![]() |
|
||||||||||||||||||||||||||
| General Descriptions |
||||||||||||||||||||||||||
Through molecular level computer simulation, we ˇ°measureˇ± the PVT properties, chemical potential, immiscibility, phase equilibrium, enthalpy, solvation structure, diffusion, metal-ion association and transport properties of geological fluids (H2O, CO2, CH4, N2, H2S, O2, HCl, NaCl, CaCl2, KCl, ore-forming metal ions and their mixtures) , with an ambitious plan to extend the experimental temperature and pressure range from less than 1000K and 10,000 atm to 2500 K and 100,000 atm. Most of these properties are traditionally measured by experiment, but those experiments are usually limited in a small temperature-pressure-composition (PTX) range because of experimental difficulties and expenses. So far tens of thousands of experimental data about geological fluids and minerals have been reported, but they are mostly limited below 5000 atm and 1273K. With the molecular level computer simulation, now it is possible for us to predict the above-mentioned properties up to 2500K and 100,000atm. |
||||||||||||||||||||||||||
| Research Cases |
||||||||||||||||||||||||||
(1) An optimized molecular potential for carbon dioxide have been proposed with remarkable accuracies for various properties of carbon dioxide (volumetric properties, phase equilibria, heat of vaporization, structural and dynamical properties) over wide range of temperatures and pressures. Interaction potentials between different molecules are always crucial in molecular simulations. As a matter of fact, the lack of appropriate intermolecular potential is often quoted as the most important barrier for applications of atomistic simulation methodologies to the problems of industrial and geochemical interests. It has long been recognized that the optimization of potential models is nontrivial. In this study, we proposed an optimized potential model for carbon dioxide on the basis of extensive simulations with the aid of the theorem of corresponding states, iterations and additional trials and errors. Superior to the models proposed by prior workers, this model turns out to accurately predict the PVT properties of carbon dioxide. In the mean time, the vapor-liquid equilibria of carbon dioxide have been predicted with remarkable accuracy and, in particular, the critical point is almost pinpointed by the new model. Moreover, compared with a number of neutron-diffraction and x-ray measurements, the simulated microscopic structures of carbon dioxide over wide temperatures and pressures are well reproduced. The predicted self-diffusion coefficients are also in good agreement with the nuclear-magnetic-resonance measurements. The results of this study have been published on the famous journal: Journal of Chemical Physics (2005, 122: 214507). |
||||||||||||||||||||||||||
![]() |
||||||||||||||||||||||||||
| Utilizing the established techniques of molecular dynamics and histogram reweighting grand canonical Monte Carlo simulations, this model is demonstrated to show excellent predictability for thermodynamic, transport, and liquid structural properties in a wide temperature-pressure range with remarkable accuracies. | ||||||||||||||||||||||||||
(2) The experimental phase behaviors of the systems CH4-C2H6 and CO2 have been successfully reproduced with Monte Carlo simulations. Phase separation of geological fluids under different thermodynamic conditions is no doubt one of the most important mechanisms controlling many geochemical processes. Experimental measurements provide many basic data points of phase equilibria and phase separations, which are generally discretely distributed within limited temperature and pressure range. Molecular level computer simulations find an alternative routine to generate reliable data of phase behaviors. During our efforts to simulate the phase equilibria properties with Gibbs Ensemble Monte Carlo (GEMC) and Grand Canonical Monte Carlo (GCMC) methodologies, we have experienced and solved several key problems. |
||||||||||||||||||||||||||
Firstly, the phase equilibria properties are extreme sensitive to the intermolecular potentials and the theorem of corresponding states are found to be effective for the optimization of potential parameters. Monte Carlo simulations with our elaborately calibrated potential models for CH4-C2H6 and CO2 show the significantly improved agreements with experiments as compared with prior workers. Secondly, the simulation algorithms and methodologies are crucial to efficiently generate reliable information applicable for geochemical research. It is one of our original objectives to find an appropriate simulation method for the study of phase behaviors and try to make some improvements. According to our experience, we find the popular GEMC method is inefficient for the simulations of the whole coexistence curve and is invalid with rough accuracy in the vicinity of critical point. |
Phase equilibria of CH4-C2H6 mixture under different conditions. |
|||||||||||||||||||||||||
Accordingly, we try to combine the non-Boltzmann sampling with GEMC simulations and improve the efficiency with a new technique of temperature scaling method. On the other hand, we find that the efficiency can also been substantially enhanced when we utilize histogram reweighting technique to analyze the GCMC simulation results. Further more, with the aid of recent mixed-field technique, the critical properties of the simulation systems can be accurately predicted. These results have already published on international journals: Geochimica et Cosmochimica Acta (2002, 66: 3431-3439) and Journal of Chemical Physics (2005, 122: 214507). |
||||||||||||||||||||||||||
(3) Through comprehensive isothermal-isobaric molecular dynamics simulations, the pressure-volume-temperature-component (PVTx) data of CO2-H2O and CH4-H2O have been extended beyond experimental range to about 2573.15 K and 10 GPa. CO2-H2O and CH4-H2O are the most frequently encountered fluid mixtures in and around the Earth, widely existent in various geospheres, as revealed from direct measurements or from the study of fluid inclusions in various rocks and ore deposits. However, the PVTx data of this system are still scarce for geological applications. To supplement the existing experimental database, with laborious and careful selection of the molecular interaction potentials, in this study we carried out more than one thousand molecular dynamics simulations of the CO2-H2O and CH4-H2O systems. We have carefully discussed the following problems: Firstly, we have confirmed the validity and predictability of molecular dynamics on the basis of comparisons with extensive experimental data. For pure water, we selected a potential model (SPCE) established at normal temperature and pressure and found remarkable reproducibility of molecular dynamics simulations over wide temperature and pressure ranges. Compared with the most recent published high-pressure experimental data up to 5.0 GPa, the simulation results show very good agreement with errors less than 1.0% and reveal a trend to be more accurate as the increase of pressures. For the mixture, we constructed the interactions of unlike molecules with ab initio potential surface and successfully reproduced a large number of experimental data published by many laboratories over the world in several decades. Even the data under the highest experimental temperature-pressure conditions (up to 1673 K and 1.94 GPa) are well predicted with the agreement within 1.0% in density. The high accuracy of the simulation results indicate a great predictive power of molecular level study and verify our attempt of using molecular dynamics simulations to generate data as an important supplement to the database of thermodynamic properties of geological fluids. Secondly, we discussed the effects of unlike interaction parameters from different approaches. In this study, we first tried two conventional combining rules and found that they are not able to accurately predict the non-ideal behavior of the mixture. We took an alternative approach to parameterize the cross mixing potential through non-linear fitting to the ab initio potential surface and proved that it substantially improves the prediction for the non-ideal mixing property. Thirdly, a new supercritical equation of state (EOS) has been established for the CO2-H2O and CH4-H2O systems. In order to integrate all the experimental and simulated data covering a wide temperature and pressure range with experimental accuracy for geochemical applications, we developed an EOS for the CO2-H2O and CH4-H2O systems valid from 673.15 K to 2573.15 K and from 0 to 10.0 GPa with errors less than 2% in density. This EOS should represent the most accurate model in a very wide temperature and pressure range. From this EOS, volumetric properties (density, volumes and excess volumes), heat properties (enthalpy), and chemical properties (fugacity, activity, chemical potential and possibly supercritical phase separation) can be derived. The results of this study have been separately published on international SCI journals: Physics of Earth Planetary Interiors (2005, 149: 335-354) and Geochimica et Cosmochimica Acta (2006, 70 (9): 2311-2324; 2007, 71 (8): 2036-2055). In particular, Dr. Wesolowski (associate editor of Geochimica et Cosmochimica Acta) and three reviewers gave positive and enthusiastic comments. They consider these results as ˇ°a triumph for computational geochemistryˇ± and ˇ°a significant contribution to Earth Scienceˇ±. |
||||||||||||||||||||||||||
![]() |
||||||||||||||||||||||||||
| Molecular dynamics simulations of the PVTx properties of CO2-H2O system. Comparisons with experiments available show the remarkable predictabilities of our simulations. | ||||||||||||||||||||||||||
(4) With molecular dynamics simulations, various properties of ionic solvations and associations have been comprehensively analyzed. These results not only agree well with experimental data and first principle calculation results, but also provide a promising approach to study the transports and deposits of ore-forming elements in our further researches. Ionic solvation, which is related with the destruction of hydrogen bond and reconstruction of new structures in the vicinity of ions, is a fundamental property of aqueous electrolyte solutions. With the ion¨Cwater potential parameters evaluated from the data of the clusters and the water¨Cwater potential predetermined from the non-rigid RWK2 model, the structural (radial distribution functions, angular distribution functions, spatial distribution functions, coordination number), dynamical (residence time) and energetic properties of the ionic solvations in bulk water were studied through a comprehensive analysis of our MD simulation outputs. These results not only agree well with experimental data and first principles calculations, but also reveal some new insights into the microscopic ionic solvation processes. Constrained molecular dynamics simulations were carried out to investigate the ionic associations in dilute aqueous solutions over a wide temperature range. Solvent mediated potentials of mean force have been carefully calculated at different thermodynamic conditions. Two intermediate states of ionic association can be well identified with an energy barrier from the oscillatory free energy profile. Clear pictures for the microscopic association structures are presented with a remarkable feature of strong hydration effect of lithium ion and the bridging role of its hydrating complex. Experimental association constants have been reasonably reproduced and a general trend of the increasing ionic association at high temperatures and low densities was observed. This study provides some complementary results for experiment and may be helpful for theoretical and semi-empirical treatment of the aqueous solutions. Moreover, this study provides a promising approach to study the transports and deposits of ore-forming elements (such as Zn2+ˇ˘Cu2+ˇ˘Fe2+) in our further researches. These results have been published on international SCI journals: Molecular Physics (2003, 101: 1501-1510) and Chemical Physics (2004, 297: 221-233). |
||||||||||||||||||||||||||
![]() |
||||||||||||||||||||||||||
| Potential of mean force (pmf) and association structures of lithium chloride in dilute aqueous solutions. The structures are revealed from the oxygen (solid) and hydrogen (dash) density contours and typical snapshots (top right window) at different ionic separations. | ||||||||||||||||||||||||||
| Related Programs |
||||||||||||||||||||||||||
|
||||||||||||||||||||||||||
Zhenhao Duan Research Group |
||||||||||||||||||||||||||