Abstract
Vinyl acetate monomer (VAM) is a crucial intermediate in the production of various polymers. While molecular Pd-acetate trimers and dimers, such as Pd3(OAc)6 and K2Pd2(OAc)6, are known to form on potassium acetate (KOAc)-promoted PdAu catalysts during heterogeneous VAM synthesis, their mechanistic role remains unclear. Here, we study the dynamics of different Pd-acetate species by utilizing in situ and operando crystallographic and spectroscopic characterizations combined with computational modeling on monometallic Pd model catalysts. The promoter-free catalyst expectedly shows low catalytic activity and VAM selectivity, corresponding to the complete reduction of Pdn(OAc)2n species to form Pd0 and PdCx nanoparticles. Conversely, noticeable quantities of KnPd2(OAc)n+4 species remain on the KOAc-promoted catalyst, leading to smaller nanoparticle formation with 10 times the activity and double the selectivity for VAM. This study reveals that molecular Pd-acetate trimers and dimers are significant indicators of catalytic performance and highlights their structurally dynamic nature in heterogeneous vinyl acetate chemistry.
Similar content being viewed by others
Introduction
Vinyl acetate monomer (VAM) is an important intermediate for polymers used for a variety of consumer goods (i.e., plastics, paints, and adhesives)1,2. VAM is commonly produced through the gas-phase reaction of ethylene with acetic acid (AcOH) (Eq. 1) over a PdAu alloy catalyst promoted by potassium acetate (KOAc)3,4,5,6,7. Although surface reaction mechanisms on the heterogeneous catalyst have been thoroughly investigated, the existence of molecular Pd-acetate trimers and dimers on the catalyst surface further complicates the catalyst system. Whereas Pd3(OAc)6 trimers have been associated with catalyst deactivation8,9,10,11,12, K2Pd2(OAc)6 dimers have been observed on catalysts with enhanced activity8,13. Despite these correlations, the specific role of these molecular Pd-acetate trimers and dimers in heterogeneous VAM chemistry remains unclear and understudied.
Studied over the past 60 years, density functional theory and spectroscopic studies on model Pd and PdAu catalysts determined that VAM is formed via coupling of acetate and ethylene on Pd active sites, followed by β-hydride elimination of an acetoxyethyl intermediate14,15,16,17,18,19,20,21. In addition, catalytic enhancements from Au in the PdAu alloys are attributed to both electronic22,23,24 and ensemble effects25,26,27. Alloying Pd with Au leads to weaker interactions between adsorbates and Pd due to shifts in the Pd d-band center from its Fermi level (Ef)23. The presence of Au also promotes the isolation of Pd surface atoms and formation of optimally spaced (~ 3 Å) Pd monomer pairs25. The addition of alkali promoters (e.g., KOAc) suppresses catalyst deactivation and enhances activity and selectivity to VAM8 by lowering the Pd d-band center due to electron transfer, altering the adsorption energies of adsorbates28,29.
However, the presence of physisorbed AcOH films on the catalyst under reaction conditions complicates the surface mechanism14,30,31,32. In liquid AcOH solutions and oxidative conditions, Pd can form trimeric Pd3(OAc)6 and dimeric Pd2(OAc)4 molecular species31,33. With the addition of KOAc, dimeric K2Pd2(OAc)6 species also exists in equilibrium34,35. Although Pd-acetate trimers and dimers have been observed on heterogeneous VAM catalysts using both ex situ9,10,36 and in situ8,11,13,37 characterizations, it is uncertain if they directly influence catalyst activity as catalytic complexes themselves. It is well-known that Pdn(OAc)2n salts are homogeneous catalysts in Suzuki-Miyaura coupling38,39,40, Mizoroki-Heck coupling41,42, Buchwald-Hartwig amination43,44,45, and C-H activation reactions46,47,48. Pdn(OAc)2n salts are also active for VAM synthesis via homogeneous acetoxylation49,50, suggesting molecular Pd-acetate trimers and dimers found on heterogeneous catalysts can generate VAM in parallel to VAM formation by the PdAu surface.
In this work, we study a simplified non-gold-containing Pd-acetate catalyst to isolate the contributions of Pd-acetate species in the VAM reaction. We synthesize model catalysts by physically mixing Pd3(OAc)6 and silica powder, then additionally mix them with KOAc. We evaluate catalytic performance through operando X-ray diffraction (XRD) while monitoring metal nanoparticle formation and restructuring verified by transmission electron microscopy (TEM), perform in situ diffuse reflectance Fourier-transform spectroscopy (DRIFTS) to observe Pd-acetate behavior under VAM conditions, and perform X-ray absorption spectroscopy (XAS) studies to assess catalyst structure and Pd oxidation state. Experimental results are complemented by density functional theory (DFT) computational studies. With fresh insights gained from these results, we uncover the reaction-induced dynamics of Pd-acetate species as well as their influence on activity and selectivity for VAM.
Results
Characterization of Fresh Catalysts
XRD profiles of fresh Pd-acetate catalysts with and without KOAc (hereby referred to as PdOAc/KOAc/SiO2 and PdOAc/SiO2, respectively) were collected to detect any Pd nanoparticles formed during synthesis. The absence of a Pd(111) peak near 40.2° for both catalysts suggests that Pd-acetate did not reduce as nanoparticles (Supplementary Fig. 1)51. Likewise, the X-ray absorption near edge structure (XANES) energies at the Pd K-edge for both catalysts were similar to palladium oxide (PdO) and palladium acetylacetonate (Pd(acac)2) references, indicating Pd was oxidized Pd2+ (Fig. 1a and Supplementary Table 1). However, the extended X-ray absorption fine structure (EXAFS) oscillations are phase-shifted compared to the PdO reference, suggesting that the Pd2+ species were not bulk PdO. The EXAFS of the catalysts shows one peak located below 2 Å in R-space, indicative of a Pd-O scattering path (Fig. 1b), whereas peaks near 3 Å characteristic of higher shell Pd-Pd scattering paths in PdO were absent for both the Pd(acac)2 reference and fresh catalysts, further indicating that Pd2+ species were organic Pd2+ species (i.e., Pd-acetate) rather than bulk PdO.
a, b Pd K edge XANES(a) and EXAFS (b) for the PdOAc/SiO2 (solid red) and PdOAc/KOAc/SiO2 (solid blue) fresh catalysts and the Pd foil (dashed black), PdO (dashed green), and Pd(acac)2 (dashed orange) reference materials. c, d Deconvoluted DRIFTS spectra in the 2000 – 1100 cm−1 acetate region of the fresh PdOAc/SiO2 (c) and PdOAc/KOAc/SiO2 (d) catalysts. A KOAc-free PdCl2/SiO2 sample was used as background for spectra in (c, d). Species present in (c, d) include K2Pd2(OAc)6 dimer (orange, 1640 cm−1), Pd3(OAc)6 trimer (purple, 1610 cm−1), KOAc (green, 1580 cm−1), Pd2(OAc)4 dimer (pink, 1570 cm−1), and bridged silyl acetate (gold, 1545 cm−1). Analysis conditions for experiments in (c, d): 1 atm; 25 °C; and 50 sccm; 100 vol% He.
DRIFTS was performed to detect Pd-acetate and other species on the fresh catalysts. The DRIFTS spectra in the acetate region (2000 to 1100 cm⁻¹; Fig. 1c, d) showed several distinct species identifiable by peaks in the asymmetric carboxylate stretching (νas(COO)) region (see Supplementary Tables 2 and 3 for a summary of ν(C = O) and νas(COO) peak locations for various acetic acid and acetate species). The KOAc-free PdOAc/SiO2 catalyst exhibited prominent νas(COO) vibrations at 1610 cm⁻¹ and 1570 cm⁻¹ (Fig. 1c), corresponding to Pd3(OAc)6 trimers and Pd2(OAc)4 dimers, respectively8,33,52. The absence of dimeric Pd2(OAc)4 in Pd3(OAc)6 powder, illustrated by a single νas(COO) vibration at 1608 cm⁻¹ in its attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectrum (see Supplementary Note 1 and Supplementary Fig. 2), suggests that mixing Pd3(OAc)6 powder with SiO2 induces the dissociation of Pd3(OAc)6 trimers into Pd2(OAc)4 dimers, likely due to interactions with residual water or surface hydroxyl groups on SiO253,54. A third peak at 1545 cm⁻¹ indicates SiO2-adsorbed bridging acetate55,56. The peaks at 1570 and 1545 cm⁻¹ together with a negative peak at 3750 cm⁻¹, characteristic of the hydroxyl stretching (ν(OH)) vibration of isolated silanols (Supplementary Fig. 3)57, indicated conversion of surface silanols to species like silyl acetates.
The addition of KOAc to the PdOAc/SiO2 catalyst revealed additional νas(COO) vibrations at 1640 and 1580 cm−1 (Fig. 1d). The peak at 1580 cm−1 was identified as ionic acetate (i.e., KOAc)11. In previous studies, we attributed the peak at 1640 cm−1 to the νas(COO) of dimeric K2Pd2(OAc)6 since it falls between ν(C = O) for terminal acetates of surface Pd-acetate (1680 cm−1) and νas(COO) for bridging acetates of Pd3(OAc)6 trimer (1610 cm−1), suggesting the species at 1640 cm−1 contains a mixture of terminal and bridging acetates13,36. Additional peaks on both catalysts located below 1450 cm−1 are attributed to symmetric carboxylate stretching (νs(COO)), CH3 deformation (δ(CH3)), and single-bond carboxyl stretching (ν(C-O)) vibrations, also characteristic of acetates11.
Since Kubelka-Munk units scale linearly with surface species concentration58, integrated peak intensities were used to quantify relative amounts of acetates with unknown extinction coefficients. Supplementary Fig. 4 shows the quantification of different Pd-acetate species based on the intensities of the νas(COO). Adding KOAc to the PdOAc/SiO2 catalyst caused the formation of K2Pd2(OAc)6 dimers and a decrease in both Pd3(OAc)6 and Pd2(OAc)4, indicating the conversion of these species to K2Pd2(OAc)6 (Eqs. 2 and 3). In addition, the relative decrease of Pd2(OAc)4 was greater than that of Pd3(OAc)6, suggesting Pd2(OAc)4 is easier to convert.
VAM Activity and selectivity
The catalytic performance for the VAM reaction was evaluated for each catalyst using operando XRD. Ethylene, AcOH and O2 conversions were less than 1, 2, and 5%, respectively. The VAM formation rate is minimal (40 μmol gPd−1 min−1 after 180 min time-on-stream; Fig. 2a) for the KOAc-free PdOAc/SiO2 catalyst. However, after KOAc was added, VAM formation increased over tenfold (410 μmol gPd−1 min−1). This is nearly double the activity for an unpromoted PdAu catalyst (PdAu/SiO2, 210 μmol gPd−1 min−1) and similar to a KOAc-promoted PdAu catalyst (PdAu/KOAc/SiO2, 420 μmol gPd−1 min−1), representative of conventional bimetallic VAM catalysts.
a, b VAM formation rate (a) and selectivity (b) during operando XRD analysis for each catalyst: PdOAc/SiO2 (closed red circles), PdOAc/KOAc/SiO2 (closed blue circles), PdAu/SiO2 (open red squares), and PdAu/KOAc/SiO2 (open blue squares). Reaction conditions during catalytic experiments in (a, b): 60 mg catalyst; 1 atm; 150 °C; 50 sccm; 15:5:2 vol% C2H4/AcOH/O2 (bal. He).
Similar trends for VAM selectivity were also observed. During the first 30 min time-on-stream, VAM selectivity was lower for all catalysts (Fig. 2b), corresponding to high CO2 formation (Supplementary Fig. 5) due to combustion side reactions. The addition of KOAc to the PdOAc/SiO2 catalyst increased VAM selectivity from 49% to 81% (Fig. 2b), similar to VAM selectivities obtained using conventional PdAu/SiO2 (75%) and PdAu/KOAc/SiO2 (78%) catalysts.
Reaction-induced formation of Pd nanoparticles
The contrasting behavior for VAM formation between different catalysts (Fig. 2) suggests variations in the catalyst structure during exposure to reaction conditions. Therefore, XRD profiles were simultaneously collected to follow the potential evolution of Pd nanoparticles during the reaction. After heating to 150 °C under the reaction mixture, XRD profiles revealed peaks in the (111) region between 38° and 40° after only 15 min for both PdOAc/SiO2 (Supplementary Fig. 6) and PdOAc/KOAc/SiO2 (Supplementary Fig. 7). Interestingly, the (111) peaks were shifted to lower 2θ values compared to the Pd(111) reference peak (40.2°). This shift, observed in PdAu catalysts due to lattice expansion from Au incorporation51,59, indicated a lattice expansion even in the absence of Au. We attribute this to the formation of Pd-carbide (PdCx, x = carbide fraction), previously observed by Goodman and co-workers on monometallic Pd catalysts after exposure to VAM reaction conditions60,61.
Deconvolution of the XRD profiles after 180 min revealed two distinct phases for both catalysts: one pure Pd, and one PdCx (Fig. 3a, b and Supplementary Fig. 8). The carbide content in the PdCx phase is related to the lattice parameter and hence the 2θ shift. Considering lattice parameters of PdCx phases with known carbide fractions determined by Chan et al.62, carbide content was estimated for the PdCx phases observed via operando XRD (see Supplementary Note 2). Carbide content of the PdCx phase for the PdOAc/SiO2 catalyst was slightly greater (PdC0.14; Table 1) compared to the PdOAc/KOAc/SiO2 catalyst (PdC0.12; Table 1). In addition, the relative XRD area of the PdCx phase for PdOAc/SiO2 was greater (69%; Supplementary Table 4) than PdOAc/KOAc/SiO2 (51%; Supplementary Table 5). These differences are also qualitatively observed in the XRD profiles, where the PdCx peak is larger for the KOAc-free catalyst (Fig. 3a, b).
a, b Operando XRD profiles of the PdOAc/SiO2 (a) and PdOAc/KOAc/SiO2(b) catalysts after 180 min time-on-stream. The blue line represents the Pd(111) reference, the blue curve represents the Pd phase, and the purple curve represents the PdCx phase in (a, b). Analysis conditions for experiments in (a, b): 1 atm; 150 °C; 50 sccm; 15:5:2 vol% C2H4/AcOH/O2 (bal. He). c, d Pd K edge XANES (c) and EXAFS (d) for the PdOAc/SiO2 (solid red) and PdOAc/KOAc/SiO2 (solid blue) catalysts collected after exposure to reaction conditions and the Pd foil (dashed black), PdO (dashed green), and Pd(acac)2 (dashed orange) reference materials.
XAS spectra were then collected for the catalysts after XRD experiment completion to obtain spectroscopic evidence for Pd and PdCx nanoparticle formation. The XANES energies for both catalysts were similar to the Pd foil, indicating the formation of a Pd0 phase (Fig. 3c and Supplementary Table 6). However, the EXAFS oscillations are phase-shifted, suggesting changes to Pd-Pd bond distance compared to Pd foil (Fig. 3c). This observation is likewise reflected in the EXAFS R-space plots, which show a slight shift in the Pd-Pd peaks to higher R values compared to Pd foil (Fig. 3d). EXAFS fittings confirmed the elongation of the first shell Pd-Pd bond (RPd-Pd) from 2.73 Å for Pd foil to 2.77 and 2.78 Å for the PdOAc/SiO2 and PdOAc/KOAc/SiO2 catalysts, respectively (Supplementary Table 6), consistent with the XRD-observed PdCx phases. To obtain further evidence of the PdCx phase, in situ XAS spectra of the catalysts were collected while under H2 gas. Exposure to H2 removes any carbide present on Pd nanoparticles, which decreases the Pd-Pd bond distance63,64. After H2 treatment at 200 °C, the Pd-Pd peaks of both catalysts shifted to lower R values in both the real and imaginary components (Supplementary Fig. 9). Likewise, RPd-Pd decreased to 2.73 Å for both catalysts, identical to the Pd foil, indicating the transition of the PdCx phase into a Pd phase (Supplementary Table 6).
The lattice expansion and bond elongation of Pd nanoparticles detected by XRD and XAS are similarly observed by high-resolution transmission electron microscopy (HR-TEM) imaging. Lattice spacings (d(111)) for the Pd(111) phase were found to be between 0.225 nm and 0.231 nm across multiple nanoparticles for both catalysts (Fig. 4a, b). This range of lattice spacings is consistent with a mixture of both Pd (d(111) = 0.224 nm, 2θ(111) = 40.2°) and PdCx (d(111) = 0.231 nm, 2θ(111) = 39.0° for PdC0.15) phases, corroborating the previously discussed crystallographic and spectroscopic results. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images and their corresponding energy-dispersive X-ray spectroscopy (EDS) maps reveal overlapping regions of bright contrast in the Pd and C elemental maps for both catalysts (Supplementary Figs. 10 and 11), indicating carbon atoms are distributed throughout Pd nanoparticles, forming large PdCx phases.
a, b HR-TEM images of PdOAc/SiO2 (a) and PdOAc/KOAc/SiO2 (b) catalysts after exposure to reaction conditions. Pd(111) lattice spacings are presented for each image in (a, b). Scale bars represent 2 nm for all images in (a, b). c, d HAADF-STEM images and their corresponding particle size distributions for PdOAc/SiO2 (c) and PdOAc/KOAc/SiO2 (d) catalysts after exposure to reaction conditions. Scale bars represent 20 nm for all images in (c, d). Particle size distributions in (c, d) were generated by measuring at least 300 particles. PdOAc/SiO2 and PdOAc/KOAc/KOAc catalysts are identified by red and blue, respectively, in (a–d).
The grain sizes of the Pd and PdCx phases on the PdOAc/SiO2 catalyst were 4.2 and 6.7 nm, respectively (Table 1). However, the PdOAc/KOAc/SiO2 catalyst contained smaller Pd (2.7 nm) and PdCx (5.2 nm) phases (Table 1). These differences are visually apparent in the XRD profiles as the peaks for the KOAc-free catalyst are sharper, indicating larger grain sizes (Fig. 3b). This suggests that the absence of KOAc on PdOAc/SiO2 leads to an increased concentration of larger PdCx particles. In addition, the large PdCx grain sizes combined with carbide fractions near saturation (i.e., x = 0.15) support the Pd and C elemental STEM-EDS maps showing interstitial carbon atoms are randomly distributed throughout nanoparticles rather than concentrated at the surface layers.
HAADF-STEM images reveal wide particle size distributions (1–10 nm) for both catalysts after exposure to reaction conditions with average particle sizes of 3.8 and 3.0 nm for the PdOAc/SiO2 and PdOAc/KOAc/SiO2 catalysts, respectively (Fig. 4c, d and Table 1), supporting trends in grain sizes obtained from operando XRD. Smaller HAADF-STEM average particle sizes compared to XRD grain sizes are attributed to XRD detection limitations for particles less than ~ 3 nm. Since previous studies have shown that first shell EXAFS M-M coordination numbers (NM-M) of Pt, Au, and Pd catalysts strongly correlate with nanoparticle dispersion and size65,66, a higher Pd-Pd coordination number for the PdOAc/SiO2 catalyst (NPd-Pd = 8.5; Table 1) indicates a larger average nanoparticle size compared the PdOAc/SiO2 catalyst (NPd-Pd = 6.2; Table 1), supporting trends found in XRD and STEM particle sizes. Supplementary Note 3 provides further details to estimate dispersion and particle size using EXAFS, with results for the PdOAc/SiO2 and PdOAc/KOAc/SiO2 catalysts summarized in Supplementary Table 7.
EXAFS of the PdOAc/KOAc/SiO2 catalyst after exposure to reaction conditions contained Pd-O scattering in addition to Pd-Pd scattering (Supplementary Table 6), with a Pd-O peak appearing below 2 Å in R-space (Fig. 3d). Interestingly, Pd-O scattering was not observed for the PdOAc/SiO2 catalyst (Fig. 3d and Supplementary Table 6), indicating that the Pd-acetate trimers and dimers were completely converted to either Pd or PdCx. On the other hand, approximately 27% of the Pd atoms on the PdOAc/KOAc/SiO2 catalyst (Table 1) remained Pd-acetate trimers and dimers as determined from the Pd-O coordination number (NPd-O) as a fraction of total possible Pd-O coordination in a bulk oxide structure (NPd-O = 4).
Dynamics of Pd-acetate species during the VAM reaction
The detection of Pd and PdCx nanoparticle phases suggests alterations to the Pd-acetate species during the VAM reaction. Therefore, in situ DRIFTS spectra were collected to monitor the dynamics of Pd-acetate species under VAM reaction conditions. In situ DRIFTS spectra revealed substantial changes in the acetate region for both catalysts within the first 30 min (Supplementary Figs. 12 and 13), suggesting changes in Pd-acetate species. After 180 min, the KOAc-free catalyst had neither trimeric Pd3(OAc)6 (1610 cm−1) nor dimeric Pd2(OAc)4 (1570 cm−1) species remaining (Fig. 5a). In contrast, numerous acetate species were present on the PdOAc/KOAc/SiO2 catalyst (Fig. 5b). These include the KOAc (1580 cm−1) and bridged silyl acetate (1540 cm−1) peaks originally present on the fresh catalyst. Additionally, a peak at 1715 cm−1 appeared, which we previously attributed to ν(C = O) of dimeric KH(OAc)2 from hydrogen bonding between AcOH and KOAc8,13,36,67. Surface-adsorbed acetate species were also detected at 1745 and 1680 cm-1. The peak at 1745 cm−1 is characteristic of ν(C = O) for AcOH adsorbed on SiO2 through hydrogen bonding with surface hydroxyls52,56. This is the only species observed on the PdOAc/SiO2 catalyst under reaction conditions (Fig. 5a). The peak at 1680 cm−1 was identified as ν(C = O) of acetate adsorbed on surface Pd in a terminal monodentate mode13,36,68.
a, b Deconvoluted in situ DRIFTS spectra in the 2000 – 1100 cm−1 acetate region of the PdOAc/SiO2 (a) and PdOAc/KOAc/SiO2 (b) catalysts after 180 min time-on-stream. A KOAc-free PdCl2/SiO2 sample was used as background for spectra in (a, b). c, d Normalized integrated DRIFTS intensities of Pd-acetate trimers and dimers for the PdOAc/SiO2 (c) and PdOAc/KOAc/SiO2 (d) catalysts during in situ experiments with exposure to reaction conditions. All intensities in (c, d) are normalized by the integrated DRIFTS intensities of the respective species on the fresh catalysts (values shown in Supplementary Fig. 4). Species present in (a–d) include monomeric AcOH gas (light gray, 1800 and 1770 cm−1), AcOH hydrogen-bound to SiO2 surface hydroxyls (blue, 1745 cm−1), dimeric AcOH gas (dark gray, 1730 cm−1), KH(OAc)2 dimer (light blue, 1715 cm−1), acetate adsorbed on surface Pd in terminal monodentate mode (dark red, 1680 cm−1), KnPd2(OAc)n+2 dimer (orange, 1640 – 1620 cm−1), Pd3(OAc)6 trimer (purple, 1610 cm−1), KOAc (green, 1580 cm−1), Pd2(OAc)4 dimer (pink, 1570 cm−1), and bridged silyl acetate (gold, 1540 cm−1). Analysis conditions for experiments in (a–d): 1 atm; 150 °C; 50 sccm; 15:5:2 vol% C2H4/AcOH/O2 (bal. He).
While absent on PdOAc/SiO2 (Fig. 5a), small amounts of trimeric Pd3(OAc)6 (1610 cm-1) and dimeric Pd2(OAc)4 (1570 cm−1) species remained on PdOAc/KOAc/SiO2 (Fig. 5b). A peak at 1620 cm−1 also appeared on the PdOAc/KOAc/SiO2 catalyst after 180 min (Fig. 5b). Although similar to νas(COO) for K2Pd2(OAc)6 dimer (1640 cm−1), the red-shift suggests a greater amount of bridging versus terminal acetates. We therefore attributed this peak to the νas(COO) vibration of KPd2(OAc)5. Interestingly, the wavenumber for dimeric KnPd2(OAc)n+4 strongly correlated to the integrated DRIFTS intensity for the KOAc peak at 1580 cm−1. As the reaction proceeds, the concentration of KOAc decreases, leading to a significant red shift of the νas(COO) vibration for dimeric KnPd2(OAc)n+4 species (Supplementary Fig. 14). Equilibrium reactions converting Pd3(OAc)6 trimers and Pd2(OAc)4 dimers to KPd2(OAc)5 dimers (Eqs. 4 and 5) require half the number of KOAc molecules compared to the equilibrium reactions for conversion to K2Pd2(OAc)6 dimers (Eqs. 2 and 3). Therefore, the formation of KPd2(OAc)5 species is favored as KOAc concentration decreases.
Peaks associated with ν(C = O) vibrations of gas-phase AcOH as either monomers (1800 and 1770 cm−1) or dimers (1730 cm−1) were also observed on both catalysts (Fig. 5a, b)11,69. Peaks below 1450 cm−1 were attributed to the νs(COO), δ(CH3), and ν(C-O) vibrations characteristic of AcOH and acetates11,69. Although a νs(COO) vibration for acetate adsorbed on surface Pd in a bidentate mode has previously been identified at 1414 cm−1 by Stacchiola et al.70, we excluded this species from our analysis due to interference with νs(COO) vibrations of other acetates. Additional sharp peaks (1890 and 1440 cm−1) observed for both catalysts are attributed to the ν7 + ν8 and v12 vibrations, respectively, of gas-phase ethylene (Fig. 5a, b)11,71.
Figure 5c, d highlights the dynamics of the Pd-acetate species in the absence of or presence of KOAc as the VAM reaction proceeds. There is an initial decrease in intensities for all species after heating to 150 °C under He atmosphere compared to the fresh catalysts (indicated by Normalized Intensity < 1 at 0 min in Fig. 5c, d), suggesting thermal decomposition occurs. After only 60 min, Pd-acetate species are entirely absent on PdOAc/SiO2 (Fig. 5c). This observation is consistent with the EXAFS indicating no Pd-O scattering (Supplementary Table 6).
Conversely, significant quantities of Pd-acetate trimers and dimers remained on PdOAc/KOAc/SiO2 (Fig. 5d), consistent with the EXAFS estimation that 27% of Pd atoms are organic Pd2+ (Table 1 and Supplementary Table 6). During the first 15 min time-on-stream, all Pd-acetate species decreased significantly (Fig. 5d) while Pd nanoparticles formed (as observed via operando XRD Supplementary Fig. 7). However, once the Pd nanoparticles stabilized (~ 30 min), DRIFTS intensities of dimeric Pd2(OAc)4 and KnPd2(OAc)n+4 began to increase whereas the intensities for Pd3(OAc)6 trimer continued to decrease (Fig. 5d), although more slowly compared to the KOAc-free catalyst (Fig. 5c). This suggests continued conversion of Pd3(OAc)6 trimers to Pd2(OAc)4 dimers, likely due to interactions with byproduct water formed from the VAM reaction (Eq. 1), and conversion to KnPd2(OAc)n+4 via reactions with KOAc (Eqs. 2 and 4). As the Pd3(OAc)6 trimer becomes nearly depleted after 120 min, the Pd2(OAc)4 dimer concentration begins to decrease while the KnPd2(OAc)n+4 dimer concentration remains relatively constant (Fig. 5d), indicating conversion of Pd2(OAc)4 dimers via reactions with KOAc (Eqs. 3 and 5).
Mechanisms Involving Pd-acetate Trimers and Dimers on Heterogeneous VAM Catalysts
The presence of Pd3(OAc)6 trimers on heterogeneous Pd and PdAu VAM catalysts has previously been associated with catalyst deactivation8,9,10,11,12. Following the addition of KOAc, heterogeneous catalysts become more active and selective, which Lercher and co-workers attributed to the formation of K2Pd2(OAc)6 dimers8. Similar behavior was observed with our model catalysts. The KOAc-free PdOAc/SiO2 catalyst consisting of Pd3(OAc)6 trimers and Pd2(OAc)4 dimers is less active and selective than the corresponding PdAu/SiO2 catalyst (Fig. 2). On the other hand, the KOAc-impregnated PdOAc/KOAc/SiO2 catalyst contained significant quantities of K2Pd2(OAc)6 dimers and had nearly identical activity as the PdAu/KOAc/SiO2 catalyst (Fig. 2).
Considering the combined XRD, XAS, TEM and DRIFTS results of our model Pd-acetate catalysts, we propose that Pd3(OAc)6 trimers and Pd2(OAc)4 dimers are not necessarily themselves catalytically less active and selective for VAM. Instead, we propose that the rapid reduction of these species under reaction conditions forming monometallic Pd and PdCx phases as either separate nanoparticles or as a Pd surface layer on PdAu nanoparticles ultimately leads to catalyst deactivation (Fig. 6a). Previously, we showed that highly clustered Pd surface atoms correlated to lower VAM activity and selectively13. Similarly, Pd/PdCx nanoparticles and/or surface layers formed from the reduction of Pdn(OAc)2n could be the cause for the observed deactivation.
a Schematic showing the complete reduction of Pdn(OAc)2n species on KOAc-free heterogeneous VAM catalysts under reaction conditions, forming large monometallic Pd and PdCx nanoparticles, which are minimally active and selective for VAM. b Schematic showing the partial reduction of KnPd2(OAc)n+4 species on KOAc-promoted heterogeneous VAM catalysts under reaction conditions, forming small monometallic Pd and PdCx nanoparticles which are highly active and selective for VAM. Pd, K, and C atoms in (a and b) are represented by black, purple, and red spheres, respectively. c Proposed redox cycle involving molecular Pd-acetate species on heterogeneous VAM catalysts.
The behavior of Pd3(OAc)6 trimers and Pd2(OAc)4 dimers on heterogeneous VAM catalysts is analogous to mechanisms observed in liquid-phase homogeneous VAM chemistry since the 1960s. After reacting with ethylene to form VAM, Pd2+ homogeneous catalysts (e.g., Pdn(OAc)2n, PdCl2) reduce to form Pd nanoparticles, and require the use of co-catalysts (e.g., CuCl2, benzoquinone) to re-oxidize Pd0 and restore activity49,50,72,73. Kragten et al. reported that K2Pd2(OAc)6 dimers formed with KOAc in solution are significantly more active and selective than Pd3(OAc)6 trimers for liquid-phase homogeneous VAM reaction31.
In the case of the heterogeneous catalyst, our studies show that the enhanced activity associated with the presence of K2Pd2(OAc)6 dimers is due to the formation of smaller Pd nanoparticles (< 3 nm) compared to those formed from trimeric Pd3(OAc)6 and dimeric Pd2(OAc)4 species (> 3 nm) under reaction conditions (Fig. 6b). We previously showed that KOAc impregnation of silica-supported PdAu catalysts in either aqueous or AcOH solutions leads to the formation of Pd-acetate species such as K2Pd2(OAc)6 dimers, whereas only trace amounts of Pd-acetate species are formed when exposed to AcOH solutions without KOAc36. Therefore, we propose that KOAc assists to re-oxidize Pd and PdCx nanoparticles formed during the reduction of Pdn(OAc)2n trimers and dimers (Fig. 6c), generating KnPd2(OAc)n+4 species which are then reduced to Pd and PdCx nanoparticles, completing the redox cycle (Fig. 6c). As a result, this redox cycle minimizes the formation of highly clustered Pd found in large Pd and PdCx nanoparticles or monometallic Pd surface layers on PdAu nanoparticles, thus preserving high catalytic activity and selectivity to VAM (see Supplementary Note 4 for discussion on the potential effects of the PdCx phase on catalyst performance).
Density Functional Theory (DFT) calculations were then performed to investigate the thermodynamics of the different steps in the catalytic system and interrogate the plausibility of our proposed redox cycle. First, the role of KOAc in assisting the re-oxidation of Pd and PdCx nanoparticles formed during the reduction of Pdn(OAc)2n trimers and dimers to form KnPd2(OAc)n+4 species was examined by considering the thermodynamics of the Pd-acetate transformations shown in Eqs. 6–8. In the presence of KOAc, bulk Pd is oxidized to the dimeric species [Pd2(OAc)6]2− (Eq. 6), while in the absence of KOAc, Pd forms the neutral dimer Pd2(OAc)4 (Eq. 7). Combining these reactions yields the overall net global reaction shown in Eq. 8, which reveals the preference for bulk Pd dimers to form as [Pd2(OAc)6]2− in the presence of KOAc.
The calculated standard Gibbs free energy (ΔG0(g)) (see Supplementary Note 5 and Supplementary Tables 8 and 9 for computational details) for the net reaction in Eq. 8 is near zero, indicating that under standard-state conditions (i.e., equal concentrations of all species) the reaction is essentially at equilibrium, favoring neither the forward nor reverse reactions (Supplementary Table 9). However, when corrected using the relative ratios of Pd2(OAc)4, (OAc)⁻, and [Pd2(OAc)6]2− obtained from DRIFTS data (Supplementary Table 10), the reaction becomes favorable (ΔG = − 3.06 and − 6.68 kcal mol−1 at the TPSSh-D3 and r2SCANh-D3 levels, respectively), demonstrating that KOAc stabilizes the [Pd2(OAc)6]2− dimer and thereby promotes reoxidation of bulk Pd during VAM formation.
We also evaluated the thermodynamic driving force for VAM formation from Pd-acetate trimers and dimers by calculating reaction energies for VAM formation according to the pathways presented in Fig. 6c, namely the reactions catalyzed by the K2Pd2(OAc)6 dimer, Pd3(OAc)6 trimer, and Pd2(OAc)4 dimer. The calculated ΔG0(g) values for these transformations were all in the range of – 59.55 to – 64.50 kcal mol−1 (Supplementary Table 11), confirming that all Pd-acetate species can readily catalyze VAM formation (see Supplementary Note 6 for computational details). Although K2Pd2(OAc)6 does not show an intrinsic energetic advantage, the actual driving force depends on the relative concentrations of surface and gaseous species. Nevertheless, the similar Gibbs free energies across all pathways to VAM further suggest that KnPd2(OAc)n+4 dimers are not inherently more active than Pdn(OAc)2n trimers and dimers. Instead, we argue that the various Pd-acetate complexes are more precisely described as indicator species of catalyst activity due to their association with Pd nanoparticle size and cluster formation.
Discussion
In this work, we prepared two heterogeneous VAM catalysts (PdOAc/SiO2 and PdOAc/KOAc/SiO2) by physically mixing Pd3(OAc)6, KOAc, and silica powders. Trimeric Pd3(OAc)6 and dimeric Pd2(OAc)4 species were initially present on the KOAc-free catalyst, whereas K2Pd2(OAc)6 dimer formed after KOAc addition. Although operando XRD studies showed Pd and PdCx phases for both catalysts under reaction conditions, the grain and particle sizes for the KOAc-free catalyst were larger than those on the KOAc-impregnated catalyst. Consequently, the PdOAc/SiO2 catalyst was essentially inactive, whereas the PdOAc/KOAc/SiO2 catalyst was nearly as active as a conventional KOAc-promoted PdAu/SiO2 catalyst. In situ DRIFTS studies confirmed the complete reduction of Pdn(OAc)2n species under reaction conditions, but dimeric KnPd2(OAc)n+4 species remained on the KOAc-impregnated catalyst. From the crystallographic and spectroscopic results, we propose that Pdn(OAc)2n and KnPd2(OAc)n+4 species actively participate in a redox cycle involving the formation and re-oxidation of Pd and PdCx nanoparticles, supported computationally by DFT studies of the thermodynamics for each step. Overall, this study of silica-supported Pd-acetate model catalysts for VAM synthesis highlights the dynamic nature of Pd-acetate trimers and dimers to form metallic Pd and PdCx phases, which are further indicative of catalyst performance.
Methods
Materials
Trimeric palladium acetate (Pd3(OAc)6, ≥ 98%), sodium tetrachloropalladate (Na2PdCl4, ≥98%), gold chloride trihydrate (HAuCl4 ∙ 3H2O, ≥ 99.9%), silica gel (SiO2, 40–63 μm particle size, 480 m2 g−1 surface area), sodium carbonate (Na2CO3, ≥ 99%), and potassium acetate (KOAc, ≥ 99%) were purchased from Sigma-Aldrich. Glacial acetic acid (AcOH, ≥ 99.7% w/w) was purchased from Fisher Scientific. Hydrogen (H2, 99.999 vol%), helium (He, 99.999 vol%), ethylene (C2H4, 99.999 vol%), and 10 vol% oxygen (10 vol% O2/bal. He) gases were supplied by Airgas. All experiments were conducted using deionized (DI) water ( ≥ 18 MΩ cm, Thermo Scientific MicroPure).
Catalyst synthesis
Pd-acetate catalysts were synthesized using a simple physical mixing approach rather than typical incipient wetness impregnation methods. Since trimeric Pd3(OAc)6 is known to dissociate into dimeric Pd2(OAc)4 and monomeric Pd(OAc)2 species when exposed to water and other solvents33,53,54,74, we implement this solution-free approach to preserve the trimeric Pd3(OAc)6 structure. After vacuum drying at 25 °C, SiO2 powder and Pd3(OAc)6 salt were physically mixed using a mortar and pestle, achieving a metal loading of 6 wt% Pd. This sample (PdOAc/SiO2) was then loaded with KOAc by an additional physical mixing step. This KOAc-impregnated sample (PdOAc/KOAc/SiO2) contained metal loadings of 6 wt% Pd and 3 wt% K.
For comparison, conventional PdAu catalysts were prepared by methods previously reported in the literature13,36,75. Na2PdCl4 and HAuCl4 metal precursors were added to SiO2 powder by incipient wetness impregnation. After drying overnight at 80 °C, 0.4 M Na2CO3 solution was added to the sample (25 mL per 5 g of sample). This was followed by another drying step at 80 °C overnight. The sample was repeatedly washed and vacuum-filtered with DI water (1 L per 5 g of sample) and dried once more overnight at 80 °C. The dried sample was reduced under flowing H2 at 150 °C for 1 hr. The reduced sample (PdAu/SiO2) was loaded with KOAc by incipient wetness impregnation. The KOAc-treated sample (PdAu/KOAc/SiO2) was then dried overnight at 80 °C. Target metal loadings of 8 wt% Pd, 4 wt% Au, and 3 wt% K were confirmed by ICP-OES. Detailed characterizations (e.g., XRD, XPS, XAS, DRIFTS, and CO chemisorption) of these catalysts have previously been reported by our group13,36.
Determination of activity and selectivity to VAM
Catalytic performance was evaluated for the VAM reaction by directing the effluent stream from the XRD reaction chamber to an online gas chromatograph (Agilent model 7890B) equipped with both a thermal conductivity detector (TCD) and a flame ionization detector (FID). A gas sampling valve (Valco Instruments) was used to inject 250 μL of gas sample into the GC inlet at a split ratio of 10:1 with He as the carrier gas at 1.8 mL min−1 constant flow. The injected sample was separated using an Agilent PoraPLOT Q-HT (25 m × 320 μm × 10 μm) capillary column. A splitter at the end of the column directed the separated sample to the TCD and FID in parallel. The oven temperature was initially set to 60 °C for 7 min, followed by a 30 °C min−1 ramp to 180 °C. XRD profiles of the catalyst sample in the reaction chamber were simultaneously collected during the reaction, allowing for operando XRD analysis.
VAM and CO2 products were quantified by converting integrated intensities of their respective GC signals to partial pressures using prepared calibration curves. Partial pressures were then normalized by total Pd loading and reported as formation rates (μmol gPd−1 min−1). Although selectivity values for the VAM reaction are commonly reported based on ethylene conversion (see Supplementary Note 7)14,76,77, we instead define selectivity to VAM (SVAM) based on total carbon conversion according to the following equation:
where pVAM and pCO2 are the partial pressures of VAM and CO2, respectively. VAM and CO2 were assumed to be the only detectable carbon-containing products.
XRD Characterization
Operando XRD profiles were collected using a Philips X’Pert Pro diffractometer equipped with a Cu Kα radiation source (1.540598 Å) generated at 45 kV and 40 mA. XRD analysis was conducted in a 2θ range of 30 – 70° with a step size of 0.0050° s−1. Operando gas-phase studies were performed by loading 60 mg of the catalyst sample into an XRK 900 (Anton-Paar) high-temperature reaction chamber. An initial XRD profile of the fresh catalyst was collected under a pure He atmosphere at 25 °C. After heating the sample to 150 °C at a rate of 10 °C min−1 under He flow, the VAM reaction was then initiated at 150 °C by introducing a mixture of 15 vol% C2H4, 5 vol% AcOH, and 2 vol% O2 (bal. He) to the reaction chamber. Additional XRD profiles were continuously collected at 150 °C. All flowrates were maintained at 50 sccm, and the total pressure maintained at 1 atm. Gas-phase AcOH was introduced by bubbling He through a temperature-controlled AcOH saturator bath. Gas lines were heated to > 110 °C to prevent condensation of volatile species.
The XRD (111) peaks for Pd phases were analyzed using OriginLab by deconvoluting the (111) planes with a Voigt function. The obtained XRD maximum peak intensity and FWHM were then used to calculate the grain size by applying Scherrer’s Equation78:
where τ is the mean grain size, K is the proportionality constant, λ is the X-ray wavelength, β is the peak full width at half maximum (FWHM) in radians, and θ is the Bragg angle in radians.
XAS Characterization
XAS experiments were performed at the 8-ID Inner Shell Spectroscopy (ISS) beamline79 at the National Synchrotron Light Source II (NSLS-II), Brookhaven National Laboratory (BNL), at the Pd K edge (24.350 keV) in fluorescence mode using a Si(111) double-crystal monochromator and a passivated implanted planar silicon (PIPS) detector. During ex situ experiments, fresh and used (i.e., collected from the XRD reaction cell after operando experiments) catalyst samples were ground into fine powders, pressed as wafers, secured by X-ray transparent Kapton tape on the sample stage, and scanned. In situ experiments were additionally performed by loading the ground sample powder into a quartz in situ cell and tightened with Swagelok fittings, allowing for control of gas atmosphere and temperature. After exposure to pure H2 gas flow at 200 °C for 15 min, the samples were cooled to 25 °C and scanned. All scans were accompanied by simultaneous acquisition of a Pd foil absorption spectrum obtained via a third ion chamber used for energy calibration. Scans of palladium oxide (PdO) and palladium acetylacetonate (Pd(acac)2) were also obtained as references for Pd2+ species.
XAS data was fit using winXAS 3.1 software80. XANES was used to determine the edge energy of each sample at the Pd K edge. Least-squares regression fits of the k2-weighted Fourier transform data from 2.8 to 10.9 Å−1 in k-space were used to obtain the EXAFS coordination parameters. The first shell was used to fit all spectra. All samples were fit using theoretical EXAFS scattering paths calculated using FEFF 6.0 software81. An S02 value of 0.80 at a fixed coordination number of 12 and bond distance of 2.73 Å was determined using the Pd foil.
DRIFTS Characterization
In situ DRIFTS experiments were performed on a Thermo Scientific Nicolet iS10 FTIR spectrometer equipped with a liquid-N2-cooled HgCdTe (MCT) detector. A high-temperature reaction chamber equipped with ZnSe windows, mounted on a Praying Mantis diffuse reflectance accessory (Harrick Scientific), was used for in situ gas-phase analysis of catalyst samples. In situ DRIFTS studies were performed using identical procedures and conditions as the operando XRD studies described previously. Background spectra were collected from a physically mixed PdCl2/SiO2 sample, synthesized using similar methods as the Pd-acetate catalysts, since it had similar IR throughput as the Pd-acetate catalysts without interference from other species in the acetate region under He atmosphere at the respective temperatures (25 or 150 °C). All spectra are reported in Kubelka-Munk units by averaging 128 scans at 4 cm−1 resolution. Peak locations and integrated intensities were determined for each vibrational mode by fitting with a Voigt function in OriginLab.
TEM Characterization
Catalyst powders were collected from either the XRD or DRIFTS in situ cells after exposure to reaction conditions for TEM analysis. HR-TEM images were acquired using a Titan (Thermo Fischer Scientific) transmission electron microscope operating at 300 kV. Lattice spacings were measured using ImageJ. HAADF-STEM images were additionally acquired with a JEM-ARM200F (JEOL) transmission electron microscope operating at 200 kV. The particle size distribution was calculated by counting and measuring at least 300 particles using ImageJ.
Computational methods
DFT calculations were carried out as implemented in ORCA v6.0.082,83,84 to evaluate the thermodynamics of key steps in the catalytic system, including KOAc-mediated reoxidation of Pd nanoparticles to KnPd2(OAc)n+4 species and their subsequent reduction, as well as the formation of vinyl acetate monomer (VAM) catalyzed by Pdn(OAc)2n and KnPd2(OAc)n+4 clusters. The Perdew-Burke-Ernzerhof hybrid functional PBE085, combined with the def2-SV(p) basis set86 and the auxiliary basis set def2/J87, was employed for geometry optimizations. The convergence strategy and criteria were configured to SlowConv and TightSCF, respectively. A defgrid2-level grid was used for numerical integration. Dispersion effects were considered using Grimme’s D3 method with the Becke-Johnson damping scheme (D3BJ)88. To accelerate DFT calculations, the RI-J approximation for the Coulomb integrals and the COSX numerical chain-of-sphere integration for the HF exchange integrals (RIJCOSX) were used89.
Analytical frequency calculations on the optimized equilibrium structures were carried out at the same level of theory and setup as geometry optimizations (PBE0-D3/def2-SV(p)). These calculations confirmed that the optimized equilibrium structures were in the minima of the potential energy surface by the absence of imaginary frequency modes. We also obtained zero-point energy (ZPE), thermal, and entropic corrections at 298.15 K (G0 →298K) for each species using the ideal gas rigid rotor-harmonic oscillator approximations.
The gas-phase electronic energies of the species were refined via subsequent single-point calculations using the hybrid meta-Generalized Gradient Approximation (meta-GGA) Tao–Perdew–Staroverov–Scuseria hybrid (TPSSh) functional90,91, as well as the range-separated hybrid r2SCANh functional92. These calculations were carried out using the def2-TZVP basis set86 combined with the def2/J auxiliary basis set87, with all other parameters set up as described for geometry and frequency calculations. Additional computational details are outlined in Supplementary Notes 5 and 6.
Data availability
The authors declare that the data supporting the findings of this study are available within the paper and its supplementary information files. Should any raw data files be needed in another format, they are available from the corresponding author upon request. Source data are provided in this paper. Input and output files for all DFT calculations are available in the following online repository: https://github.com/tsenftle/VAM.
References
Ben Halima, N. Poly(vinyl alcohol): review of its promising applications and insights into biodegradation. RSC Adv. 6, 39823–39832 (2016).
Speetjens, F. W. & Mahanthappa, M. K. Synthesis and rheological characterization of poly(vinyl acetate- b -vinyl alcohol- b -vinyl acetate) triblock copolymer hydrogels. Macromolecules 48, 5412–5422 (2015).
Ertl, G., Knozinger, H. & Weitkamp, J. Handbook of Heterogeneous Catalysis. (Wiley-VCH, Weinheim, Germany, 1997).
Couves, J. W. & Kitchen, S. J. Process for the production of vinyl acetate. US6492299B1 (2002).
Wang, T. Vinyl acetate catalyst prepared with potassium aurate and comprising metallic palladium and gold on a carrier precoated with copper. US6017847A (2000).
Wang, T. & Broussard, J. A. Heterogeneous bimetallic palladium-gold catalyst for vinyl acetate production. US5700753A (1997).
Mayer, R. et al. Catalyst. US20090093653A1 (2009).
Hanrieder, E. K., Jentys, A. & Lercher, J. A. Atomistic engineering of catalyst precursors: dynamic reordering of PdAu nanoparticles during vinyl acetate synthesis enhanced by potassium acetate. ACS Catal. 5, 5776–5786 (2015).
Nakamura, S. & Yasui, T. The mechanism of the palladium-catalyzed synthesis of vinyl acetate from ethylene in a heterogeneous gas reaction. J. Catal. 17, 366–374 (1970).
Nakamura, S. & Yasui, T. Formation of palladous acetate and stability of catalyst in palladium-metal-catalyzed synthesis of vinyl acetate from ethylene. J. Catal. 23, 315–320 (1971).
Augustine, S. M. & Blitz, J. P. The use of DRIFTS-MS and kinetic studies to determine the role of acetic acid in the palladium-catalyzed vapor-Phase synthesis of vinyl acetate. J. Catal. 142, 312–324 (1993).
Pohl, M.-M. et al. Bimetallic PdAu–KOac/SiO2 catalysts for vinyl acetate monomer (VAM) synthesis: Insights into deactivation under industrial conditions. J. Catal. 262, 314–323 (2009).
Jacobs, H. P. et al. Promotional role of acid sites on aluminosilicate-supported PdAu for vinyl acetate synthesis. ACS Catal. 14, 211–226 (2024).
Samanos, B., Boutry, P. & Montarnal, R. The mechanism of vinyl acetate formation by gas-phase catalytic ethylene acetoxidation. J. Catal. 23, 19–30 (1971).
Calaza, F., Stacchiola, D., Neurock, M. & Tysoe, W. T. Coverage effects on the palladium-catalyzed synthesis of vinyl acetate: comparison between theory and experiment. J. Am. Chem. Soc. 132, 2202–2207 (2010).
Calaza, F., Mahapatra, M., Neurock, M. & Tysoe, W. T. Disentangling ensemble, electronic and coverage effects on alloy catalysts: Vinyl acetate synthesis on Au/Pd(111). J. Catal. 312, 37–45 (2014).
Neurock, M. et al. First principle analysis of the catalytic reaction pathways in the synthesis of vinyl acetate. Chem. Eng. Sci. 51, 1691–1699 (1996).
Neurock, M. Perspectives on the first principles elucidation and the design of active sites. J. Catal. 216, 73–88 (2003).
Neurock, M. & Tysoe, W. T. Mechanistic insights in the catalytic synthesis of vinyl acetate on palladium and gold/palladium alloy surfaces. Top. Catal. 56, 1314–1332 (2013).
Neurock, M. & Tysoe, W. T. Effect of coverage on catalytic selectivity and activity on metallic and alloy catalysts; vinyl acetate monomer synthesis. Top. Catal. 61, 722–735 (2018).
Stacchiola, D. et al. Elucidation of the reaction mechanism for the palladium-catalyzed synthesis of vinyl acetate. Angew. Chem. Int. Ed. 44, 4572–4574 (2005).
Zhang, M., Hao, Q. & Yu, Y. Pd/Au(100) alloy for vinyl acetate synthesis: Effects of surface properties on reagents adsorption. Comput. Theor. Chem. 1019, 33–38 (2013).
Liu, P. & Nørskov, J. K. Ligand and ensemble effects in adsorption on alloy surfaces. Phys. Chem. Chem. Phys. 3, 3814–3818 (2001).
Yi, C.-W., Luo, K., Wei, T. & Goodman, D. W. The composition and structure of Pd−Au surfaces. J. Phys. Chem. B 109, 18535–18540 (2005).
Chen, M., Kumar, D., Yi, C.-W. & Goodman, D. W. The promotional effect of gold in catalysis by palladium-gold. Science 310, 291–293 (2005).
Chen, M. S. et al. The nature of the active site for vinyl acetate synthesis over Pd–Au. Catal. Today 117, 37–45 (2006).
Kumar, D., Chen, M. S. & Goodman, D. W. Synthesis of vinyl acetate on Pd-based catalysts. Catal. Today 123, 77–85 (2007).
Gravelle-Rumeau-Maillot, M., Pitchon, V., Martin, G. A. & Praliaud, H. Complementary study by calorimetry and infrared spectroscopy of alkali metal doped Pd/SiO2 solids: adsorption of hydrogen and carbon monoxide. Appl. Catal. A Gen. 98, 45–59 (1993).
Pitchon, V., Guenin, M. & Praliaud, H. X-ray photoelectron spectroscopic study of the electronic state of palladium in alkali metal doped Pd/SiO2 solids. Appl. Catal. 63, 333–343 (1990).
Crathorne, E. A., Macgowan, D., Morris, S. R. & Rawlinson, A. P. Application of isotopic transient kinetics to vinyl acetate catalysis. J. Catal. 149, 254–267 (1994).
Kragten, D. D., van Santen, R. A., Crawford, M. K., Provine, W. D. & Lerou, J. J. A spectroscopic study of the homogeneous catalytic conversion of ethylene to vinyl acetate by palladium acetate. Inorg. Chem. 38, 331–339 (1999).
Reilly, C. R. & Lerou, J. J. Supported liquid phase catalysis in selective oxidation. Catal. Today 41, 433–441 (1998).
Stoyanov, E. S. Ir study of the structure of palladium(ii) acetate in chloroform, acetic acid, and their mixtures in solution and in liquid-solid subsurface layers. J. Struct. Chem. 41, 440–445 (2000).
Pandey, R. N. & Henry, P. M. Interaction of palladium(II) acetate with sodium and lithium acetate in acetic acid. Can. J. Chem. 52, 1241–1247 (1974).
Pandey, R. N. & Henry, P. M. Equilibria between palladium(II) acetate and olefins in acetic acid containing sodium acetate. Can. J. Chem. 53, 1833–1841 (1975).
Jacobs, H. P. et al. Impregnation of KOAc on PdAu/SiO2 causes Pd-acetate formation and metal restructuring. J. Mater. Chem. A 11, 6918–6933 (2023).
Gonzalez Caranton, A. R. et al. Nanostructured Pd−Cu catalysts supported on Zr−Al and Zr−Ti for synthesis of vinyl acetate. ChemCatChem 10, 5256–5269 (2018).
Miyaura, N. orio & Suzuki, A. kira Palladium-catalyzed cross-coupling reactions of organoboron compounds. Chem. Rev. 95, 2457–2483 (1995).
Tan, J., Chen, Y., Li, H. & Yasuda, N. Suzuki-miyaura cross-coupling reactions of unprotected haloimidazoles. J. Org. Chem. 79, 8871–8876 (2014).
Wolfe, J. P., Singer, R. A., Yang, B. H. & Buchwald, S. L. Highly active palladium catalysts for suzuki coupling reactions. J. Am. Chem. Soc. 121, 9550–9561 (1999).
Heck, R. F. & Nolley, J. P. Jr. Palladium-catalyzed vinylic hydrogen substitution reactions with aryl, benzyl, and styryl halides. J. Org. Chem. 37, 2320–2322 (1972).
Xu, H.-J., Zhao, Y.-Q. & Zhou, X.-F. Palladium-catalyzed heck reaction of aryl chlorides under mild conditions promoted by organic ionic bases. J. Org. Chem. 76, 8036–8041 (2011).
Shen, Q., Ogata, T. & Hartwig, J. F. Highly reactive, general and long-lived catalysts for palladium-catalyzed amination of heteroaryl and aryl chlorides, bromides, and iodides: scope and structure–activity relationships. J. Am. Chem. Soc. 130, 6586–6596 (2008).
Topchiy, M. A., Asachenko, A. F. & Nechaev, M. S. Solvent-free buchwald–hartwig reaction of aryl and heteroaryl halides with secondary amines. Eur. J. Org. Chem. 2014, 3319–3322 (2014).
Wolfe, J. P., Tomori, H., Sadighi, J. P., Yin, J. & Buchwald, S. L. Simple, efficient catalyst system for the palladium-catalyzed amination of aryl chlorides, bromides, and triflates. J. Org. Chem. 65, 1158–1174 (2000).
Gorelsky, S. I., Lapointe, D. & Fagnou, K. Analysis of the concerted metalation-deprotonation mechanism in palladium-catalyzed direct arylation across a broad range of aromatic substrates. J. Am. Chem. Soc. 130, 10848–10849 (2008).
Kim, S. H. & Chang, S. Highly efficient and versatile Pd-catalyzed direct alkynylation of both azoles and azolines. Org. Lett. 12, 1868–1871 (2010).
McNally, A., Haffemayer, B., Collins, B. S. L. & Gaunt, M. J. Palladium-catalysed C–H activation of aliphatic amines to give strained nitrogen heterocycles. Nature 510, 129–133 (2014).
Smidt, J. et al. The oxidation of olefins with palladium chloride catalysts. Angew. Chem. Int. Ed. Engl. 1, 80–88 (1962).
Moiseev, I. I., Vargaftik, M. N. & Syrkin, Y. K. On the mechanism of the reaction of palladium salts with olefines in hydroxyl-containing solutions. Dokl. Akad. Nauk SSSR 133, 377–380 (1960).
Owen, E. A. & Yates, E. L. XLI. Precision measurements of crystal parameters. Lond., Edinb. Dublin Philos. Mag. J. Sci. 15, 472–488 (1933).
Giannakakis, G. et al. Single-atom alloy formation via reaction-driven catalyst restructuring. J. Am. Chem. Soc. 146, 20989–20995 (2024).
Li, W., Ivanov, S., Mozaffari, S., Shanaiah, N. & Karim, A. M. Palladium acetate trimer: understanding its ligand-induced dissociation thermochemistry using isothermal titration calorimetry, X-ray absorption fine structure, and 31 P nuclear magnetic resonance. Organometallics 38, 451–460 (2019).
Adrio, L. A., Nguyen, B. N., Guilera, G., Livingston, A. G. & Hii, K. K. Mimi). Speciation of Pd(OAc)2 in ligandless Suzuki–Miyaura reactions. Catal. Sci. Technol. 2, 316–323 (2012).
Brijaldo, M. H., Rojas, H. A., Martínez, J. J. & Passos, F. B. Effect of support on acetic acid decomposition over palladium catalysts. J. Catal. 331, 63–75 (2015).
Green, I. X., Tang, W., Neurock, M. & Yates, J. T. Mechanistic insights into the partial oxidation of acetic acid by O2 at the dual perimeter sites of a Au/TiO2 catalyst. Faraday Discuss 162, 247–265 (2013).
McDonald, R. S. Surface functionality of amorphous silica by infrared spectroscopy. J. Phys. Chem. 62, 1168–1178 (1958).
Sirita, J., Phanichphant, S. & Meunier, F. C. Quantitative analysis of adsorbate concentrations by diffuse reflectance FT-IR. Anal. Chem. 79, 3912–3918 (2007).
Spreadborough, J. & Christian, J. W. High-temperature X-ray diffractometer. J. Sci. Instrum. 36, 116–118 (1959).
Han, Y.-F., Kumar, D. & Goodman, D. W. Particle size effects in vinyl acetate synthesis over Pd/SiO2. J. Catal. 230, 353–358 (2005).
Han, Y.-F., Kumar, D., Sivadinarayana, C., Clearfield, A. & Goodman, D. W. The formation of PdCx over Pd-based catalysts in vapor-phase vinyl acetate synthesis: Does a Pd–Au alloy catalyst resist carbide formation? Catal. Lett. 94, 131–134 (2004).
Chan, C. W. A. et al. New environmentally friendly catalysts containing Pd–interstitial carbon made from Pd–glucose precursors for ultraselective hydrogenations in the liquid phase. Chem. Commun. 47, 7971–7973 (2011).
Bugaev, A. L. et al. Palladium carbide and hydride formation in the bulk and at the surface of palladium nanoparticles. J. Phys. Chem. C. 122, 12029–12037 (2018).
Bugaev, A. L. et al. In situ formation of hydrides and carbides in palladium catalyst: When XANES is better than EXAFS and XRD. Catal. Today 283, 119–126 (2017).
Miller, J. T. et al. The effect of gold particle size on AuAu bond length and reactivity toward oxygen in supported catalysts. J. Catal. 240, 222–234 (2006).
Breckner, C. J. et al. Controlled site coverage of strong metal–support interaction (SMSI) on Pd NP catalysts. Catal. Sci. Technol. 13, 157–169 (2023).
Davidson, A. W. & McAllister, W. H. Solutions of salts in pure acetic acid. II. solubilities of acetates. J. Am. Chem. Soc. 52, 507–519 (1930).
Rivalta, I., Mazzone, G., Russo, N. & Sicilia, E. Adsorption of ethylene, vinyl, acetic acid, and acetate species on PdAu(111) and PdAu(100) surface alloys: A cluster model sudy. J. Chem. Theory Comput. 5, 1350–1360 (2009).
Weltner, W. Jr. The vibrational spectrum, associative and thermodynamic properties of acetic acid vapor. J. Am. Chem. Soc. 77, 3941–3950 (1955).
Stacchiola, D., Calaza, F., Burkholder, L. & Tysoe, W. T. Vinyl Acetate Formation by the Reaction of Ethylene with Acetate Species on Oxygen-Covered Pd(111). J. Am. Chem. Soc. 126, 15384–15385 (2004).
Herzberg, G. Molecular Spectra and Molecular Structure. II, Infrared and Raman Spectra of Polyatomic Molecules. (D. Van Nostrand Company, Toronto, 1945).
Plata, J. J., García-Mota, M., Braga, A. A. C., López, N. & Maseras, F. Vinyl acetate synthesis on homogeneous and heterogeneous Pd-based catalysts: a theoretical analysis on the reaction mechanisms. J. Phys. Chem. A 113, 11758–11762 (2009).
Grover, G. S. & Chaudhari, R. V. Kinetics of oxidation of ethylene to vinyl acetate using a homogeneous palladium complex catalyst. Chem. Eng. J. 32, 93–99 (1986).
Carole, W. A. & Colacot, T. J. Understanding palladium acetate from a user perspective. Chem. – A Eur. J. 22, 7686–7695 (2016).
Simson, S., Jentys, A. & Lercher, J. A. Dynamic self-organization of supported Pd/Au catalysts during vinyl acetate synthesis. J. Phys. Chem. C 117, 8161–8169 (2013).
Han, Y.-F., Kumar, D., Sivadinarayana, C. & Goodman, D. W. Kinetics of ethylene combustion in the synthesis of vinyl acetate over a Pd/SiO2 catalyst. J. Catal. 224, 60–68 (2004).
Han, Y.-F., Wang, J.-H., Kumar, D., Yan, Z. & Goodman, D. W. A kinetic study of vinyl acetate synthesis over Pd-based catalysts: kinetics of vinyl acetate synthesis over Pd–Au/SiO2 and Pd/SiO2 catalysts. J. Catal. 232, 467–475 (2005).
Scherrer, P. Bestimmung der inneren Struktur und der Größe von Kolloidteilchen mittels Röntgenstrahlen. in Kolloidchemie Ein Lehrbuch (ed. Zsigmondy, R.) 387–409 (Springer, Berlin, Heidelberg, 1912).
Leshchev, D. et al. The inner shell spectroscopy beamline at NSLS-II: a facility for in situ and operando X-ray absorption spectroscopy for materials research. J. Synchrotron Rad. 29, 1095–1106 (2022).
Ressler, T. WinXAS: a Program for X-ray Absorption Spectroscopy Data Analysis under MS-Windows. J. Synchrotron Rad. 5, 118–122 (1998).
Rehr, J. J., Booth, C. H., Bridges, F. & Zabinsky, S. I. X-ray-absorption fine structure in embedded atoms. Phys. Rev. B 49, 12347–12350 (1994).
Neese, F. The ORCA program system. Wiley Interdiscip. Rev. Comput. Mol. Sci. 2, 73–78 (2012).
Neese, F. Software update: The ORCA program system—Version 5.0. WIREs Comput. Mol. Sci. 12, e1606 (2022).
Neese, F. Software update: the ORCA program system, version 4.0. WIREs Comput. Mol. Sci. 8, 4–9 (2018).
Adamo, C. & Barone, V. Toward reliable density functional methods without adjustable parameters: The PBE0 model. J. Chem. Phys. 110, 6158–6170 (1999).
Weigend, F. & Ahlrichs, R. Balanced basis sets of split valence, triple zeta valence and quadruple zeta valence quality for H to Rn: Design and assessment of accuracy. Phys. Chem. Chem. Phys. 7, 3297–3305 (2005).
Weigend, F. Accurate Coulomb-fitting basis sets for H to Rn. Phys. Chem. Chem. Phys. 8, 1057 (2006).
Grimme, S., Antony, J., Ehrlich, S. & Krieg, H. A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu. J. Chem. Phys. 132, 154104 (2010).
Neese, F., Wennmohs, F., Hansen, A. & Becker, U. Efficient, approximate and parallel Hartree–Fock and hybrid DFT calculations. A ‘chain-of-spheres’ algorithm for the Hartree–Fock exchange. Chem. Phys. 356, 98–109 (2009).
Staroverov, V. N., Scuseria, G. E., Tao, J. & Perdew, J. P. Comparative assessment of a new nonempirical density functional: Molecules and hydrogen-bonded complexes. J. Chem. Phys. 119, 12129–12137 (2003).
Staroverov, V. N., Scuseria, G. E., Tao, J. & Perdew, J. P. Erratum: “Comparative assessment of a new nonempirical density functional: Molecules and hydrogen-bonded complexes”.J. Chem. Phys. 121, 11507 (2004).
Bursch, M., Neugebauer, H., Ehlert, S. & Grimme, S. Dispersion corrected r2SCAN based global hybrid functionals: r2SCANh, r2SCAN0, and r2SCAN50. J. Chem. Phys. 156, 134105 (2022).
Acknowledgements
The authors gratefully acknowledge support from Celanese Corporation and ACS-PRF Grant #61790-ND5. The authors thank Ted T. Kim and Wei-Ling Huang for their assistance with XAS data collection. This research used resources at the 8-ID beamline of the National Synchrotron Light Source II, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Brookhaven National Laboratory under Contract No. DE-SC0012704. H.P.J. and A.C.F. additionally acknowledge work performed in collaboration with ChemCatBio, a member of the Energy Materials Network (EMN) and supported by the U.S. Department of Energy (DOE) Bioenergy Technology Office under Contract DE-AC05-00OR22725 (ORNL), Contract DE-AC36-08-GO28308 (NREL), and Contract DE-AC02-06CH11357 (ANL). Microscopy was performed as part of a user project at the Center for Nanophase Materials Sciences (CNMS), which is a U.S. DOE Office of Science User Facility.
Author information
Authors and Affiliations
Contributions
H.P.J. and W.C.E. contributed equally to this work. H.P.J.: conceptualization, investigation, formal analysis, visualization, writing – original draft. W.C.E.: conceptualization, investigation, formal analysis, visualization, writing – review & editing. K.N.H.: conceptualization, supervision, writing – review & editing. S.J.: investigation, formal analysis, visualization, writing – review & editing. J.J.D.: supervision, writing – review & editing. C.S.-P.: investigation, formal analysis, writing – review & editing. A.C.F.: investigation, formal analysis, visualization, writing – review & editing. B.J.C.: conceptualization, investigation, writing – review & editing. J.H.A.: methodology. L.C., S.G.M., and S.R.A.: supervision, writing – review & editing. T.P.S.: resources, supervision, writing – review & editing. J.T.M.: resources, supervision, writing – review & editing. M.S.W.: conceptualization, funding acquisition, project administration, supervision, writing – review & editing.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests
Peer review
Peer review information
Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Jacobs, H.P., Elias, W.C., Heck, K.N. et al. Dynamic behavior of molecular Pd-acetate trimers and dimers in heterogeneous vinyl acetate synthesis. Nat Commun 17, 127 (2026). https://doi.org/10.1038/s41467-025-66820-7
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s41467-025-66820-7
