Periodic Reporting for period 4 - DYNPOR (First principle molecular dynamics simulations for complex chemical transformations in nanoporous materials)
Période du rapport: 2020-02-01 au 2021-07-31
Résumé du contexte et des objectifs généraux du projet
DYNPOR aims in simulating chemical and physical transformation taking place in nanoporous materials at operating conditions. Nanoporous materials are omnipresent in our daily society, they are vital in catalysis to convert chemical feedstocks to useful building blocks in a sustainable way, they are ubiquitous in the field of separation and to induce clean energy conversions. DYNPOR starts from the hypothesis that the function of nanoporous materials is critically dependent on the conditions in which they do the work. Such conditions may relate to true reaction temperature, guest loading, external pressure, etc. Within DYNPOR methods have been developed to follow chemical transformations at operating conditions. To this end molecular dynamics methods have been developed which allow to account for the full complexity of the free energy surface. The processes under investigation are mostly activated, which means that the probability of sampling these processes during regular molecular dynamics simulations as too low to be observed within a reasonable time. For this reason, methods are developed to sample such rare events.
The applications of DYNPOR are situated in the field of heterogeneous catalysis within zeolites and phase transformations of metal-organic frameworks and other flexible materials. Chemical conversions taking place in zeolites are omnipresent in chemical industry and are vital for future technologies to convert non-fossil based feedstocks. Metal-Organic Frameworks are materials which are built from metal nodes and organic linkers and which show a unique dynamic response upon external stimuli making them tractable for applications in gas storage, separation, catalysis.
The work has been performed in close collaboration with experimental partners. On one hand extensive benchmarking of developed methods has been performed with experimental kinetic data to predict reaction rate constants at operating conditions for important reactions. Based on the new methods, we have charted reaction mechanisms for important chemical processes such as the methanol to olefin conversion, alkene cracking in zeolites, CO2 to hydrocarbons conversion, etc. The methods have also been extended to study transport of reactants, products from and away from the active site taking into account realistic loadings within the nanoporous frameworks and the nature of active sties. This led to unique insights into the activity, selectivity and lifetime of heterogeneous catalysts and thanks to the new approach molecular level suggestions can be made to propose new selective and active catalysts. Thanks to DYNPOR a unique set of methods has been developed which allowed to make the transition from a static description of realistic nanoporous materials towards a molecular dynamics view.
The work is essential in designing of new catalysts and materials for sustainable chemical and clean energy conversions. Our society is confronted with important challenges, such as how to provide a clean and affordable energy system without endangering future generation, how to protect our limited natural resources, how to combat environmental pollution, how to mitigate climate change. Functional nanomaterials play a pivotal role in the development of future technological solutions to provide answers to these questions. From a purely experimental point of view it is very difficult to establish a causal relation between nanometer scale structural modifications and the observed macroscopic function of the material. Modeling functional nanostructured materials in close synergy with experimental groups may give unique nanoscopic insight into the desired function, provided the material is modelled in a realistic way and is studied at the conditions in which it does the work. DYNPOR has induced a paradigm shift in modelling nanoporous materials at real working conditions enabling design of functional nanomaterials for sustainable chemical conversions.
The applications of DYNPOR are situated in the field of heterogeneous catalysis within zeolites and phase transformations of metal-organic frameworks and other flexible materials. Chemical conversions taking place in zeolites are omnipresent in chemical industry and are vital for future technologies to convert non-fossil based feedstocks. Metal-Organic Frameworks are materials which are built from metal nodes and organic linkers and which show a unique dynamic response upon external stimuli making them tractable for applications in gas storage, separation, catalysis.
The work has been performed in close collaboration with experimental partners. On one hand extensive benchmarking of developed methods has been performed with experimental kinetic data to predict reaction rate constants at operating conditions for important reactions. Based on the new methods, we have charted reaction mechanisms for important chemical processes such as the methanol to olefin conversion, alkene cracking in zeolites, CO2 to hydrocarbons conversion, etc. The methods have also been extended to study transport of reactants, products from and away from the active site taking into account realistic loadings within the nanoporous frameworks and the nature of active sties. This led to unique insights into the activity, selectivity and lifetime of heterogeneous catalysts and thanks to the new approach molecular level suggestions can be made to propose new selective and active catalysts. Thanks to DYNPOR a unique set of methods has been developed which allowed to make the transition from a static description of realistic nanoporous materials towards a molecular dynamics view.
The work is essential in designing of new catalysts and materials for sustainable chemical and clean energy conversions. Our society is confronted with important challenges, such as how to provide a clean and affordable energy system without endangering future generation, how to protect our limited natural resources, how to combat environmental pollution, how to mitigate climate change. Functional nanomaterials play a pivotal role in the development of future technological solutions to provide answers to these questions. From a purely experimental point of view it is very difficult to establish a causal relation between nanometer scale structural modifications and the observed macroscopic function of the material. Modeling functional nanostructured materials in close synergy with experimental groups may give unique nanoscopic insight into the desired function, provided the material is modelled in a realistic way and is studied at the conditions in which it does the work. DYNPOR has induced a paradigm shift in modelling nanoporous materials at real working conditions enabling design of functional nanomaterials for sustainable chemical conversions.
Travail effectué depuis le début du projet jusqu’à la fin de la période considérée dans le rapport et principaux résultats atteints jusqu’à présent
With DYNPOR first principle molecular dynamics methods have been developed to simulate chemical reactions within nanoporous materials, i.e. zeolites, metal-organic frameworks at operating conditions. To this end we have developed free energy methods based on enhanced sampling techniques to simulate rare events which have a low probability of occurring during a normal molecular dynamics run.
We have benchmarked various enhanced sampling methods such as thermodynamic integration, umbrella sampling, metadynamics, etc. to accurately reproduce kinetic data. For some zeolite catalyzed reactions, very good benchmark kinetic data are available in literature to validate the developed methods. A key example is the methylation of olefins within Bronsted acidic zeolites, for which accurate reaction rates are available from experiment. We have developed first principle molecular dynamics methods that enable to predict reaction rates with chemical accuracy for well defined reactions taking place at well defined active sites.
The new methods have been used to unravel timely applications within zeolite catalysis in close collaboration with experimental partners. Within the methanol to olefin process, we have shown how presence of water or other protic molecules such as methanol, may induce a dynamic behaviour of the active sites. These findings are important to optimize catalysts towards enhanced lifetimes. Together with experimental partners, we have developed new catalysts, with enhanced lifetimes and better selectivity for propene. To this end zeolites were postsynthetically modified by the experimental partners and with theoretical methods we have unraveled how such postsynthetic modifications impact reaction mechanisms and stabilities of important reaction intermediates. Also with the methanol to olefin process, we have followed on the fly the formation of important coke precursors using operando Raman spectroscopy and molecular dynamics simulations. We have obtained unique insight into the species giving rise to deactivation of the catalysts. This insight has been shown instrumental to design catalysts with the optimal topology to reduce deactivation. For some catalysts it was observed how the selectivity changes substantially with time on stream. To this end we have developed new free energy methods that allow to simulate hindered diffusion of product species such as olefins through small pore zeolites. We have found an important effect of acid site density and presence of hydrocarbon pool species within the zeolite on the transport properties. The findings are instrumental for designing new materials which allow good separation properties.
A second important pillar of the project has focused on stimuli responsive behavior in metal-organic frameworks.
MOFs have the unique property to respond in a cooperative way to large external triggers, due to their large variety of nanoscale interactions with different strengths. Soft porous crystals (SPCs) or flexible MOFs can undergo a phase transitions between various crystalline phases, while rigid MOFs undergo transformations between a crystalline and a largely amorphous phase. For applications situated in sensing, separation, it is of utmost importance to understand and predict under which conditions a material becomes flexible, when a material can adsorb in an efficient way molecules and understand the stability window of MOFs. Within this program we have developed a variety of new methods that allow to construct the underlying thermodynamic potential in terms of the important state variables dictating the observed behaviour. To this end we have developed a force field protocol QUICKFF which allows to deduce classical force fields from underlying ab initio data. The generated force fields are used in molecular dynamics simulations to generate the pressure versus volume equations of state. To this end we have developed specific thermodynamic ensembles for flexible frameworks where the volume can be fixed but where the shape can vary. The developed software procedures have been made available via open source (http://molmod.ugent.be/software(s’ouvre dans une nouvelle fenêtre)). The work has been widely picked up the scientific community and the new thermodynamic ensembles were integrated in large scale molecular dynamics engines such as LAMMPS, OPENMM,etc. With the new thermodynamic protocol, we were able to identify the external conditions leading to single-crystal-to-single-crystal transitions or transitions that are accompanied by a loss of crystallinity.
The approach has largely been applied on a broad set of materials in collaboration with experimental partners with application in sensing and actuating, development of nanodampers, shock adsorbers, thermal triggers, etc.
Very recently, we have discovered how the size of the crystal impacts the phase transformations within MOFs. Experimentally it was observed that defects and finite size of the crystal may largely impact the flexible behavior of MOFs. We have substantially pushed the limits of our force field simulations to enable simulation of very large systems where disorder may spontaneously occur during a molecular dynamics run.
Throughout the program, an important connection was made with operando spectroscopic measurements to characterize the behavior of the materials. We have developed and benchmarked various methods that allow to calculate operando spectroscopic signals from computational point of view, such as vibrational fingerprints, UV-Vis spectra, etc. As such it became possible to identify important intermediates observed experimentally and to follow chemical conversion with time on stream.
The work has been extensively disseminated via various channels. First of all the program lead to a substantial number of papers in high profile papers, which received high impact from the scientific community. Various papers were selected as Very Important Papers and have been highlighted in editorials or as covers. Second various researchers of DYNPOR have extensively communicated their results by oral contributions, invited talks on leading international conferences. We have also disseminated the results to a broader audience by launching on various occasions press releases, which were picked up by various media channels.
Our software has been made available by open access for the whole scientific community and is now used by various labs worldwide, furthermore various of our new methods were integrated in other large scale molecular dynamics engines.
We have benchmarked various enhanced sampling methods such as thermodynamic integration, umbrella sampling, metadynamics, etc. to accurately reproduce kinetic data. For some zeolite catalyzed reactions, very good benchmark kinetic data are available in literature to validate the developed methods. A key example is the methylation of olefins within Bronsted acidic zeolites, for which accurate reaction rates are available from experiment. We have developed first principle molecular dynamics methods that enable to predict reaction rates with chemical accuracy for well defined reactions taking place at well defined active sites.
The new methods have been used to unravel timely applications within zeolite catalysis in close collaboration with experimental partners. Within the methanol to olefin process, we have shown how presence of water or other protic molecules such as methanol, may induce a dynamic behaviour of the active sites. These findings are important to optimize catalysts towards enhanced lifetimes. Together with experimental partners, we have developed new catalysts, with enhanced lifetimes and better selectivity for propene. To this end zeolites were postsynthetically modified by the experimental partners and with theoretical methods we have unraveled how such postsynthetic modifications impact reaction mechanisms and stabilities of important reaction intermediates. Also with the methanol to olefin process, we have followed on the fly the formation of important coke precursors using operando Raman spectroscopy and molecular dynamics simulations. We have obtained unique insight into the species giving rise to deactivation of the catalysts. This insight has been shown instrumental to design catalysts with the optimal topology to reduce deactivation. For some catalysts it was observed how the selectivity changes substantially with time on stream. To this end we have developed new free energy methods that allow to simulate hindered diffusion of product species such as olefins through small pore zeolites. We have found an important effect of acid site density and presence of hydrocarbon pool species within the zeolite on the transport properties. The findings are instrumental for designing new materials which allow good separation properties.
A second important pillar of the project has focused on stimuli responsive behavior in metal-organic frameworks.
MOFs have the unique property to respond in a cooperative way to large external triggers, due to their large variety of nanoscale interactions with different strengths. Soft porous crystals (SPCs) or flexible MOFs can undergo a phase transitions between various crystalline phases, while rigid MOFs undergo transformations between a crystalline and a largely amorphous phase. For applications situated in sensing, separation, it is of utmost importance to understand and predict under which conditions a material becomes flexible, when a material can adsorb in an efficient way molecules and understand the stability window of MOFs. Within this program we have developed a variety of new methods that allow to construct the underlying thermodynamic potential in terms of the important state variables dictating the observed behaviour. To this end we have developed a force field protocol QUICKFF which allows to deduce classical force fields from underlying ab initio data. The generated force fields are used in molecular dynamics simulations to generate the pressure versus volume equations of state. To this end we have developed specific thermodynamic ensembles for flexible frameworks where the volume can be fixed but where the shape can vary. The developed software procedures have been made available via open source (http://molmod.ugent.be/software(s’ouvre dans une nouvelle fenêtre)). The work has been widely picked up the scientific community and the new thermodynamic ensembles were integrated in large scale molecular dynamics engines such as LAMMPS, OPENMM,etc. With the new thermodynamic protocol, we were able to identify the external conditions leading to single-crystal-to-single-crystal transitions or transitions that are accompanied by a loss of crystallinity.
The approach has largely been applied on a broad set of materials in collaboration with experimental partners with application in sensing and actuating, development of nanodampers, shock adsorbers, thermal triggers, etc.
Very recently, we have discovered how the size of the crystal impacts the phase transformations within MOFs. Experimentally it was observed that defects and finite size of the crystal may largely impact the flexible behavior of MOFs. We have substantially pushed the limits of our force field simulations to enable simulation of very large systems where disorder may spontaneously occur during a molecular dynamics run.
Throughout the program, an important connection was made with operando spectroscopic measurements to characterize the behavior of the materials. We have developed and benchmarked various methods that allow to calculate operando spectroscopic signals from computational point of view, such as vibrational fingerprints, UV-Vis spectra, etc. As such it became possible to identify important intermediates observed experimentally and to follow chemical conversion with time on stream.
The work has been extensively disseminated via various channels. First of all the program lead to a substantial number of papers in high profile papers, which received high impact from the scientific community. Various papers were selected as Very Important Papers and have been highlighted in editorials or as covers. Second various researchers of DYNPOR have extensively communicated their results by oral contributions, invited talks on leading international conferences. We have also disseminated the results to a broader audience by launching on various occasions press releases, which were picked up by various media channels.
Our software has been made available by open access for the whole scientific community and is now used by various labs worldwide, furthermore various of our new methods were integrated in other large scale molecular dynamics engines.
Progrès au-delà de l’état des connaissances et impact potentiel prévu (y compris l’impact socio-économique et les conséquences sociétales plus larges du projet jusqu’à présent)
Before the onset of DYNPOR, most theoretical simulations of chemical reactions relied on the existence of one single transition state and reaction rates were determined in a static way by calculating a few points on the potential energy surface. Within DYNPOR, we undeniably showed that the textbook concept of a single transition state is far too simplistic for complex chemical conversions taking place in nanoporous materials. At real operating conditions, chemical reactions taking place within confined environments have a very complex nature and their outcome is critically affected by the working temperatures, assisting molecules within the pores of the material. Within DYNPOR we pioneered the simulations of complex chemical conversions at operating conditions using enhanced sampling molecular dynamics methods capturing the full complexity of the free energy surface. We discovered how the nature of elusive intermediates may be critically dependent on the conditions, how the active site may become mobile at operating conditions. A range of new methods has been developed, which are applicable for a broad set of applications and which have been made available for the scientific community. DYNPOR induced a paradigm shift from static to molecular dynamics methods for the simulations of complex chemical conversions in nanoporous materials.