see Fig. 4b and Table 2). The reduction of the Pd(II) complex was related to the dechlorination–reduction during the thermal treatment under an inert gas [28]. The Pd species in Sample N2 was mostly in the metallic form, but some PdAO bonds were also present. After the thermal treatment under vacuum, the XANES spectrum of Sample Vac showed an absorption-edge profile
Table 2 Structural parameters of the Pd catalysts after heat treatment.a. Samples
Shell
C.N.
R (Å)
r2 (Å2)
Eshift (eV)
R-factor (%)
PdCl2(MeCN)2
PdAN PdACl PdAN PdACl PdAO PdACl PdAPd PdAPd PdAO PdAPd PdAO PdAPd PdACl PdAPd PdAO PdAPd PdAPd
2.0 2.0 1.1 2.9 3.0 0.5 1.1 0.5 0.7 7.3 1.0 2.8 4.0 12.0 4.0 4.0 8.0
1.95 2.31 2.00 2.30 2.03 2.30 3.03 3.42 2.00 2.74 2.01 2.70 2.31 2.75 2.03 3.03 3.42
0.003 0.003 0.003 0.005 0.006 0.001 0.006 0.001 0.002 0.006 0.003 0.009 0.004 0.005 0.003 0.005 0.005
5.3 3.3 0.1 6.9 5.7 6.0 1.4 1.1 8.0 3.2 17.7 6.0 22.1 6.2 5.6 12.3 11.4
1.4
Imp. Air
N2 Vac PdCl2 Pd PdO
a
C.N.: coordination number; R: bond length; r2: Debye–Waller factor; Eshift: inner potential correction.
0.8 3.4
1.7 5.0 3.0 1.9 1.4
26
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
different from that of the impregnated sample. In its FT profile, there were two broad signals attributed to PdAO coordination (R = 2.01 Å, NPdAO = 1.0) and PdAPd coordination (R = 2.70 Å, NPdAPd = 2.8). The lack of signal associated with PdACl coordination indicated the total decomposition of the Pd precursor. For the subsequently reduced samples, their Pd K-edge XANES spectra and the corresponding FT profiles of k3-weighted EXAFS data are shown in Fig. 5 with reference spectra of Pd foil and PdO for comparison. Table 3 summarizes the structural parameters from the EXAFS fit of those samples. After the subsequent reduction and storage under atmospheric conditions, the Sample X–H2 series were measured in air. The XANES spectra of Sample H2 and Sample N2–H2 showed absorption-edge profiles similar to that of metallic Pd foil. Their FT profiles showed only a sharp signal attributed to PdAPd coordination (R = 2.74 Å), indicating that the Pd species in these two samples were reduced to metallic form.
For Samples Air–H2 and Vac–H2, the signals at 24,367 and 24,390 eV in their XANES spectra were less sharp than those in the spectrum of Pd foil. In their FT profiles, a broad signal attributed to PdAO coordination and a signal attributed to PdAPd coordination were observed. These results indicated that the Pd species in these samples were Pd and PdO. To elucidate the oxidation state of palladium after the hydrogen reduction in 350 °C, in situ Pd Kedge XAS measurements were performed (see Fig. S5 and Table S1). The result showed that before exposure to the air the Pd species in Samples Air–H2 and Vac–H2 were in the metallic state and the Pd species in these samples was only Pd. As with the previous observation in the XPS results, the formation of PdO is expected to be associated with the reoxidation of the surface of the small metallic Pd nanoparticles when the Pd materials were stored under air [4,19]. 3.3. Catalyst performance
Fig. 5. (a) Pd K-edge XANES spectra and (b) Fourier transform profiles of Pd K-edge k3-weighted EXAFS data (–) and fitted results () of SBA-15-supported Pd samples after thermal reduction. Data ranges for Fourier transformation (without phase correction) and fitting are Dk = 3.2–13.6 Å 1 and DR = 1.0–3.8 Å, respectively.
We applied the reduced samples (i.e., Samples H2, Air–H2, N2–H2 and, Vac–H2) to the oxidation of benzyl alcohol at 120 °C; the corresponding profiles of conversion of benzyl alcohol versus time are exhibited in Fig. 6. In that figure, the conversion of benzyl alcohol increased sharply in the first 2 h and then gradually during the following 1 h. The catalytic results are summarized in Table 4, which shows the results for conversion of benzyl alcohol, product selectivity, and turnover frequency (TOF). The catalysts produced benzaldehyde as the main product and toluene as the main byproduct. The selectivities of benzaldehyde for these catalysts were >85%. Among four catalysts, Sample Vac–H2 showed the highest activity; the conversion achieved 99% in 3 h. Blank reactions were also carried out and the corresponding conversion was <2% under our reaction conditions. When this sample was also reused three times under the same reaction conditions, the conversion remained above 96%. Overall, the trend of TOF was Sample Vac–H2 > Sample Air–H2 > Sample H2 > Sample N2–H2, which shows a trend the same as for the size of Pd nanoparticle: Sample Vac–H2, which possessed the smallest Pd size, exhibited the highest TOF. Sample Vac–H2 was further tested at varied reaction temperatures; the catalytic results are shown in Table 5 and Fig. S6. As the reaction temperature increased, the selectivity of benzaldehyde decreased a little from 94 to 86%. For temperatures <100 °C, the conversion and TOF were moderate (i.e., 47% and 383 h 1). When the reaction temperature was increased to 100 °C, the conversion and TOF increased sharply to 82% and 3953 h 1. The increased TOF might be due to the reduction of surface PdO to metallic Pd by benzyl alcohol at 100 °C. When the reaction temperature was further increased to 130 °C, the conversion and TOF increased steadily to 99% and 10,672 h 1, but when the reaction temperature became 140 °C, the conversion and TOF abruptly decreased to 79% and 6127 h 1. Such decreased catalytic activity
Table 3 Structural parameters of the Pd catalysts after H2 reduction.a. Samples
Shell
C.N.
R (Å)
r2 (Å2)
H2 Air–H2
PdAPd PdAO PdAPd PdAPd PdAO PdAPd PdAPd PdAO PdAPd PdAPd
7.8 1.4 4.1 8.7 1.0 2.5 12.0 4.0 4.0 8.0
2.74 2.00 2.73 2.74 2.02 2.70 2.75 2.03 3.03 3.42
0.006 0.004 0.005 0.006 0.001 0.009 0.005 0.003 0.005 0.005
N2–H2 Vac–H2 Pd PdO
a
C.N.: coordination number; R: bond length; r2: Debye–Waller factor; Eshift: the inner potential correction.
Eshift (eV) 4.8 7.3 4.7 4.6 20.4 6.4 6.2 5.6 12.3 11.4
R-factor (%) 0.2 0.6 0.3 5.0 1.9 1.4
27
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
the ion-exchange method (90 °C, 1120 h 1) [4], as both catalysts possessed a similar Pd size (1.9 nm), but, compared with the carbon-supported Pd catalysts, such as Pd@N-doped carbon (120 °C, 153,551 h 1) [5] and Pd/Graphene (110 °C, 30,137 h 1) [11], the oxide-supported Pd catalysts showed poor activity, which was due to the strong adsorption of benzyl alcohol on the supports [5,11].
3.4. XAS of sample Vac–H2 for varied reaction temperature
Fig. 6. Profiles of benzyl alcohol conversion vs. time. Reaction conditions: benzyl alcohol (0.87 mL, 8.4 mmol), SBA-15-supported Pd catalyst (9 mg, 0.01 mol.% of Pd relative to benzyl alcohol), and p-xylene (2.0 mL) under flowing O2 (20 sccm) at 120 °C.
at a high reaction temperature is unexpected; this phenomenon might have resulted from the oxidation of metallic Pd nanoparticles [1] or the decreased adsorption property of Pd nanoparticles for benzyl alcohol [29]. The activity of Sample Vac–H2 is greater than for some reported supported Pd catalysts, such as Pd/Al2O3 (120 °C, 3615 h 1) [7], Pd/Al2O3 (100 °C, 3300 h 1) [9], and Pd/ NaX zeolite (100 °C, 626 h 1) [10]. It showed a TOF similar to that for the reported SBA-15-supported Pd catalyst synthesized with
As it is generally recognized that the oxidation state of Pd greatly affects the catalytic activity [7,30,31], we recorded the XAS of Sample Vac–H2 ex situ after the reaction at varied temperatures in the range 25–140 °C to observe the oxidation and the coordination structure of Pd under working conditions. After each reaction for 30 min, part of the solution was extracted and cooled in air to near 23 °C for the recording of XAS in a fluorescent mode. The results are shown in Fig. 7; the structural parameters from the EXAFS fits are listed in Table S2. When the reaction temperature was < 100 °C, the oxidation state and the coordination structure of the Pd species were constant; the XANES spectra and the FT profiles were nearly identical to those of Sample Vac–H2 before the reaction. When the reaction temperature attained 100 °C, the XANES spectrum showed a metallic feature at 24,390 eV; in its FT profile, the intensity of the signal attributed to PdAPd coordination (R = 2.76 Å, NPdAPd = 3.1) increased whereas the intensity of the signal attributed to PdAO coordination (R = 2.01 Å, NPdAO = 1.3) decreased. This result indicated the reduction of Pd(II) to Pd(0) at 100 °C. When the reaction temperature was increased to 140 °C, the metallic nature and the size of the Pd particles did not change; the XANES spectra and the FT profiles after 110–140 °C are similar. This observation indicated that, for reaction temperatures >100 °C, metallic Pd was the major phase, and it was suspected to be the active phase under our reaction conditions; for reaction temperatures <100 °C, PdO was still present, which resulted in poor
Table 4 The catalytic performance of the SBA-15-supported Pd catalysts.a.
a b c
Conv.
Selectivity (%)
Catalyst
(%)
Benzaldehyde
Othersb
TOF (h
H2 Air–H2 N2–H2 Vac–H2
40 77 28 99
93 85 87 90
7 15 13 10
1779 3953 791 9684
1 c
)
Conv.: conversion; TOF: turnover frequency. Including toluene, benzoic acid, dibenzyl ether, and benzylbenzoate. Calculated as moles of benzyl alcohol converted per mole of total Pd per hour, measured at time on stream of 30 min.
Table 5 The catalytic performance of sample Vac–H2 at different reaction temperatures.a.
a b c
T
Conv.
Selectivity (%)
Catalyst
(°C)
(%)
Benzaldehyde
Othersb
(h
Vac–H2
80 90 100 110 120 130 140
8 47 82 94 99 99 79
94 95 86 90 90 87 86
6 5 14 10 10 13 14
296 1383 3953 8498 9684 10,672 6127
Conv.: conversion; TOF: turnover frequency. Including toluene, benzoic acid, dibenzyl ether, and benzylbenzoate. Calculated as moles of benzyl alcohol converted per mole of total Pd per hour, measured at time on stream of 30 min.
TOF 1 c
)
28
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
4. Conclusions SBA-15-supported Pd catalysts were synthesized and applied to the oxidation of benzyl alcohol with molecular oxygen as oxidant. Depending on the thermal pretreatment, SBA-15-supported catalysts with Pd nanoparticles of size 1.3 to 10.3 nm were synthesized. The catalysts were applied to the oxidation of benzyl alcohol, which showed that Sample Vac–H2 with the smallest Pd possessed the largest TOF. The catalyst was reused three times without loss of activity (i.e., conversion >96%). From the catalytic results and the XAS ex situ, when the reaction temperature was greater than 100 °C, the Pd species in Sample Vac–H2 was in metallic form and the catalytic activity increased sharply, indicating that metallic Pd might be the active phase in the reactions. This result indicates that tuning the thermal pretreatment condition enables the synthesis of highly dispersed and active Pd catalysts; this method might be applied to synthesize other supported highly dispersed Pd catalysts for which the interaction of the Pd precursor and the support is weak. Electronic supplementary information (ESI) available includes N2 physisorption isotherms, size distributions from the TEM images, XPS data, the results of catalytic tests, and structural parameters from EXAFS fits. Acknowledgments We thank the NSRRC staff for their technical support. NSRRC and the Ministry of Science and Technology of Republic of China under grant Nos. MOST102-2113-M-213 -004 -MY3 and MOST105-2113-M-213 -005 -MY3 provided support of this research. Appendix A. Supplementary material 3
Fig. 7. (a) Pd K-edge XANES spectra and (b) Fourier transforms of Pd K-edge k weighted EXAFS data (–) and fitted results () of sample Vac–H2 at varied reaction temperature. Data ranges for Fourier transformation (without phase correction) and fitting are Dk = 3.2–10.7 Å 1 and DR = 1.0–3.8 Å, respectively.
catalytic activity. In addition, at 140 °C the metallic Pd nanoparticles were not oxidized or aggregated, while the catalytic activity decreased from 130 to 140 °C (see Tables 5 and S2). The decreased catalytic activity was obviously not related to the modification of the active phase at a high reaction temperature. The XAS results ex situ indicate that metallic Pd might be the active phase at a high reaction temperature (i.e., 100–140 °C). In the literature, it had been suspected that metallic Pd was the active phase. A tentative reaction mechanism on the surface of metallic Pd was proposed: An oxidative addition of an OAH bond from alcohol to the metallic Pd affords a Pd–alcoholate species and the species undergoes b-hydride elimination to produce the corresponding aldehyde [2,3,5,6,10]. Nevertheless, in recent surface analysis, high-pressure XPS measurements over metallic and oxidized Pd(1 1 1) single crystals suggested that aldehyde formation was disfavored over metallic Pd(1 1 1) but favored on oxidized Pd(1 1 1) surfaces [32]. The results of time-resolved liquid and vapor phase XAS measurements also revealed that surface palladium oxide was the catalytically active phase [31,33–35]. The alcohol reacted with the surface palladium oxide on the metallic Pd to form the corresponding aldehyde. To determine the active site, XAS results ex situ might not be sufficient, because the surface palladium oxide was unstable and might be further reduced by benzyl alcohol under our measurement conditions [32]. Further experiments need to be performed to distinguish the active site and elucidate the mechanism of benzyl alcohol oxidation in our catalytic condition.
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcat.2017.01.019. References [1] T. Mallat, A. Baiker, Oxidation of alcohols with molecular oxygen on solid catalysts, Chem. Rev. 104 (2004) 3037–3058. [2] Z. Guo, B. Liu, Q. Zhang, W. Deng, Y. Wang, Y. Yang, Recent advances in heterogeneous selective oxidation catalysis for sustainable chemistry, Chem. Soc. Rev. 43 (2014) 3480–3524. [3] J. Chen, Q. Zhang, Y. Wang, H. Wan, Size-dependent catalytic activity of supported palladium nanoparticles for aerobic oxidation of alcohols, Adv. Synth. Catal. 350 (2008) 453–464. [4] C.M.A. Parlett, D.W. Bruce, N.S. Hondow, M.A. Newton, A.F. Lee, K. Wilson, Mesoporous silicas as versatile supports to tune the palladium-catalyzed selective aerobic oxidation of allylic alcohols, ChemCatChem 5 (2013) 939– 950. [5] P. Zhang, Y. Gong, H. Li, Z. Chen, Y. Wang, Solvent-free aerobic oxidation of hydrocarbons and alcohols with Pd@N-doped carbon from glucose, Nat. Commun. 4 (2013) 1593–1604. [6] K. Mori, T. Hara, T. Mizugaki, K. Ebitani, K. Kaneda, Hydroxyapatite-supported palladium nanoclusters: a highly active heterogeneous catalyst for selective oxidation of alcohols by use of molecular oxygen, J. Am. Chem. Soc. 126 (2004) 10657–10666. [7] X. Wang, G. Wu, N. Guan, L. Li, Supported Pd catalysts for solvent-free benzyl alcohol selective oxidation: effects of calcination pretreatments and reconstruction of Pd sites, Appl. Catal. B 115–116 (2012) 7–15. [8] Y. Chen, Z. Guo, T. Chen, Y. Yang, Surface-functionalized TUD-1 mesoporous molecular sieve supported palladium for solvent-free aerobic oxidation of benzyl alcohol, J. Catal. 275 (2010) 11–24. [9] H. Wu, Q. Zhang, Y. Wang, Solvent-free aerobic oxidation of alcohols catalyzed by an efficient and recyclable palladium heterogeneous catalyst, Adv. Synth. Catal. 347 (2005) 1356–1360. [10] F. Li, Q. Zhang, Y. Wang, Size dependence in solvent-free aerobic oxidation of alcohols catalyzed by zeolite-supported palladium nanoparticles, Appl. Catal. A 334 (2008) 217–226. [11] G. Wu, X. Wang, N. Guan, L. Li, Palladium on graphene as efficient catalyst for solvent-free aerobic oxidation of aromatic alcohols: role of graphene support, Appl. Catal. B 136–137 (2013) 177–185.
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29 [12] M.L. Toebes, J.A. van Dillen, K.P. de Jong, Synthesis of supported palladium catalysts, J. Mol. Catal. A Chem. 173 (2001) 75–98. [13] Q. Wang, D.F. Shantz, Ordered mesoporous silica-based inorganic nanocomposites, J. Solid State Chem. 181 (2008) 1659–1669. [14] M. Opanasenko, P. Stepnicka, J. Cejka, Heterogeneous Pd catalysts supported on silica matrices, RSC Adv. 4 (2014) 65137–65162. [15] C.P. Mehnert, D.W. Weaver, J.Y. Ying, Heterogeneous Heck catalysis with palladium-grafted molecular sieves, J. Am. Chem. Soc. 120 (1998) 12289– 12296. [16] R.L. Oliveira, J.B.F. Hooijmans, P.E. de Jongh, R.J.M. Klein Gebbink, K.P. de Jong, Stabilization of palladium catalysts for the Heck reaction by support functionalization and solvent selection, ChemCatChem 6 (2014) 3223–3230. [17] C. Li, Q. Zhang, Y. Wang, H. Wan, Preparation, characterization and catalytic activity of palladium nanoparticles encapsulated in SBA-15, Catal. Lett. 120 (2008) 126–136. [18] C.M. Yang, H.A. Lin, B. Zibrowius, B. Spliethoff, F. Schüth, S.C. Liou, M.W. Chu, C. H. Chen, Selective surface functionalization and metal deposition in the micropores of mesoporous silica SBA-15, Chem. Mater. 19 (2007) 3205–3211. [19] C.J. Lin, S.H. Huang, N.C. Lai, C.M. Yang, Efficient room-temperature aqueousphase hydrogenation of phenol to cyclohexanone catalyzed by Pd nanoparticles supported on mesoporous MMT-1 silica with unevenly distributed functionalities, ACS Catal. 5 (2015) 4121–4129. [20] I. Yuranov, P. Moeckli, E. Suvorova, P. Buffat, L. Kiwi-Minsker, A. Renken, Pd/ SiO2 catalysts: synthesis of Pd nanoparticles with the controlled size in mesoporous silicas, J. Mol. Catal. A: Chem. 192 (2003) 239–251. [21] C.H. Liu, N.C. Lai, J.F. Lee, C.S. Chen, C.M. Yang, SBA-15-supported highly dispersed copper catalysts: vacuum–thermal preparation and catalytic studies in propylene partial oxidation to acrolein, J. Catal. 316 (2014) 231–239. [22] J.R.A. Sietsma, J.D. Meeldijk, J.P. den Breejen, M. Versluijs-Helder, A.J. Van Dillen, P.E. de Jongh, K.P. de Jong, The preparation of supported NiO and Co3O4 nanoparticles by the nitric oxide controlled thermal decomposition of nitrates, Angew. Chem. Int. Ed. 46 (2007) 4547–4549. [23] W. Zou, R.D. Gonzalez, The preparation of silica supported Pd catalysts: the effect of pretreatment variables on particle size, Catal. Lett. 12 (1992) 73–86. [24] I. Yuranov, L. Kiwi-Minsker, P. Buffat, A. Renken, Selective synthesis of Pd nanoparticles in complementary micropores of SBA-15, Chem. Mater. 16 (2004) 760–761.
29
[25] C.H. Liu, C.Y. Lin, J.L. Chen, N.C. Lai, C.M. Yang, J.M. Chen, K.T. Lu, Metal oxidecontaining SBA-15-supported gold catalysts for base-free aerobic homocoupling of phenylboronic acid in water, J. Catal. 336 (2016) 49–57. [26] C.H. Liu, N.C. Lai, S.C. Liou, M.W. Chu, C.H. Chen, C.M. Yang, Deposition and thermal transformation of metal oxides in mesoporous sba-15 silica with hydrophobic mesopores, Micropor. Mesopor. Mater. 179 (2013) 40–47. [27] R. Palkovits, C.M. Yang, S. Olejnik, F. Schüth, Active sites on SBA-15 in the Beckmann rearrangement of cyclohexanone oxime to e-caprolactam, J. Catal. 243 (2006) 93–98. [28] W.K. Józ´wiak, T.P. Maniecki, Influence of atmosphere kind on temperature programmed decomposition of noble metal chlorides, Thermochim. Acta 435 (2005) 151–161. [29] A. Savara, C.E. Chan-Thaw, I. Rossetti, A. Villa, L. Prati, Benzyl alcohol oxidation on carbon-supported Pd nanoparticles: elucidating the reaction mechanism, ChemCatChem 6 (2014) 3464–3473. [30] J.-D. Grunwaldt, M. Caravati, A. Baiker, Oxidic or metallic Palladium: which is the active phase in Pd-catalyzed aerobic alcohol oxidation?, J Phys. Chem. B 110 (2006) 25586–25589. [31] A.F. Lee, S.F.J. Hackett, J.S.J. Hargreaves, K. Wilson, On the active site in heterogeneous palladium selox catalysts, Green Chem. 8 (2006) 549–555. [32] A.F. Lee, J.N. Naughton, Z. Liu, K. Wilson, High-pressure XPS of crotyl alcohol selective oxidation over metallic and oxidized Pd(111), ACS Catal. 2 (2012) 2235–2241. [33] A.F. Lee, C.V. Ellis, J.N. Naughton, M.A. Newton, C.M.A. Parlett, K. Wilson, Reaction-driven surface restructuring and selectivity control in allylic alcohol catalytic aerobic oxidation over Pd, J. Am. Chem. Soc. 133 (2011) 5724–5727. [34] C.V. Gaskell, C.M.A. Parlett, M.A. Newton, K. Wilson, A.F. Lee, Redox-controlled crotyl alcohol selective oxidation: in situ oxidation and reduction dynamics of catalytic Pd nanoparticles via synchronous XANES/MS, ACS Catal. 2 (2012) 2242–2246. [35] C.M.A. Parlett, L.J. Durndell, A. Machado, G. Cibin, D.W. Bruce, N.S. Hondow, K. Wilson, A.F. Lee, Alumina-grafted SBA-15 as a high performance support for Pd-catalysed cinnamyl alcohol selective oxidation, Catal. Today 229 (2014) 46–55.
Contents lists available at ScienceDirect
Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat
SBA-15-supported Pd catalysts: The effect of pretreatment conditions on particle size and its application to benzyl alcohol oxidation Chun-Hsia Liu a, Chih-Yu Lin b, Jeng-Lung Chen a, Kueih-Tzu Lu a, Jyh-Fu Lee a, Jin-Ming Chen a,⇑ a b
National Synchrotron Radiation Research Center, Hsinchu 30076, Taiwan Department of Chemistry, National Tsing Hua University, Hsinchu 30013, Taiwan
a r t i c l e
i n f o
Article history: Received 11 November 2016 Revised 30 December 2016 Accepted 24 January 2017
Keywords: Palladium catalyst Mesoporous material Alcohol oxidation X-ray absorption spectra
a b s t r a c t SBA-15-supported Pd catalysts were prepared by impregnation, thermal pretreatment, and H2 reduction. The pretreatment conditions have an important influence on the resulting size (from 1.3 to 10.3 nm) of the Pd particles; the catalyst prepared under vacuum conditions possessed the smallest particles (size 1.3 nm). X-ray absorption spectra were used to analyze the transformation of the Pd precursor in each step of the preparation. These catalysts were applied to the oxidation of benzyl alcohol by molecular oxygen; the catalyst with the smallest Pd particles exhibited the highest turnover frequency (9684 h 1). The catalyst was reusable three times without loss of activity (with conversion >96%). X-ray absorption spectra were recorded ex situ to monitor the oxidation and the coordination structure of Pd at varied temperatures of reaction; when the reaction temperature was above 100 °C, the Pd species were in the metallic state. This result indicates that metallic Pd might be the active phase in the oxidation of benzyl alcohol by molecular oxygen. Ó 2017 Elsevier Inc. All rights reserved.
1. Introduction The oxidation of alcohols to aldehydes or ketones is an important reaction in the production of fine and specialty chemicals [1,2]. In contrast to a traditional method using stoichiometric oxidation with mineral oxidizing agents, a recently developed method using a noble metal (e.g., Ru, Pt, Pd, or Au) as a catalyst and molecular oxygen as the oxidant constitutes a sustainable and green process [1,2]. Among those catalysts, Pd-containing catalysts appear to be promising because both high activity and high selectivity can be obtained concurrently [3–11]. As Pd is recognized to be a scarce resource, the synthesis of a Pd catalyst with high dispersion and high efficiency is an attractive topic in catalytic research [12–15]. Among various methods of synthesizing a supported Pd catalyst are impregnation [7,12,16,17], a ligand-assisted method [14–16,18,19], and an ion-exchange method [4,12,17,20]. With the latter two methods and on improving the interaction of the Pd precursor and the support through grafting a hydrophilic or hydrophobic functional group (e.g., ANH, ASOH, ACH@CH2, or ACH3) or adjusting the surface charge of the support, catalysts consisting of Pd nanoparticles (size <2 nm) were synthesized [4,12,14–20], but the impregnation ⇑ Corresponding author. E-mail address: [email protected] (J.-M. Chen). http://dx.doi.org/10.1016/j.jcat.2017.01.019 0021-9517/Ó 2017 Elsevier Inc. All rights reserved.
method yielded Pd catalysts that generally showed a broad distribution of size of Pd particles larger than 2 nm [7,12,16,17,19]. A synthesis of a Pd catalyst that possesses a narrow distribution of particle size <2 nm with a simple impregnation method is attractive [16]. When we synthesized SBA-15-supported copper catalysts by a vacuum-thermal method, we found that the vacuum condition was vital for preparing highly dispersed copper catalysts; the transformation of the copper species during the heat treatment was altered by the vacuum condition [21]. The atmosphere during the thermal decomposition is reported to be an important factor in tuning the resulting metal or metal-oxide size in supported catalysts [12,22–24]. Sietsma et al. found that gaseous dinitrogen monoxide prevented the rapid decomposition of Ni3(NO3)2(OH)4 that is observed in air, resulting in smaller NiO nanoparticles than those prepared in air [22]. Zou et al. discovered that, when the chemical structure of the adsorbed palladium complex is varied during the pretreatment (i.e., with H2, O2, or N2), the resulting silica-supported Pd catalysts possessed varied particle sizes [23]. Yuranov et al. showed that reducing the adsorbed palladium species before oxidation gave the resulting SBA-15-supported PdO catalyst smaller PdO nanoparticles than without that reduction [24]. To the best of our knowledge, there is no report of supported Pd catalysts from a vacuum-thermal preparation.
22
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
Here we report the preparation of SBA-15-supported Pd catalysts with an impregnation method and subsequent varied pretreatment conditions and H2 reduction, and tests of oxidation of benzyl alcohol with molecular oxygen. SBA-15 was selected as a support because it has great chemical stability, a large surface area for metal dispersion, and porosity of nanometer size to facilitate molecular diffusion [18,21,25]. The pretreatment condition was found to have an important influence on the resulting Pd particle size (from 1.3 to 10.3 nm); among those catalysts the catalyst pretreated under vacuum conditions possessed the smallest particles (1.3 nm). X-ray absorption spectroscopy (XAS) served for analysis of the transformation of the Pd precursor in each preparation step. These catalysts were applied to the oxidation of benzyl alcohol with molecular oxygen, which showed that the catalyst with the smallest Pd particles exhibited the greatest activity. The catalyst was reusable three times without loss of activity (with conversion >96%). XAS measurements were recorded ex situ to monitor the oxidation and the coordination structure of Pd at varied reaction temperatures.
2. Experiments
drate (ZrOCl28H2O) in hydrochloric acid (HCl). The mixture was stirred at 35 °C for 24 h. After sulfuric acid (H2SO4) was added to the mixture in a calculated proportion to remove the triblock copolymer [26], the mixture was aged at 90 °C for 24 h. The molar proportions of the final mixture were 1 TEOS:0.017 P123:5.9HCl:193 H2O:2.9 H2SO4:0.05ZrOCl2. The solid product was washed with propanone, filtered, and dried in air. The product was heated further at 350 °C to produce the SBA-15 support. To impregnate Pd nanoparticles, we loaded the SBA-15 support with a solution of dichlorobis(acetonitrile) palladium (PdCl2(MeCN)2) in propanone (nominally 1 mass% Pd by an incipient wetness method) and dried it. The sample was subsequently heated under varied gases (hydrogen, air, nitrogen), or vacuum, at 350 °C for 2 h with a ramp rate of 1 °C min 1. The resulting samples are referred to as Sample X, in which X denotes the conditions of thermal treatment (with X = H2, air, N2, or Vac for treatment under hydrogen, air, nitrogen, or vacuum (ca. 0.1 Torr), respectively). Each sample X was further reduced under pure hydrogen at 350 °C for 2 h. After cooling to 23 °C, the resulting samples were stored under atmospheric conditions. The resulting samples are referred to as Sample X–H2, in which X denotes the conditions of thermal treatment (with X = Air, N2, and Vac for pretreatment in air, nitrogen or vacuum, respectively).
2.1. Preparation of the catalyst The synthesis of mesoporous SBA-15 silica involved the addition of tetraethoxysilane (TEOS) to a solution of triblock copolymer (Pluronic P123, EO20PO70EO20) and zirconium oxychloride octahy-
2.2. Characterization of the catalyst X-ray diffraction (XRD) patterns were recorded on a diffractometer (Mac Science 18MPX) using Cu Ka radiation. N2 physisorption isotherms were measured (QuantachromeAutosorb-1-MP) at 77 K. The isotherms were analyzed with nonlocal densityfunctional theory (NLDFT) to evaluate the pore sizes of the samples using the kernel of NLDFT equilibrium capillary condensation isotherms of nitrogen at 77 K on silica (adsorption branch, assuming cylindrical pore geometry). The BET surface areas were calculated from the adsorption branches in the relative pressure range 0.05–0.30; the total pore volumes were evaluated at relative pressure 0.95. Inductively coupled plasma mass spectral (ICP-MS) data were recorded (Perkin–Elmer SCIEX-ELAN 5000). Transmission electron microscopy (TEM) images were taken from ultramicrotomed samples (thickness 80–100 nm) with an electron microscope (JEOL JEM-ARM 200 FTH) equipped with an energydispersive X-ray (EDX) spectrometer and a high-angle annular dark-field (HAADF) detector. X-ray photoelectron spectra (XPS) were recorded on an Ulvac-PHI PHI Quantera device. 2.3. XAS
Fig. 1. Small-angle and wide-angle XRD patterns of SBA-15 and SBA-15-supported Pd materials.
Pd K-edge XAS were recorded at beamline 01C1 at Taiwan Light Source (storage-ring energy 1.5 GeV, ring current 360 mA) in the National Synchrotron Radiation Research Center. In a typical experiment, the Pd catalyst (about 120 mg) was pressed to form self-supporting wafers and mounted on a sample holder for measurements. Spectra were recorded near 23 °C in a transmission
Table 1 Physical properties of materials.a.
a b c
Sample
a (nm)
D (nm)
W (nm)
S (m2 g
SBA-15 H2 Air-H2 N2-H2 Vac-H2
12.1 12.1 12.1 12.1 12.1
8.1 8.1 8.1 8.1 8.1
4.0 4.0 4.0 4.0 4.0
1023 1001 1022 1008 1024
1
)
Vt (cm3 g 1.38 1.31 1.31 1.31 1.31
a: unit cell parameter; D: mesopore diameter; W: wall thickness; S: BET surface area; Vt: total pore volume. DPd: Pd particle size calculated from XRD pattern. DPd: average particle size observed by TEM.
1
)
DPdb (nm)
DPdc (nm)
— 5.6 4.5 10.2 —
— 6.6 5.2 10.3 1.3
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
23
mode with two gas-filled ionization chambers in series to measure the intensities of the incident beam and the transmitted beam. A Pd foil that served as a reference for energy calibration was measured simultaneously with a third ionization chamber. X-ray absorption near-edge structure (XANES) spectra and extended X-ray absorption fine-structure (EXAFS) spectra were all recorded near 23 °C. For the measurement of the XAS of the Pd catalyst under working conditions, spectra were recorded ex situ under ambient conditions and in a fluorescent mode. In a typical experiment, a mixture of benzyl alcohol (0.87 mL, 8.4 mmol), p-xylene (2.0 mL), and Sample Vac–H2 (9 mg, 0.01 mol.% of Pd relative to benzyl alcohol) was stirred (800 rpm) in flowing O2 (20 sccm) at a reaction temperature in a range from 25 to 140 °C for 0.5 h. After the reaction, the flow of the oxygen was stopped and the solution was cooled near 23 °C under ambient conditions for the measurement of XAS. For the in situ Pd K-edge XAS measurement of the Pd catalyst in pure H2, a custom-made stainless steel cell was used. Two holes were made in the cell, one on the front and the other on the back. After the solid samples were placed inside the cell, the holes were closed with Kapton film to avoid exposure of the sample to the ambient atmosphere. Before the XAS measurement, the samples were reduced at 350 °C in pure H2 and then cooled to room temperature for the measurement. 2.4. Testing the catalyst In a typical reaction, a mixture of benzyl alcohol (0.87 mL, 8.4 mmol), p-xylene (2.0 mL), and the SBA-15-supported Pd catalyst (9 mg, 0.01 mol.% of Pd relative to benzyl alcohol) was stirred (at 800 rpm) under flowing O2 (20 sccm) at 120 °C for 3 h. After the reaction, the solid catalyst was filtered off while the liquid phase (with decane added as internal standard) was analyzed on a gas chromatograph (Shimadzu GC-2014 equipped with a FID detector and a SGE BP-5 column). To test reuse, the catalysts were collected with centrifugation, washed with p-xylene, dried, and then reused in further runs. 3. Results and discussion 3.1. Characterization of SBA-15-supported Pd materials Mesoporous SBA-15 with short mesochannels was prepared as a support for the deposition of palladium. The synthesis involved the addition of Zr(IV) salt to the synthesis solution, which was removed on the addition of sulfuric acid to the synthesis solution before the aging step [26]. The Zr/Si ratio of the resulting SBA-15 was 0.002 (i.e., 0.4 mass% ZrO2) determined with ICP-MS; the remaining Zr species were inaccessible to the acidic solution and would be inside the silica framework rather than on the mesopore surface. The resulting SBA-15 support exhibited a highly ordered p6mm mesostructure with unit cell parameter a = 12.1 nm and uniform mesopores of width 8.1 nm (see Fig. 1 for the XRD pattern, Fig. S1 for the N2 physisorption isotherm, and Table 1). The SBA-15 support was impregnated with PdCl2(MeCN)2 in propanone solution and heated at 350 °C under varied atmospheres. ICP-MS analysis indicated that all Pd-containing materials had Pd about 0.95 mass%, near the nominal value 1 mass% Pd. The small-angle XRD data in Fig. 1 confirmed the retention of the ordered p6mm mesostructure after the Pd deposition and thermal treatment. The wide-angle XRD data revealed that the Pd species in Sample H2 consisted of Pd nanocrystals (5.6 nm, JCPDS file 870638). An XRD signal contributed from PdO (JCPDS file 75-584) was observed in the XRD pattern of Sample Air, which indicated the formation of PdO. Signals contributed from Pd and PdO were
Fig. 2. Bright-field TEM images of (a) sample H2, (b) sample Air–H2, (c) sample N2– H2, and (d) sample Vac–H2.
observed in the XRD pattern of Sample N2, indicating the formation of metallic Pd and PdO simultaneously after heating under an inert gas. In the XRD pattern of Sample Vac, no signal was contributed from a Pd species after the thermal treatment, indicating the formation of amorphous Pd species or highly dispersed nanoparticles.
24
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
Fig. 3. Dark-field TEM images of sample Vac–H2 viewed along (a, b) and perpendicular to (c, d) the axis of the hexagonally arranged mesopores. Scale bars indicate 20 nm.
After the subsequent reduction at 350 °C, the signal contributed by PdO in the XRD patterns of Samples Air and N2 vanished and signals from metallic Pd were observed. The Pd particle sizes calculated with the Scherrer equation were 4.5 nm for Sample Air–H2 and 10.2 nm for Sample N2–H2. For the XRD pattern of Sample Vac–H2, no additional peaks were formed after the reduction in H2. In N2 physisorption isotherms of the Sample X–H2 series (see Fig. S1), they exhibited sharp steps and hysteresis loops of type H1 and isotherms similar to that of the SBA-15 support. The minor change of the isotherms occurred because of the small content of Pd. The reduced samples were characterized with a TEM to observe the Pd nanoparticles; the results appear in Figs. 2 and S2. In four samples, hexagonally arranged mesopores of channel type were clearly observed; the average pore diameter was near 8 nm, in agreement with the value derived from the N2 physisorption data (see Table 1). For the sample with direct reduction (i.e., Sample H2) the Pd particles were large (7 nm) and aggregation of the particles was observed. For Sample Air–H2, the Pd particle size decreased to 5.0 nm. For Sample N2–H2, the Pd particles were larger than for Sample H2 and were inhomogeneously dispersed in the silica support. The distribution of particle sizes was broad, from 4 to 20 nm. For Sample Vac–H2, no discrete Pd nanoparticles were observed in the silica support, although the EDS data confirmed the presence of Pd. We further characterized the Sample Vac–H2 with TEM-HAADF viewed along and perpendicular to the axis of the hexagonal mesopores; the results appear in Figs. 3 and S3. Bright spots, in the range 0.6–2.1 nm with average size 1.3 nm, were homogeneously distributed in the SBA-15 support. Most spots were incorporated into the wall of SBA-15. The particles in Sample Vac–H2 had a narrow distribution of particle size, average size 1.3 nm (cf. Figs. 3 and S2). Overall, the Pd nanoparticle size decreased in the sequence Sample N2–H2 (10.3 nm) > Sample H2 (6.6 nm) > Sample Air–H2 (5.2 nm) > Sample Vac–H2 (1.3 nm). It should be noticed that for the synthesis of the catalysts with small Pd size (<2 nm) and a narrow distribution of particle size, the exis-
tence of a functional group or adjustment of the surface charge of the support is generally required to improve the interaction between the Pd precursor and the support [4,12,14,16–18]. It is interesting that the impregnation method and pretreatment under vacuum achieved a narrow distribution of particle size with Pd size <2 nm. We recorded XPS to characterize the oxidation state of Pd; Fig. S4 shows XPS of the reduced samples in the Pd3d region. For Sample H2 and Sample N2–H2, the signals of binding energy were observed at 340.8 and 335.5 eV, corresponding to Pd3d3/2 and Pd3d5/2, respectively. Pd in both samples was in metallic form after reduction with hydrogen at 350 °C. For Samples Air–H2 and Vac– H2, the signals of binding energy were observed at 342.1, 340.8, 336.8, and 335.5 eV, which are assigned to PdO3d3/2, Pd3d3/2, PdO3d5/2, and Pd3d5/2, respectively: both metallic Pd and PdO were hence present in the samples. Under such reduction, the Pd species are expected to be totally reduced, so that the formation of PdO is expected to be associated with the reoxidation of the surface of the metallic Pd nanoparticles when the Pd materials were stored under air [4,19]. The result indicated that the Pd nanoparticles formed in Samples Air–H2 and Sample Vac–H2 were smaller than those in Samples H2 and N2–H2, consistent with the TEM analysis. The inverse correlation between Pd particle size and oxidation state over SBA-15 in as-prepared catalysts was reported [4]. The molar ratios of surface PdO in total palladium were 18% for Sample Air– H2 and 28 % for Sample Vac–H2. 3.2. XAS of SBA-15-supported Pd materials According to our experimental results, the pretreatment conditions of PdCl2(MeCN)2 were important for the formation of small and highly dispersed Pd nanoparticles in SBA-15. We further characterized with XAS the SBA-15-supported Pd catalysts before and after pretreatment and subsequent hydrogen reduction to evaluate the transformation of the oxidation state and the coordination
25
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
Fig. 4. (a) Pd K-edge XANES spectra and (b) Fourier transform profiles of Pd K-edge k3-weighted EXAFS data (-) and fitted results () of SBA-15-supported Pd samples before and after thermal pretreatments. Data ranges for Fourier transformation (without phase correction) and fitting are Dk = 3.2–13.6 Å 1 and DR = 1.0–3.8 Å, respectively.
structure of palladium at each preparation step. Fig. 4 shows Pd Kedge XANES spectra and the corresponding Fourier transform (FT) profiles of k3-weighted EXAFS data of the impregnated sample, Sample Air, Sample N2, and Sample Vac with the reference spectra of PdCl2 (MeCN)2, PdCl2, Pd foil, and PdO for comparison. The structural parameters from the EXAFS fit of samples are listed in Table 2.
After the SBA-15 support was impregnated with the solution of PdCl2(MeCN)2 in propanone and dried under atmospheric conditions, the XANES spectrum of the impregnated sample (labeled imp.) showed an absorption-edge profile different from that of PdCl2(MeCN)2, indicating a change of the coordination structure of the impregnated Pd(II) complex. The intensities of the signals at 24,369 and 24,386 eV in the spectrum decreased; the spectrum exhibited an absorption-edge profile similar to that of PdCl2 rather than of PdCl2(MeCN)2 (see Fig. 4a). In its FT profile and fitted value (see Fig. 4b and Table 2), the intensity of the signal attributed to PdAN coordination (R = 2.00 Å, NPdAN = 1.1) decreased while the intensity of the signal attributed to PdACl coordination (R = 2.30 Å, NPdACl = 2.9) increased, relative to the signals attributed to PdAN coordination (R = 1.95 Å, NPdAN = 2.0) and PdACl coordination (R = 2.31 Å, NPdACl = 2.0) in the spectrum of PdCl2(MeCN)2. These results indicated that, when the PdCl2(MeCN)2 complex was impregnated in the pores of the SBA-15 support, one MeCN ligand in PdCl2(MeCN)2 left and the Pd center became further coordinated with a chlorine atom from a nearby Pd complex. The elimination of MeCN from PdCl2(MeCN)2 might be due to the weakness of the PdAN bond in that a p-electron associated with the C„N triple bond was withdrawn by the Lewis acid site on the surface of SBA-15 [27]. The impregnated sample was further subjected to thermal treatments under air, N2, or vacuum. After the thermal treatment in air, the XANES spectrum of Sample Air changed. The first signal at 24,371 eV became sharper and shifted to decreased energy. In its FT profile and fitted value (see Fig. 4b and Table 2), the intensity of the signal attributed to PdACl coordination (R = 2.30 Å, NPdACl = 0.5) decreased and two signals appeared, attributed to the PdAO shell (R = 2.03 Å, NPdAO = 3.0) and the PdAPd shell (R = 3.03 Å, NPdAPd = 1.1). The fitted result indicated that, after the thermal treatment in air at 350 °C, the Pd species in Sample Air had mostly a local structure of PdO type, but some PdACl bonds remained (see Fig. 4b and Table 2). After the thermal treatment under N2, the XANES spectrum of Sample N2 showed an absorption edge profile similar to that of the Pd foil, indicating the reduction of Pd(II). The FT profile exhibited a sharp signal attributed to PdAPd coordination (R = 2.74 Å, NPdAPd = 7.3) and a broad signal attributed to PdAO coordination (R = 2.00 Å, NPdAO = 0.7) (see Fig. 4b and Table 2). The reduction of the Pd(II) complex was related to the dechlorination–reduction during the thermal treatment under an inert gas [28]. The Pd species in Sample N2 was mostly in the metallic form, but some PdAO bonds were also present. After the thermal treatment under vacuum, the XANES spectrum of Sample Vac showed an absorption-edge profile
Table 2 Structural parameters of the Pd catalysts after heat treatment.a. Samples
Shell
C.N.
R (Å)
r2 (Å2)
Eshift (eV)
R-factor (%)
PdCl2(MeCN)2
PdAN PdACl PdAN PdACl PdAO PdACl PdAPd PdAPd PdAO PdAPd PdAO PdAPd PdACl PdAPd PdAO PdAPd PdAPd
2.0 2.0 1.1 2.9 3.0 0.5 1.1 0.5 0.7 7.3 1.0 2.8 4.0 12.0 4.0 4.0 8.0
1.95 2.31 2.00 2.30 2.03 2.30 3.03 3.42 2.00 2.74 2.01 2.70 2.31 2.75 2.03 3.03 3.42
0.003 0.003 0.003 0.005 0.006 0.001 0.006 0.001 0.002 0.006 0.003 0.009 0.004 0.005 0.003 0.005 0.005
5.3 3.3 0.1 6.9 5.7 6.0 1.4 1.1 8.0 3.2 17.7 6.0 22.1 6.2 5.6 12.3 11.4
1.4
Imp. Air
N2 Vac PdCl2 Pd PdO
a
C.N.: coordination number; R: bond length; r2: Debye–Waller factor; Eshift: inner potential correction.
0.8 3.4
1.7 5.0 3.0 1.9 1.4
26
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
different from that of the impregnated sample. In its FT profile, there were two broad signals attributed to PdAO coordination (R = 2.01 Å, NPdAO = 1.0) and PdAPd coordination (R = 2.70 Å, NPdAPd = 2.8). The lack of signal associated with PdACl coordination indicated the total decomposition of the Pd precursor. For the subsequently reduced samples, their Pd K-edge XANES spectra and the corresponding FT profiles of k3-weighted EXAFS data are shown in Fig. 5 with reference spectra of Pd foil and PdO for comparison. Table 3 summarizes the structural parameters from the EXAFS fit of those samples. After the subsequent reduction and storage under atmospheric conditions, the Sample X–H2 series were measured in air. The XANES spectra of Sample H2 and Sample N2–H2 showed absorption-edge profiles similar to that of metallic Pd foil. Their FT profiles showed only a sharp signal attributed to PdAPd coordination (R = 2.74 Å), indicating that the Pd species in these two samples were reduced to metallic form.
For Samples Air–H2 and Vac–H2, the signals at 24,367 and 24,390 eV in their XANES spectra were less sharp than those in the spectrum of Pd foil. In their FT profiles, a broad signal attributed to PdAO coordination and a signal attributed to PdAPd coordination were observed. These results indicated that the Pd species in these samples were Pd and PdO. To elucidate the oxidation state of palladium after the hydrogen reduction in 350 °C, in situ Pd Kedge XAS measurements were performed (see Fig. S5 and Table S1). The result showed that before exposure to the air the Pd species in Samples Air–H2 and Vac–H2 were in the metallic state and the Pd species in these samples was only Pd. As with the previous observation in the XPS results, the formation of PdO is expected to be associated with the reoxidation of the surface of the small metallic Pd nanoparticles when the Pd materials were stored under air [4,19]. 3.3. Catalyst performance
Fig. 5. (a) Pd K-edge XANES spectra and (b) Fourier transform profiles of Pd K-edge k3-weighted EXAFS data (–) and fitted results () of SBA-15-supported Pd samples after thermal reduction. Data ranges for Fourier transformation (without phase correction) and fitting are Dk = 3.2–13.6 Å 1 and DR = 1.0–3.8 Å, respectively.
We applied the reduced samples (i.e., Samples H2, Air–H2, N2–H2 and, Vac–H2) to the oxidation of benzyl alcohol at 120 °C; the corresponding profiles of conversion of benzyl alcohol versus time are exhibited in Fig. 6. In that figure, the conversion of benzyl alcohol increased sharply in the first 2 h and then gradually during the following 1 h. The catalytic results are summarized in Table 4, which shows the results for conversion of benzyl alcohol, product selectivity, and turnover frequency (TOF). The catalysts produced benzaldehyde as the main product and toluene as the main byproduct. The selectivities of benzaldehyde for these catalysts were >85%. Among four catalysts, Sample Vac–H2 showed the highest activity; the conversion achieved 99% in 3 h. Blank reactions were also carried out and the corresponding conversion was <2% under our reaction conditions. When this sample was also reused three times under the same reaction conditions, the conversion remained above 96%. Overall, the trend of TOF was Sample Vac–H2 > Sample Air–H2 > Sample H2 > Sample N2–H2, which shows a trend the same as for the size of Pd nanoparticle: Sample Vac–H2, which possessed the smallest Pd size, exhibited the highest TOF. Sample Vac–H2 was further tested at varied reaction temperatures; the catalytic results are shown in Table 5 and Fig. S6. As the reaction temperature increased, the selectivity of benzaldehyde decreased a little from 94 to 86%. For temperatures <100 °C, the conversion and TOF were moderate (i.e., 47% and 383 h 1). When the reaction temperature was increased to 100 °C, the conversion and TOF increased sharply to 82% and 3953 h 1. The increased TOF might be due to the reduction of surface PdO to metallic Pd by benzyl alcohol at 100 °C. When the reaction temperature was further increased to 130 °C, the conversion and TOF increased steadily to 99% and 10,672 h 1, but when the reaction temperature became 140 °C, the conversion and TOF abruptly decreased to 79% and 6127 h 1. Such decreased catalytic activity
Table 3 Structural parameters of the Pd catalysts after H2 reduction.a. Samples
Shell
C.N.
R (Å)
r2 (Å2)
H2 Air–H2
PdAPd PdAO PdAPd PdAPd PdAO PdAPd PdAPd PdAO PdAPd PdAPd
7.8 1.4 4.1 8.7 1.0 2.5 12.0 4.0 4.0 8.0
2.74 2.00 2.73 2.74 2.02 2.70 2.75 2.03 3.03 3.42
0.006 0.004 0.005 0.006 0.001 0.009 0.005 0.003 0.005 0.005
N2–H2 Vac–H2 Pd PdO
a
C.N.: coordination number; R: bond length; r2: Debye–Waller factor; Eshift: the inner potential correction.
Eshift (eV) 4.8 7.3 4.7 4.6 20.4 6.4 6.2 5.6 12.3 11.4
R-factor (%) 0.2 0.6 0.3 5.0 1.9 1.4
27
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
the ion-exchange method (90 °C, 1120 h 1) [4], as both catalysts possessed a similar Pd size (1.9 nm), but, compared with the carbon-supported Pd catalysts, such as Pd@N-doped carbon (120 °C, 153,551 h 1) [5] and Pd/Graphene (110 °C, 30,137 h 1) [11], the oxide-supported Pd catalysts showed poor activity, which was due to the strong adsorption of benzyl alcohol on the supports [5,11].
3.4. XAS of sample Vac–H2 for varied reaction temperature
Fig. 6. Profiles of benzyl alcohol conversion vs. time. Reaction conditions: benzyl alcohol (0.87 mL, 8.4 mmol), SBA-15-supported Pd catalyst (9 mg, 0.01 mol.% of Pd relative to benzyl alcohol), and p-xylene (2.0 mL) under flowing O2 (20 sccm) at 120 °C.
at a high reaction temperature is unexpected; this phenomenon might have resulted from the oxidation of metallic Pd nanoparticles [1] or the decreased adsorption property of Pd nanoparticles for benzyl alcohol [29]. The activity of Sample Vac–H2 is greater than for some reported supported Pd catalysts, such as Pd/Al2O3 (120 °C, 3615 h 1) [7], Pd/Al2O3 (100 °C, 3300 h 1) [9], and Pd/ NaX zeolite (100 °C, 626 h 1) [10]. It showed a TOF similar to that for the reported SBA-15-supported Pd catalyst synthesized with
As it is generally recognized that the oxidation state of Pd greatly affects the catalytic activity [7,30,31], we recorded the XAS of Sample Vac–H2 ex situ after the reaction at varied temperatures in the range 25–140 °C to observe the oxidation and the coordination structure of Pd under working conditions. After each reaction for 30 min, part of the solution was extracted and cooled in air to near 23 °C for the recording of XAS in a fluorescent mode. The results are shown in Fig. 7; the structural parameters from the EXAFS fits are listed in Table S2. When the reaction temperature was < 100 °C, the oxidation state and the coordination structure of the Pd species were constant; the XANES spectra and the FT profiles were nearly identical to those of Sample Vac–H2 before the reaction. When the reaction temperature attained 100 °C, the XANES spectrum showed a metallic feature at 24,390 eV; in its FT profile, the intensity of the signal attributed to PdAPd coordination (R = 2.76 Å, NPdAPd = 3.1) increased whereas the intensity of the signal attributed to PdAO coordination (R = 2.01 Å, NPdAO = 1.3) decreased. This result indicated the reduction of Pd(II) to Pd(0) at 100 °C. When the reaction temperature was increased to 140 °C, the metallic nature and the size of the Pd particles did not change; the XANES spectra and the FT profiles after 110–140 °C are similar. This observation indicated that, for reaction temperatures >100 °C, metallic Pd was the major phase, and it was suspected to be the active phase under our reaction conditions; for reaction temperatures <100 °C, PdO was still present, which resulted in poor
Table 4 The catalytic performance of the SBA-15-supported Pd catalysts.a.
a b c
Conv.
Selectivity (%)
Catalyst
(%)
Benzaldehyde
Othersb
TOF (h
H2 Air–H2 N2–H2 Vac–H2
40 77 28 99
93 85 87 90
7 15 13 10
1779 3953 791 9684
1 c
)
Conv.: conversion; TOF: turnover frequency. Including toluene, benzoic acid, dibenzyl ether, and benzylbenzoate. Calculated as moles of benzyl alcohol converted per mole of total Pd per hour, measured at time on stream of 30 min.
Table 5 The catalytic performance of sample Vac–H2 at different reaction temperatures.a.
a b c
T
Conv.
Selectivity (%)
Catalyst
(°C)
(%)
Benzaldehyde
Othersb
(h
Vac–H2
80 90 100 110 120 130 140
8 47 82 94 99 99 79
94 95 86 90 90 87 86
6 5 14 10 10 13 14
296 1383 3953 8498 9684 10,672 6127
Conv.: conversion; TOF: turnover frequency. Including toluene, benzoic acid, dibenzyl ether, and benzylbenzoate. Calculated as moles of benzyl alcohol converted per mole of total Pd per hour, measured at time on stream of 30 min.
TOF 1 c
)
28
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29
4. Conclusions SBA-15-supported Pd catalysts were synthesized and applied to the oxidation of benzyl alcohol with molecular oxygen as oxidant. Depending on the thermal pretreatment, SBA-15-supported catalysts with Pd nanoparticles of size 1.3 to 10.3 nm were synthesized. The catalysts were applied to the oxidation of benzyl alcohol, which showed that Sample Vac–H2 with the smallest Pd possessed the largest TOF. The catalyst was reused three times without loss of activity (i.e., conversion >96%). From the catalytic results and the XAS ex situ, when the reaction temperature was greater than 100 °C, the Pd species in Sample Vac–H2 was in metallic form and the catalytic activity increased sharply, indicating that metallic Pd might be the active phase in the reactions. This result indicates that tuning the thermal pretreatment condition enables the synthesis of highly dispersed and active Pd catalysts; this method might be applied to synthesize other supported highly dispersed Pd catalysts for which the interaction of the Pd precursor and the support is weak. Electronic supplementary information (ESI) available includes N2 physisorption isotherms, size distributions from the TEM images, XPS data, the results of catalytic tests, and structural parameters from EXAFS fits. Acknowledgments We thank the NSRRC staff for their technical support. NSRRC and the Ministry of Science and Technology of Republic of China under grant Nos. MOST102-2113-M-213 -004 -MY3 and MOST105-2113-M-213 -005 -MY3 provided support of this research. Appendix A. Supplementary material 3
Fig. 7. (a) Pd K-edge XANES spectra and (b) Fourier transforms of Pd K-edge k weighted EXAFS data (–) and fitted results () of sample Vac–H2 at varied reaction temperature. Data ranges for Fourier transformation (without phase correction) and fitting are Dk = 3.2–10.7 Å 1 and DR = 1.0–3.8 Å, respectively.
catalytic activity. In addition, at 140 °C the metallic Pd nanoparticles were not oxidized or aggregated, while the catalytic activity decreased from 130 to 140 °C (see Tables 5 and S2). The decreased catalytic activity was obviously not related to the modification of the active phase at a high reaction temperature. The XAS results ex situ indicate that metallic Pd might be the active phase at a high reaction temperature (i.e., 100–140 °C). In the literature, it had been suspected that metallic Pd was the active phase. A tentative reaction mechanism on the surface of metallic Pd was proposed: An oxidative addition of an OAH bond from alcohol to the metallic Pd affords a Pd–alcoholate species and the species undergoes b-hydride elimination to produce the corresponding aldehyde [2,3,5,6,10]. Nevertheless, in recent surface analysis, high-pressure XPS measurements over metallic and oxidized Pd(1 1 1) single crystals suggested that aldehyde formation was disfavored over metallic Pd(1 1 1) but favored on oxidized Pd(1 1 1) surfaces [32]. The results of time-resolved liquid and vapor phase XAS measurements also revealed that surface palladium oxide was the catalytically active phase [31,33–35]. The alcohol reacted with the surface palladium oxide on the metallic Pd to form the corresponding aldehyde. To determine the active site, XAS results ex situ might not be sufficient, because the surface palladium oxide was unstable and might be further reduced by benzyl alcohol under our measurement conditions [32]. Further experiments need to be performed to distinguish the active site and elucidate the mechanism of benzyl alcohol oxidation in our catalytic condition.
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcat.2017.01.019. References [1] T. Mallat, A. Baiker, Oxidation of alcohols with molecular oxygen on solid catalysts, Chem. Rev. 104 (2004) 3037–3058. [2] Z. Guo, B. Liu, Q. Zhang, W. Deng, Y. Wang, Y. Yang, Recent advances in heterogeneous selective oxidation catalysis for sustainable chemistry, Chem. Soc. Rev. 43 (2014) 3480–3524. [3] J. Chen, Q. Zhang, Y. Wang, H. Wan, Size-dependent catalytic activity of supported palladium nanoparticles for aerobic oxidation of alcohols, Adv. Synth. Catal. 350 (2008) 453–464. [4] C.M.A. Parlett, D.W. Bruce, N.S. Hondow, M.A. Newton, A.F. Lee, K. Wilson, Mesoporous silicas as versatile supports to tune the palladium-catalyzed selective aerobic oxidation of allylic alcohols, ChemCatChem 5 (2013) 939– 950. [5] P. Zhang, Y. Gong, H. Li, Z. Chen, Y. Wang, Solvent-free aerobic oxidation of hydrocarbons and alcohols with Pd@N-doped carbon from glucose, Nat. Commun. 4 (2013) 1593–1604. [6] K. Mori, T. Hara, T. Mizugaki, K. Ebitani, K. Kaneda, Hydroxyapatite-supported palladium nanoclusters: a highly active heterogeneous catalyst for selective oxidation of alcohols by use of molecular oxygen, J. Am. Chem. Soc. 126 (2004) 10657–10666. [7] X. Wang, G. Wu, N. Guan, L. Li, Supported Pd catalysts for solvent-free benzyl alcohol selective oxidation: effects of calcination pretreatments and reconstruction of Pd sites, Appl. Catal. B 115–116 (2012) 7–15. [8] Y. Chen, Z. Guo, T. Chen, Y. Yang, Surface-functionalized TUD-1 mesoporous molecular sieve supported palladium for solvent-free aerobic oxidation of benzyl alcohol, J. Catal. 275 (2010) 11–24. [9] H. Wu, Q. Zhang, Y. Wang, Solvent-free aerobic oxidation of alcohols catalyzed by an efficient and recyclable palladium heterogeneous catalyst, Adv. Synth. Catal. 347 (2005) 1356–1360. [10] F. Li, Q. Zhang, Y. Wang, Size dependence in solvent-free aerobic oxidation of alcohols catalyzed by zeolite-supported palladium nanoparticles, Appl. Catal. A 334 (2008) 217–226. [11] G. Wu, X. Wang, N. Guan, L. Li, Palladium on graphene as efficient catalyst for solvent-free aerobic oxidation of aromatic alcohols: role of graphene support, Appl. Catal. B 136–137 (2013) 177–185.
C.-H. Liu et al. / Journal of Catalysis 350 (2017) 21–29 [12] M.L. Toebes, J.A. van Dillen, K.P. de Jong, Synthesis of supported palladium catalysts, J. Mol. Catal. A Chem. 173 (2001) 75–98. [13] Q. Wang, D.F. Shantz, Ordered mesoporous silica-based inorganic nanocomposites, J. Solid State Chem. 181 (2008) 1659–1669. [14] M. Opanasenko, P. Stepnicka, J. Cejka, Heterogeneous Pd catalysts supported on silica matrices, RSC Adv. 4 (2014) 65137–65162. [15] C.P. Mehnert, D.W. Weaver, J.Y. Ying, Heterogeneous Heck catalysis with palladium-grafted molecular sieves, J. Am. Chem. Soc. 120 (1998) 12289– 12296. [16] R.L. Oliveira, J.B.F. Hooijmans, P.E. de Jongh, R.J.M. Klein Gebbink, K.P. de Jong, Stabilization of palladium catalysts for the Heck reaction by support functionalization and solvent selection, ChemCatChem 6 (2014) 3223–3230. [17] C. Li, Q. Zhang, Y. Wang, H. Wan, Preparation, characterization and catalytic activity of palladium nanoparticles encapsulated in SBA-15, Catal. Lett. 120 (2008) 126–136. [18] C.M. Yang, H.A. Lin, B. Zibrowius, B. Spliethoff, F. Schüth, S.C. Liou, M.W. Chu, C. H. Chen, Selective surface functionalization and metal deposition in the micropores of mesoporous silica SBA-15, Chem. Mater. 19 (2007) 3205–3211. [19] C.J. Lin, S.H. Huang, N.C. Lai, C.M. Yang, Efficient room-temperature aqueousphase hydrogenation of phenol to cyclohexanone catalyzed by Pd nanoparticles supported on mesoporous MMT-1 silica with unevenly distributed functionalities, ACS Catal. 5 (2015) 4121–4129. [20] I. Yuranov, P. Moeckli, E. Suvorova, P. Buffat, L. Kiwi-Minsker, A. Renken, Pd/ SiO2 catalysts: synthesis of Pd nanoparticles with the controlled size in mesoporous silicas, J. Mol. Catal. A: Chem. 192 (2003) 239–251. [21] C.H. Liu, N.C. Lai, J.F. Lee, C.S. Chen, C.M. Yang, SBA-15-supported highly dispersed copper catalysts: vacuum–thermal preparation and catalytic studies in propylene partial oxidation to acrolein, J. Catal. 316 (2014) 231–239. [22] J.R.A. Sietsma, J.D. Meeldijk, J.P. den Breejen, M. Versluijs-Helder, A.J. Van Dillen, P.E. de Jongh, K.P. de Jong, The preparation of supported NiO and Co3O4 nanoparticles by the nitric oxide controlled thermal decomposition of nitrates, Angew. Chem. Int. Ed. 46 (2007) 4547–4549. [23] W. Zou, R.D. Gonzalez, The preparation of silica supported Pd catalysts: the effect of pretreatment variables on particle size, Catal. Lett. 12 (1992) 73–86. [24] I. Yuranov, L. Kiwi-Minsker, P. Buffat, A. Renken, Selective synthesis of Pd nanoparticles in complementary micropores of SBA-15, Chem. Mater. 16 (2004) 760–761.
29
[25] C.H. Liu, C.Y. Lin, J.L. Chen, N.C. Lai, C.M. Yang, J.M. Chen, K.T. Lu, Metal oxidecontaining SBA-15-supported gold catalysts for base-free aerobic homocoupling of phenylboronic acid in water, J. Catal. 336 (2016) 49–57. [26] C.H. Liu, N.C. Lai, S.C. Liou, M.W. Chu, C.H. Chen, C.M. Yang, Deposition and thermal transformation of metal oxides in mesoporous sba-15 silica with hydrophobic mesopores, Micropor. Mesopor. Mater. 179 (2013) 40–47. [27] R. Palkovits, C.M. Yang, S. Olejnik, F. Schüth, Active sites on SBA-15 in the Beckmann rearrangement of cyclohexanone oxime to e-caprolactam, J. Catal. 243 (2006) 93–98. [28] W.K. Józ´wiak, T.P. Maniecki, Influence of atmosphere kind on temperature programmed decomposition of noble metal chlorides, Thermochim. Acta 435 (2005) 151–161. [29] A. Savara, C.E. Chan-Thaw, I. Rossetti, A. Villa, L. Prati, Benzyl alcohol oxidation on carbon-supported Pd nanoparticles: elucidating the reaction mechanism, ChemCatChem 6 (2014) 3464–3473. [30] J.-D. Grunwaldt, M. Caravati, A. Baiker, Oxidic or metallic Palladium: which is the active phase in Pd-catalyzed aerobic alcohol oxidation?, J Phys. Chem. B 110 (2006) 25586–25589. [31] A.F. Lee, S.F.J. Hackett, J.S.J. Hargreaves, K. Wilson, On the active site in heterogeneous palladium selox catalysts, Green Chem. 8 (2006) 549–555. [32] A.F. Lee, J.N. Naughton, Z. Liu, K. Wilson, High-pressure XPS of crotyl alcohol selective oxidation over metallic and oxidized Pd(111), ACS Catal. 2 (2012) 2235–2241. [33] A.F. Lee, C.V. Ellis, J.N. Naughton, M.A. Newton, C.M.A. Parlett, K. Wilson, Reaction-driven surface restructuring and selectivity control in allylic alcohol catalytic aerobic oxidation over Pd, J. Am. Chem. Soc. 133 (2011) 5724–5727. [34] C.V. Gaskell, C.M.A. Parlett, M.A. Newton, K. Wilson, A.F. Lee, Redox-controlled crotyl alcohol selective oxidation: in situ oxidation and reduction dynamics of catalytic Pd nanoparticles via synchronous XANES/MS, ACS Catal. 2 (2012) 2242–2246. [35] C.M.A. Parlett, L.J. Durndell, A. Machado, G. Cibin, D.W. Bruce, N.S. Hondow, K. Wilson, A.F. Lee, Alumina-grafted SBA-15 as a high performance support for Pd-catalysed cinnamyl alcohol selective oxidation, Catal. Today 229 (2014) 46–55.