Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Anh Ngo Project Name: Mullite Division: MSD Project title: The structure of disordered Mullite studied using a ab initio hybrid reversed Monte Carlo algorithm Associated funding: LDRD Other Systems: Science: Mullite is a material with an important role in the technology of classical and advanced ceramics for both traditional and advanced applications. The material is characterized by a very high thermal stability especially under oxidizing conditions, outstanding chemical stability in severe environments. Consequently, they are used in a great number of applications. The mullite structure can be written as Al2[Al2x+2Si2-2x]O10-x where x denotes the number of oxygen atoms missing per average unit cell and generally ranging between 0.2 and 0.5, x = 0.25 corresponds to 3:2 mullite, and x = 0.4 corresponds to 2:1 mullite. In mullite phases vacancies are introducing into the structures by replacing Si with Al while maintaining the overall charge neutrality . With higher alumina content, increasing x, the complexity of the structures increases with an increased number of O vacancies. The mullite structure becomes more and more disordered. Project description: Understanding the nature of the disordered structures and physical/chemical properties of mullite is therefore an area of increasing importance and has been the focus of investigation for many decades. In this project we would like to give a complete description of the real mullite. We will use ab initio energetic cluster expansion techniques to explore all distinct interactions/mechanisms between/of Al, Si, Oxygen and vacancy in the mullite. We then integrate these effective interaction parameters to build a fully microscopic model with our new developed ab initio hybrid reversed Monte Carlo. The project will be realized by collaboration between Anh Ngo (MSD, argonne) and Peter Zapol (MSD, argonne) and Ray Osborn (MSD, Argonne). Our results obtained from ab initio hybrid reversed Monte Carlo can be directly compared to diffuse scattering data measured at APS from Osborn’s group. To obtain our goal the following calculations will be performed: (a) The initial calculations will focus on finding the ab initio cluster expansion energies (CE) for all pair, triplet and higher order interactions. (b) To find real disordered structures, in our further investigations we will use CE obtained from step 1 to perform ab initio hybrid reversed Monte Carlo. This work will plan to use the high-performance computing facilities at LCRC. The first-principles calculations will be carried out using the electronic-structure code VASP. The Blues cluster is particularly suitable for this research. In order to obtain good Cluster expansion nergies we will plan to do 200 energy optimized calculations each calculation contains 60 atoms takes approximately 3072 CPU hours (16*4*48 ). The ab initio hybrid reversed Monte Carlo calculation takes roughly 26880 CPU hours (16*10*168) we plan to run 8 of these calculations. So our total request is 200*3072 + 8*26880 = 829440. In summary, the total requested allocations for this project is 829440 CPU hours. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 930000 Q1: 250000 Q2: 250000 Q3: 250000 Q4: 180000 Justification: VASP performance on Blues is well established, we use the latest version maintained on Blues. Storage requirements: 1 TB is sufficient Thank You, The LCRC Accounts System