We could proceed with 100K for Q4 Classification: P Chem Wagner ----------------------------------------- Ray Bair Computing, Environment, and Life Sciences Argonne National Laboratory and the University of Chicago TCS Building 240, Room 4122 9700 South Cass Avenue Argonne, IL 60439 email: rbair(at)anl.gov Phone: (630)252-5751 On 9/11/15, 4:13 PM, "[email protected] on behalf of [email protected]" <[email protected] on behalf of [email protected]> wrote:

Hello,

A new project on the LCRC cluster has been requested. Please forward the information on to the LCRC Allocation sub-committee.

Applicant's name: Baofu Qiao Applicant's institution: ANL Applicant's division: CSE Project Name: Lanthanide_FF Project title: Force Field Development and Transferability for Metalloamphiphile-in-Oil Solutions Associated funding: DOE-BES Other Systems: The PI is currently a user on LCRC since 2014, where around 1 million of CPU hours have been used. Additionally, the high-performance cluster TARDIS at Prof. Monica Olvera de la Cruz¹s group at Northwestern University is also available. The PI has used some millions of CPU hours on TARDIS. Science: Water-based metal ion separations, known as hydrometallurgy, underpin a broad range of industrial processes from nuclear fuel reprocessing to strategic metals refining, and to water treatment. Some examples are the removal of toxic heavy metals, e.g., cadmium and lead, from industrial waste water, the production of rare earth elements, to name just two. These processes generally involve the transfer of metal ions from aqueous phase to organic phase using a phase transfer agent (e.g., extractant). At the Heavy Element Separation and Science (HESS) group at ANL, we are focusing on the solvent extraction process for the refinery of f-block metals. Such solvent extraction is a necessary post-process treatment of nuclear waste water from the economic, environmental and energy point of view. The fundamental understanding of the solvent extraction of f-block metal ions from aqueous phase to organic phase is a pre-requisite to improve the industrial processes, which has a long history even since World War II and is still very limited understood. A variety of experimental technologies have been developed and/or employed for the structures particularly in the organic phase (after extraction). Such experiments include, but not limited to, small-angle X-ray scattering (SAXS) and extended X-ray absorption fine structure (EXAFS), both of which have been extensively used in the APS at Argonne. However, all the existing experimental technologies have some advantages and disadvantaged . For SAXS, it provides the real experimental data regarding the mesoscale morphology of micelles. Nevertheless, the resolution of SAXS is at the nm length scale, which is not high enough for the supramecular structures in organic phase that is also at the nm length scale. In terms of EXAFS, it reflects the atomic scale coordination behavior of metal ions, but the unique interpretation of the EXAFS data doesn¹t exist in the complicated metalloamphiphile-in-oil systems. Thanks to the fast development of high-performance computing clusters, the recent years have witnessed the rapid growth in computer simulations to reveal the atomistic structures in the metalloamphiphile-in-oil systems. In particular, the PI has shown the atomistic molecular dynamics (MD) simulations are capable of semi-quantitatively reproducing experimental SAXS and EXAFS data. [(a) Qiao, et al., J. Phys. Chem. Lett., 2014, 5, 1440. (b) Qiao, et al., ACS Central Science, 2015, submitted. (c) Ferru, et al., 2015, to be submitted.] However, due to the limited existing force field parameters, the capability of atomistic MD simulations is still hampered. In the present proposal, the PI is planning to systematically develop force field parameters for lanthanide and actinide metals, as well as necessary ligands and some small aliphatic molecules (solvent in the organic phase). In addition to SAXS and EXAFS, the high-energy X-ray scattering (HEXS) will be employed to calibrate the newly developed force field parameters. HEXS technology has the unique advantage that it is capable of unveiling the atomic structures at the angstrom resolution. Historically, HEXS technology has been extensively employed in the aqueous phase for heavy metal­bearing structures. The experimental experts at HESS group at ANL are currently extensively developing the HEXS technology in the organic phase. The complementary of atomistic MD simulations and experiment technologies (HEXS, SAXS, EXAFS, etc.) will surely provide a complete understanding for the structures in both the aqueous phase and the organic phase for heavy metal solvent extraction. Subse quently, the atomistic simulations will be employed to unveil the basic knowledge of the driving force of these structures, and provide predictive insights for the optimal design of elevated solvent extraction process in the near future. Project description: In this project, we are aiming to: 1. Develop the polarizable and non-polarizable force field parameters for lanthanide and actinide metals, as well as necessary ligands (nitrate, amphiphilic extractants, etc) and aliphatic solvent molecules. Based on the existing literature and the experience of the PI, the polarizable atomistic force fields, are very slow, even though they are ³expected² to provide better agreement with experiments. Therefore, only small simulation box (< 10 nm) and short simulation time (around 10 ns) were used for polarizable atomistic simulations on metalloamphiphile-in-oil solutions. Nevertheless, longer time scale and larger length scale, which are affordable using the non-polarizable simulations, are expected for comparison with experimental SAXS, etc. We are thus planning to quantitatively compare the effects of the polarizability (polarizable vs. non-polarizable) on the agreements with experimental data. This will surely guide the further development and applications of atomistic force fields not just for us, but for the whole community of simulation scientists. 2. Investigate the transferability of force fields, which are originally developed in aqueous phase (or vacuum), from aqueous phase to organic phase. Ever since the simulation approaches were proposed several decades ago, aqueous solutions were extensively investigated, leaving alone the organic phase. Recently, it has been shown that organic phases are also becoming the focus of a variety of important technologies, not only for meal solvent extraction investigated here, but also for protein retention (see. e.g., Panganiban, et al., 2015, to be submitted.), etc. It has been well established that the relative permittivity of aqueous solutions (~ 80) is quite different from that of organic solutions (generally < 10, depending on the solvent type). Therefore, one question is: Are the force field originally developed in a high dielectric media (aqueous phase) applicable to a much lower dielectric media (organic phase)? This question becomes extremely important especially when th e small reverse micelles with the dimeter of around 1 nm dominates in the organic phase, which are composed of polar cores containing metal, anions (nitrate, etc.), amphiphile headgroups, embedded in nonpolar environment, forming nm-size (hydrophilic)core­(hydrophobic)shell morphology.

In this project, density functional theory (DFT) and atomistic MD simulations will be combined. To develop the (polarizable or non-polarizable) force fields, the DFT calculations will be performed to get the necessary chemical bond parameters, as well as the atomic partial charges for the Coulomb interactions. The existing Lennard-Jones (LJ) parameters can be reasonably used due to the relatively weaker contribution of LJ interactions as respect to the Coulomb interactions. However the LJ parameters of most of the lanthanide and actinide metals are missing, which need to be parameterized in aqueous solutions in the next step. Given the obtained force field parameters and assumed LJ parameters of metals based on existing related literature, (polarizable and non-polarizable) atomistic simulations will be performed on aqueous solutions and compared to HEXS data. The LJ parameters of metals will be systematically adjusted till the best agreement with experimental HEXS data is achieved. It is noteworthy that such a force field parameterization methodology has been employed in our recent work (Ferru. 2015, to be submitted.), where the polarizable force field parameters of Ce3+ and Yb3+ have been developed based on the existing EXAFS data and previous simulations from other groups. In the present project, the HEXS technology will be employed in aqueous solutions instead to calibrate the force field parameters, which can provide a much higher resolution on the coordination behavior of metal ions than EXAFS. The force field parameters developed in vacuum and aqueous solutions will be transferred into organic phases. Atomistic simulations using polarizable and non-polarizable force fields will be performed, and compared to experimental HEXS, SAXS and EXAFS in corresponding organic solutions. The performance of both force fields will be consequently compared in reproducing as many as possible experimental data. Gaussian or NWchem packages will be used for the DFT calculations. AMBER and GROMACS will be used for the atomistic simulations using polarizable force field and non-polarizable force field, respectively. AMBER supports the polarizable force field using induced dipoles. GROMACS, though doesn¹t support such a polarizable field field, provides a much higher calculation speed than AMBER (~ 5 times faster than AMBER using the same non-polarizable force field, based on the experience of the PI). The packages are open source, or ANL has purchased the required licenses for the commercial software. The simulation box edge length will be around 8-20 nm in each dimension, containing around 50,000 ­ 500,000 atoms. Smaller simulation boxes are good enough for the comparisons in terms of the local coordination behaviors from HEXA and EXAFS, whereas large enough simulation boxes are required for the comparison with SAXS data. The simulation duration will probably vary from tens of ns up to some hundreds of ns, depending on how fast the systems will converge in the simulations, and how many simulation snapshot are necessary for good data analysis. According to the experience of PI, AMBER and GROMACS can be well scaled up to at least 96 CPU cores on LCRC per job containing around 50,000 atoms. GROMACS has shown good scalability for around 300 CPU cores on TARDIS cluster (Northwestern University) for solutions containing around 300,000 atoms. Some of the experimental experts at the HESS group at ANL will be involved in the related experiments, e.g., Dr. Lynne Soderholm, Dr. Ross Ellis, Dr. Geoferroy Ferru, Dr. Suntharalingam Skanthakumar. Dr. Baofu Qiao will perform the related calculated and simulations, and will be the PI of this project and the only member of this project on LCRC. Industry partnership: Project URL: Requested allocation: 200000 Q1: 0 Q2: 0 Q3: 0 Q4: 200000 Justification: Storage requirements:

The requester has used undetermined amount hours of their initial startup project.

In addition to approving an initial amount, please specify a Category and Subcategory for this project. For a list of the current selection of approved categories, please see:

https://wiki.lcrc.anl.gov/wiki/Processes/Categories

Once the Allocation committee has approved the project, please go to the Project Management page to create it:

https://accounts.lcrc.anl.gov/projects.php

Thank You, The LCRC Accounts System