Final Report Summary - SIMDEPRO (Deprotonation of organic molecules in solution by ab initio MD and rare events simulation techniques)
The scientific objective of this project is to study the dissociation reaction of weak acids in bulk and confined systems. Dissociation of weak acids is very important in chemistry, biology and environmental science. For example, the rate of the dissociation reaction of organic acids is one of the input parameters needed to develop chemical models of pollutants formation and degradation (see, for example, the SPARC project of the USA Environmental Protection Agency). These data are typically difficult to obtain from experiments. Computer simulations, based on reliable and efficient techniques, can represent a complementary source of these data. The development of these techniques was the objective of this project. We expect that, in the long run, our results will have an impact on the fields mentioned above.
The computational study of the dissociation of weak acids in bulk and confined systems is very difficult as reactive events of this kind occur with a low frequency (typically one event in 100 ns). Thus, their study by standard techniques, such as molecular dynamics or Monte Carlo, is beyond the capacity of even the most powerful current supercomputer, and of supercomputers that will emerge in a foreseeable future. During this project we developed simulation techniques that allow overcoming this problem. These techniques have been implemented in the open source package CP2K (see http://www.cp2k.org(odnośnik otworzy się w nowym oknie) online) and made available to the scientific community. We applied these techniques to several systems, hydrogen fluoride (HF) in the HF(H2O)7, and methanol (CH3OH) and formic acid (HCOOH) in bulk water. In the following we will focus on the results of the study of the dissociation of HF in the HF(H2O)7, which is completed. The study of the other two systems is under completion.
Summary of the scientific results of the study of the dissociation reaction in the HF(H2O)7 system.
- We found only two 'equilibrium' states (metastable states, i.e. minima of the free energy), the first corresponding to the undissociated HF (the reactant) and the second to the state in which fluorine ion (F-) and H ion (H+) are separated by three hydrogen (H) bonds (the product). This means that in cluster conditions the 'contact ion pair' state, i.e. the state in which F- and hydronium ion (H3O+) form an H-bond (no solvent molecules between them), is not formed or is only a 'passage' state.
- The transition state, i.e. the highest (free) energy state the system has to pass though along the reaction, corresponds to a H3O+ at a distance of two hydrogen bonds from F-.
- HF in HF(H2O)7 becomes more acid (lower pKa) at lower T. This behaviour is analogous to that of HF in bulk water. In bulk, this phenomenon is thought to be due to the lower entropy (less disorder) of the product with respect to the reactant. We computed the variation of entropy along the reaction and verified that also in the cluster case the weakness of the HF and the trend of its acidity with T has an entropic origin. In the bulk, the negative variation of entropy along the reaction is assumed to be of configurational origin, i.e. that to the reactant state can be associated more configurations than the product one. This explanation cannot be invocated for the cluster because both, the reactant and product, have only one configuration each associated. The other possible explanation is that the variation of entropy is of vibrational origin. We computed the variation of vibrational entropy from data extracted from the normal mode analysis and the finite temperature power spectrum of the system, and found a good agreement with the results of our free energy simulations.
- The mechanism of the dissociation reaction resulted to be a concerted one, with the three hydrogen atoms of the relevant H-bond chain moving together. Parrinello and co-workers (PNAS 108, 20410 (2011)) proposed an analogous concerted mechanism for the self-dissociation of water (H2O) in bulk water. They proposed that this event is triggered by a compression of the H-bond chain, i.e. a shortening of the distance between the acid and water molecules forming the chain. We tested this hypothesis and found that, indeed, a significant compression of the relevant H-bond chain is observed during the early stage of the reaction. The chain remains compressed while the H+ atoms migrate along it, and finally relaxes at the end of the reaction. However, the compression of the chain is not the result of a uniform compression of the individual H-bonds forming it. We observed that the first two H-bonds get compressed at the beginning of the reaction, while the third H-bond gets compressed more toward the end. This results in an H+ transfer that, although concerted, is not synchronous as hypothesised by the authors mentioned above.
- The rate of deprotonation of HF in the HF(H2O)7 cluster at 150 and 225 K resulted to be between one and two orders of magnitude higher than in bulk water. This implies, for example, that this reaction in the conditions found in the atmosphere is much faster than what is estimated on the basis of data on bulk systems. From the comparison between our results with those reported by Joutsuka and Ando (J. Phys. Chem. A, 115, 671 (2011)) we deduce that this difference is mostly due to a lower (free) energy barrier the system must overcome to get dissociated in the cluster case (approximately 0.28 eV in bulk versus 0.14 eV in the cluster). This is not surprising because the formation of the transition state in bulk conditions requires a significant rearrangement of the local structure around the HF molecule. On the contrary, as explained above, the transition state in the cluster does not require such a large change in the configuration of the system but just a compression of the H-bond chain.
We also used the simulation techniques developed during this project to study other subjects: hydrogenation / dehydrogenation of hydrogen storage materials, change of morphology of materials for semi-permeable membranes for fuel cells, clathrate hydrate nucleation, and wetting transitions on nanopatterned membranes. The first three topics are relevant to the field of (renewable) energy. The third is relevant to the engineering of novel self-cleaning materials, and ultra-low friction materials for nanofluidics.
Concerning the success of the training, within the 18 months of the project the fellow achieved a large autonomy in the use of the computational techniques that are at the core of this project. This is confirmed by the publication of several papers authored together and without the scientist in charge. Moreover, the fellow participated (and will keep participating) in the co-supervision of three Doctor of Philosophy (PhD) students of the group of Prof Ciccotti. Finally, the fellow gave a two-lecture course entitled 'Theory and computational techniques to study rare events in atomistic and molecular simulations' for the joint Trinity College Dublin / University College doctoral programme on 'Simulation Science'.
The computational study of the dissociation of weak acids in bulk and confined systems is very difficult as reactive events of this kind occur with a low frequency (typically one event in 100 ns). Thus, their study by standard techniques, such as molecular dynamics or Monte Carlo, is beyond the capacity of even the most powerful current supercomputer, and of supercomputers that will emerge in a foreseeable future. During this project we developed simulation techniques that allow overcoming this problem. These techniques have been implemented in the open source package CP2K (see http://www.cp2k.org(odnośnik otworzy się w nowym oknie) online) and made available to the scientific community. We applied these techniques to several systems, hydrogen fluoride (HF) in the HF(H2O)7, and methanol (CH3OH) and formic acid (HCOOH) in bulk water. In the following we will focus on the results of the study of the dissociation of HF in the HF(H2O)7, which is completed. The study of the other two systems is under completion.
Summary of the scientific results of the study of the dissociation reaction in the HF(H2O)7 system.
- We found only two 'equilibrium' states (metastable states, i.e. minima of the free energy), the first corresponding to the undissociated HF (the reactant) and the second to the state in which fluorine ion (F-) and H ion (H+) are separated by three hydrogen (H) bonds (the product). This means that in cluster conditions the 'contact ion pair' state, i.e. the state in which F- and hydronium ion (H3O+) form an H-bond (no solvent molecules between them), is not formed or is only a 'passage' state.
- The transition state, i.e. the highest (free) energy state the system has to pass though along the reaction, corresponds to a H3O+ at a distance of two hydrogen bonds from F-.
- HF in HF(H2O)7 becomes more acid (lower pKa) at lower T. This behaviour is analogous to that of HF in bulk water. In bulk, this phenomenon is thought to be due to the lower entropy (less disorder) of the product with respect to the reactant. We computed the variation of entropy along the reaction and verified that also in the cluster case the weakness of the HF and the trend of its acidity with T has an entropic origin. In the bulk, the negative variation of entropy along the reaction is assumed to be of configurational origin, i.e. that to the reactant state can be associated more configurations than the product one. This explanation cannot be invocated for the cluster because both, the reactant and product, have only one configuration each associated. The other possible explanation is that the variation of entropy is of vibrational origin. We computed the variation of vibrational entropy from data extracted from the normal mode analysis and the finite temperature power spectrum of the system, and found a good agreement with the results of our free energy simulations.
- The mechanism of the dissociation reaction resulted to be a concerted one, with the three hydrogen atoms of the relevant H-bond chain moving together. Parrinello and co-workers (PNAS 108, 20410 (2011)) proposed an analogous concerted mechanism for the self-dissociation of water (H2O) in bulk water. They proposed that this event is triggered by a compression of the H-bond chain, i.e. a shortening of the distance between the acid and water molecules forming the chain. We tested this hypothesis and found that, indeed, a significant compression of the relevant H-bond chain is observed during the early stage of the reaction. The chain remains compressed while the H+ atoms migrate along it, and finally relaxes at the end of the reaction. However, the compression of the chain is not the result of a uniform compression of the individual H-bonds forming it. We observed that the first two H-bonds get compressed at the beginning of the reaction, while the third H-bond gets compressed more toward the end. This results in an H+ transfer that, although concerted, is not synchronous as hypothesised by the authors mentioned above.
- The rate of deprotonation of HF in the HF(H2O)7 cluster at 150 and 225 K resulted to be between one and two orders of magnitude higher than in bulk water. This implies, for example, that this reaction in the conditions found in the atmosphere is much faster than what is estimated on the basis of data on bulk systems. From the comparison between our results with those reported by Joutsuka and Ando (J. Phys. Chem. A, 115, 671 (2011)) we deduce that this difference is mostly due to a lower (free) energy barrier the system must overcome to get dissociated in the cluster case (approximately 0.28 eV in bulk versus 0.14 eV in the cluster). This is not surprising because the formation of the transition state in bulk conditions requires a significant rearrangement of the local structure around the HF molecule. On the contrary, as explained above, the transition state in the cluster does not require such a large change in the configuration of the system but just a compression of the H-bond chain.
We also used the simulation techniques developed during this project to study other subjects: hydrogenation / dehydrogenation of hydrogen storage materials, change of morphology of materials for semi-permeable membranes for fuel cells, clathrate hydrate nucleation, and wetting transitions on nanopatterned membranes. The first three topics are relevant to the field of (renewable) energy. The third is relevant to the engineering of novel self-cleaning materials, and ultra-low friction materials for nanofluidics.
Concerning the success of the training, within the 18 months of the project the fellow achieved a large autonomy in the use of the computational techniques that are at the core of this project. This is confirmed by the publication of several papers authored together and without the scientist in charge. Moreover, the fellow participated (and will keep participating) in the co-supervision of three Doctor of Philosophy (PhD) students of the group of Prof Ciccotti. Finally, the fellow gave a two-lecture course entitled 'Theory and computational techniques to study rare events in atomistic and molecular simulations' for the joint Trinity College Dublin / University College doctoral programme on 'Simulation Science'.