Loading...
Computational Characterization of Amorphous Silica Supported Catalysts
Jystad, Amy
Jystad, Amy
Citations
Altmetric:
Abstract
Metal-doped amorphous silicates are effective acidic catalysts for numerous industrial applications, including dehydration of alcohols, hydrodeoxygenation of biomass, and metathesis of olefins. The catalytic activity of metal-doped amorphous silicates is directly related to their acidity, which is imparted by the metal dopant. However, the structure of the metal dopant site is difficult to char- acterize experimentally as it is influenced by the irregular, amorphous silica support. Insights from simulations can greatly aid the structural characterization of metal-doped amorphous silica, which in turn may lead to improved catalyst design. To provide atomistic insight into these systems, we have developed the METal-doped Amorphous SIlicate Library (METASIL), comprising 70 Zr-, W-, and Nb-doped amorphous silicate cluster models, which may be used for the simulation of the catalytic activity of real mesoporous silicates. The validity of this library is tested with simulations of various types of experimental measurements.
We study the adsorption of ammonia molecules on the metal (M), MOH and SiOH sites to model measurements of NH3 temperature-programmed desorption (NH3-TPD). Our calculations reproduce the experimental trends of acidity strength across metals: Zr > W > Nb, where Zr sites are predominantly Lewis acids, while Nb and W sites are both Brønsted and Lewis acids. Metal atoms that are more grafted into the silica (i.e., have more M–O–Si bonds) exhibit stronger Lewis acidity. Pseudo bridging silanols, stabilized by a nearby metal atom, exhibit the strongest affinity for ammonia, thus representing the most likely source of Brønsted acidity in these materials.
We also simulate pyridine adsorbed Fourier-transform Infrared (Py-FTIR) and 15N nuclear magnetic resonance (NMR) spectra to further characterize the local structure and acidity strength of metal-doped silicate clusters. By matching the calculated normal mode frequencies with ex- perimental IR spectra, we reproduce the differentiation between Lewis and strong Brønsted acid sites. However, our simulations show that the coordination of Py to weak Brønsted acids is indistinguishable from that to metal sites, as they elicit similar frequency shifts. On the other hand, 15N NMR can distinguish between Py interacting with metal centers, weak and strong Brønsted acid sites, as well as showing potential in identifying the structure of metal dopants.
We use DFT simulations to obtain NMR spectra and quadrupolar coupling constants of METASIL to provide more detailed structural information on the metal sites. The shieldings are calculated for 29Si, 17O, and 1H, which reproduce available experimental data and provide predictions for future measurements. In particular, our simulations show that 17O NMR can provide direct information regarding the structure of the metal dopant, as bonding to the metal center (MOSi and MOH) is quantitatively distinguishable from that to the silica (SiOSi and SiOH).
We characterize W-doped amorphous silica, a catalyst the Subramaniam group at the University of Kansas used to convert 2-butene and ethylene to propene. We simulate the 29Si chemical shifts of W-doped amorphous silica clusters to confirm the experimental assignment of the W–O–Si species, i.e., the expected active species in olefin metathesis. We further examine bimetallic doped amorphous silica clusters that contain W and an additional metal dopant of Nb, Zr, Hf, or Ta, and compare structural descriptors for possible correlations with reactivity.
The immobilization of a molecular palladium(II) complex on amorphous silica is investigated in collaboration with the Vannucci group at the University of South Carolina. Due to the irregu- larity of the support, the binding mode and strength of the catalyst onto the support is difficult to characterize experimentally. Our simulations show that the catalysts can adhere to the oxide sur- face through hydrogen bonding interaction, Coulombic attraction between the charged molecule and the oxide surface, or covalent bonding.
In order to improve the methodological approach to contrast realistic models, we investigated the geometry optimization protocol used to relax metal-doped amorphous silica, in collaboration with the Peters group at the University of Illinois Urbana–Champaign. As a large number of structurally different amorphous silica clusters are needed to obtain macroscopic data, such as rate constants, an efficient and robust computational protocol must be developed. We find that a functionalized silica slab obtained from MD simulation must be pre-optimized with DFT before doping with a metal center.
This work has established a procedure to model amorphous silica supported catalysts, which
has provided direct insight into the irregular support’s impact on the catalytic sites.
We study the adsorption of ammonia molecules on the metal (M), MOH and SiOH sites to model measurements of NH3 temperature-programmed desorption (NH3-TPD). Our calculations reproduce the experimental trends of acidity strength across metals: Zr > W > Nb, where Zr sites are predominantly Lewis acids, while Nb and W sites are both Brønsted and Lewis acids. Metal atoms that are more grafted into the silica (i.e., have more M–O–Si bonds) exhibit stronger Lewis acidity. Pseudo bridging silanols, stabilized by a nearby metal atom, exhibit the strongest affinity for ammonia, thus representing the most likely source of Brønsted acidity in these materials.
We also simulate pyridine adsorbed Fourier-transform Infrared (Py-FTIR) and 15N nuclear magnetic resonance (NMR) spectra to further characterize the local structure and acidity strength of metal-doped silicate clusters. By matching the calculated normal mode frequencies with ex- perimental IR spectra, we reproduce the differentiation between Lewis and strong Brønsted acid sites. However, our simulations show that the coordination of Py to weak Brønsted acids is indistinguishable from that to metal sites, as they elicit similar frequency shifts. On the other hand, 15N NMR can distinguish between Py interacting with metal centers, weak and strong Brønsted acid sites, as well as showing potential in identifying the structure of metal dopants.
We use DFT simulations to obtain NMR spectra and quadrupolar coupling constants of METASIL to provide more detailed structural information on the metal sites. The shieldings are calculated for 29Si, 17O, and 1H, which reproduce available experimental data and provide predictions for future measurements. In particular, our simulations show that 17O NMR can provide direct information regarding the structure of the metal dopant, as bonding to the metal center (MOSi and MOH) is quantitatively distinguishable from that to the silica (SiOSi and SiOH).
We characterize W-doped amorphous silica, a catalyst the Subramaniam group at the University of Kansas used to convert 2-butene and ethylene to propene. We simulate the 29Si chemical shifts of W-doped amorphous silica clusters to confirm the experimental assignment of the W–O–Si species, i.e., the expected active species in olefin metathesis. We further examine bimetallic doped amorphous silica clusters that contain W and an additional metal dopant of Nb, Zr, Hf, or Ta, and compare structural descriptors for possible correlations with reactivity.
The immobilization of a molecular palladium(II) complex on amorphous silica is investigated in collaboration with the Vannucci group at the University of South Carolina. Due to the irregu- larity of the support, the binding mode and strength of the catalyst onto the support is difficult to characterize experimentally. Our simulations show that the catalysts can adhere to the oxide sur- face through hydrogen bonding interaction, Coulombic attraction between the charged molecule and the oxide surface, or covalent bonding.
In order to improve the methodological approach to contrast realistic models, we investigated the geometry optimization protocol used to relax metal-doped amorphous silica, in collaboration with the Peters group at the University of Illinois Urbana–Champaign. As a large number of structurally different amorphous silica clusters are needed to obtain macroscopic data, such as rate constants, an efficient and robust computational protocol must be developed. We find that a functionalized silica slab obtained from MD simulation must be pre-optimized with DFT before doping with a metal center.
This work has established a procedure to model amorphous silica supported catalysts, which
has provided direct insight into the irregular support’s impact on the catalytic sites.
Description
Date
2020-01-01
Journal Title
Journal ISSN
Volume Title
Publisher
University of Kansas
Collections
Archive Status
This item contains archived web content.
Files
Jystad_ku_0099D_17318.pdf
Adobe PDF, 40.91 MB
- Embargoed until 2170-05-31
Research Projects
Organizational Units
Journal Issue
Keywords
amorphous silica, cluster model, DFT, heterogeneous catalysis, metal-doped