Simulation-based engineering and science

This is a five-year program funded by CREST-JST at the level of ¥190 M and it is a joint effort between UT, Tokyo University of Science, Nihon University, and Kobe University. The main objective is to develop quantum simulators to understand nanoscale electron transport for helping accurate interpretation of nanoscale measurements. These simulators include single- and multiple-probe measurements, capacitance evaluation, and analysis on effects from other external factors. A grand challenge remains the simulation of non-steady dynamics of atoms and electrons in non-equilibrium open systems. The computational resources employed are both PC clusters as well as supercomputers at UT.

The focus of this presentation is on multiscale modeling. A new method for molecular collisions and evaluation of physical properties (e.g., viscosity, conductivity) was presented that takes into account rotational energy. The results shown suggested great improvement in conductivity predictions. In addition, topics on multiscale modeling of surface reactions, bubble-bubble interactions, and nano-bubbles were discussed. Finally, the human simulator was described as a joint project with RIKEN.

Various applications of single- and multi-walled carbon nanotubes (CNT) were discussed, including applications at Toyota. Results with 2 nm standing CNT were presented, apparently the first such achievement. The objective of this project is to understand metal-carbon interaction inside a CNT using MD. Issues associated with MD acceleration via molecule positioning were discussed. There is a close MD-experiment interaction. A typical MD simulation on a small (5-PC) cluster may take up to six months.

Prof. Kasagi is a leader of the 21st COE Program. He first presented a comprehensive view of Japan’s strategic basic plan on science & technology, and the goal of UT programs’ on mechanical systems innovation and its history. The third period of the national basic plan is underway (2006–2010), funded at the level of ¥25 trillion. It supports research with emphasis on eight fields such as life sciences, information sciences, nanosciences and the environment, and the education of young scientists. It also supports two key technological areas, namely the next-generation supercomputer at RIKEN and the space transportation system at JAXA. Prof. Kasagi then proceeded to describe his own work, which comprises two main areas: energy systems and shear flow control. Interesting concepts related to a microturbine 30 kW energy generation system combined with a solid oxide fuel cell were presented, and the power supply needs for mobile systems were addressed. On the shear flow control, direct numerical simulations at the highest currently possible Reynolds number were presented, and MEMS-based techniques were shown to effectively control near-wall turbulence.

The main objective of this project is full-scale simulation of components but also of buildings subject to earthquakes. Most of the simulations presented involved one-way coupling on flow-structure interactions. A fully coupled simulation will require access to the next-generation supercomputer system at RIKEN.

The purpose of this project is the development of a simulation capability for disaster propagation in chemical plants considering the effects of tank fire and wind speed and direction. Another key question is how to estimate associated damages using the results of simulations in conjunction with information from the Geographical Information System (GIS). Monte Carlo methods of estimating radiation heat transfer were presented, and examples from tank fires and gas explosions in realistic scenarios were discussed. A validation example was presented, but it is clear that long-term prediction of such events remains a very difficult task. This project is well funded by a MEXT special project on earthquake disaster mitigation in urban areas, by the Petroleum Association of Japan, and by industry.

The purpose of this project is simulation of tsunamis and also the estimation of damages caused, based on simulation results and survey data. Advanced methods in solving the shallow water equations were developed for realistic terrains, and a “fragility function” that estimates the probability of damage was constructed for first time for Banda Aceh.

The objective of this project is to develop middleware for the main generic operations in finite elements, including global assembly and solvers, for a JST-CREST (2006-2011) project focusing on predictive capability for earthquakes and tsunamis. The usual parallel libraries such as MPI and OpenMP are hidden from the user, and optimization is pursued both on scalar and vector architectures. Very good performance results were achieved on the Earth Simulator for the structural analysis code, but the viscous flow codes does not scale as well. The software analyzer of the structure code (FrontSTR) is selected as a target software package to be optimized at RIKEN’s next-generation supercomputer.

The purpose of the “Revolutionary Simulation Software (RSS)” project is to develop Japanese-made software for life sciences, nanosciences, engineering, and urban environments. The project spans the period 2005-2007, it is supported at ¥1.2 billion per year, and it involves 120 researchers. The software developed is free to be used in Japan or abroad, and to the date of this writing, 40,000 downloads have been reported. Continuous improvement and maintenance of software will be accomplished through commercialization; currently, seven companies, including Advancesoft Corporation, Mizuho Information & Research Institute, Inc., and NEC Soft, Ltd., are pursuing such commercialization.

Research at the University of Tokyo (UT) is of very high quality. The 21st century COE Program has provided the resources to develop new simulation capabilities in energy systems, disaster modeling and damage evaluation, life sciences, nanotechnology, and urban environments. The Institute of Industrial Science at UT is supported generously by MEXT to work closely with researchers conducting fundamental research in order to develop free software with the potential of commercialization by the private sector. Initial results of this collaboration scheme are very encouraging, and the simulation capability developed in diverse areas is very impressive. Some of the researchers reported isolated efforts in verification and validation, but no separate funding and hence no systematic effort is underway to address this issue.

Site: Autonomous University of Barcelona and

Edifici CM7; Facultat de Ciencies

Email: ordejon@icmab.es

Dr. Alberto García

Email: albertog@icmab.es

Dr. Eduardo Hernández

Email: ehe@icmab.es

Background

The Research Center for Nanoscience and Nanotechnology (Centre d'investigació en nanociència i nanotecnologia), also known as CIN2, at the Autonomous University of Barcelona (Universitat Autònoma de Barcelona) is a mixed research institute that has as partners the nationally funded Spanish National Research Council (Consejo Superior de Investigaciones Científicas; CSIC; http://www.csic.es) and the local (Barcelona) Catalan Institute of Nanotechnology (Institut Català de Nanotecnologia; ICN; http://www.nanocat.org).

CSIC is Spain’s largest national institute for research, comprising about 100 institutes between the sciences and the arts. The central government provides money for the basic functions of these institutes, but research money must come from external funding. CSIC allocates budgets for a four-year period. People hired under CSIC are “public servants” with tenure for life, although there is talk now to substitute this mode for long-term positions followed by evaluation.

The legal structure of the ICN, on the other hand, is that of a private foundation with public capital. ICN was founded by the Catalan government in 2003 and has two partners, the Ministry of Universities, Research, and Information Society of the Catalan Government, and the Autonomous University of Barcelona. ICN has the flexibility to hire personnel as needed, and institute management decides the salaries and contracts.

From the point of view of scientific research policy, the CIN2, launched in September 2007, is a relevant example of collaboration between Catalonia and the central government of Spain. Its formal and legal structure allows exploiting the synergies between both partners (CSIC and ICN), following the principles of scientific excellence that are key to the CIN2's foundational deed. CIN2 combines the best of both worlds: the stability of the central government with the funding and flexibility of ICN.

Research

CIN2 activities are focused on the theory and simulation of processes at the nanoscale. They aim at understanding the basic physics behind the behavior of nanoscale systems at the atomistic level. The techniques focus on the electronic and atomic structure and dynamics, and their evolution in response to external conditions. CIN2 activities include the development of methods and simulation tools for the description of these systems, from semiempirical to fully first-principles approaches. Current interests include, among other topics, the understanding of electronic transport at the nanoscale (including inelastic effects and strong electronic correlations), the simulation of STM images and STS spectra, the interaction of molecules with surfaces, and different aspects of nanostructures such as nanoclusters and nanotubes.

The CIN2 group is also involved in the SIESTA project. This was developed as an approach to compute the electronic properties and perform atomistic simulations of complex materials from first principles. Very large systems, with an unprecedented number of atoms, can be studied while keeping the computational cost at a reasonable level. The SIESTA code is freely available for the academic community (http://www.uam.es/siesta), and this has made it a widely used tool for the study of materials. It has been applied to a large variety of systems, including surfaces, adsorbates, nanotubes, nanoclusters, biological molecules, amorphous semiconductors, ferroelectric films, low-dimensional metals, etc. There are strong interactions with the Electronic Structure Lab at the Materials Science Institute of Barcelona (Institut de Ciència de Materials de Barcelona; ICMAB-CSIC; Prof. Canadell et al.).

A more “itemized” research list includes:

• 
  Methods and codes for atomic scale simulations: SIESTA
  

• 
  Electronic structure methods & molecular dynamics
  
• 
  Quantum electronic transport (TranSIESTA)
  

• 
  Electronic transport
  

• 
  Fundamental problems (NDR, Inelastic transport—weak and strong coupling, Kondo, ferroelectric tunnel junctions, and so on)
  
• 
  Applications (STM, molecular electronics, nanowires, surfaces, and so on)
  

• 
  Molecules on surfaces
  

• 
  Molecule-molecule and molecule-substrate interactions
  
• 
  Mainly metallic surfaces
  

• 
  Simulation of electronic dynamics processes in dye-sensitized thin-film solar cells
  

• 
  Electron excitations in dye molecule
  
• 
  Electron transfer processes
  

• 
  Ion dynamics in oxides
  

• 
  Oxygen/vacancy mobility in nanostructured functional oxides
  
• 
  Electrochemical devices, fuel cells, gas sensors
  

At present, the group has funding from the following sources:

• 
  The coordinated Spanish Ministry of Education and Science (MEC) project (2006–2009) provides a total of €532,000, with collaborators in the Autonomous University of Madrid, the University of Oviedo, and Materials Science Institute of Madrid (ICMM-CSIC). The topic is “Quantum-mechanical simulations and scanning probe microscopy in current problems in surfaces, complex materials, biomolecules and nanostructures.”
  
• 
  A “CONSOLIDER project” (support for supercomputing and e-science) has a budget of €5 million for 5 years and involves 23 groups around the country. The title of the project is “Strategic research action to advance in the development of supercomputing environments and their application to grand challenges projects in the areas of Life Sciences, Earth Sciences, Astrophysics, Engineering and Materials/ Condensed Matter Physics.” CIN2’s task is to address some of the challenges in Materials Sciences and Condensed Matter Physics and to carry out massive parallelization of SIESTA. Most of the money goes to post-docs’ salaries.
  

The Siesta program has the following key characteristics:

• 
  Density Functional Theory (DFT) code
  
• 
  logorithmically designed for efficiency, scales linearly with system size
  
• 
  The code has become standard in the Quantum community and competes with other codes (VASP, CASTEP, CPMD, Gaussian, etc.)
  
• 
  A large development team from several institutions (8 researchers just in the restricted “code team”)
  
• 
  Very long-term project with several sources of funding
  
• 
  Freely available for the academic community (~3500 users worldwide)
  
• 
  Nonacademic license available for a fee (Motorola, Sumitomo Chem., Samsung, Air Products, Fujitsu, and others)
  
• 
  Currently 300+ publications referencing the code per year
  
• 
  Currently SIESTA has excellent performance in serial machines but poor parallel scaling
  
• 
  Current developments
  

• 
  Optimization and parallelization, in collaboration with the Barcelona Supercomputing Center
  
• 
  Implementation of new tools and methods (hybrid QM/MM, electronic transport with inelastic effects, free-energy issues)
  

The group investigating the electronic structure of materials works in close collaboration with CIN2. The main lines of research are organic conductors; low-dimensional metals; electronic transport; fullerenes, nanotubes, and other “nano”; magnetoelectrics; multiferroics; computational crystallography; materials for hydrogen storage; phase behavior and phase diagrams; TD-DFT; and novel simulation techniques.

The research tools of the group include expertise in both first-principles and semi-empirical electronic structure; empirical force fields, molecular dynamics, Monte Carlo, and structural relaxation.

The main sources of funding are the Spanish Ministry of Science and Education; the Catalan regional government, and several EU funding programs.

• 
  Barcelona Supercomputing Center: 10240 processors, 94 TFl; 3rd largest in Europe, 13th largest in the world; funded by the Spanish central government, the Catalan government, and through special pricing from IBM
  
• 
  Diverse clusters are shared by the group
  

• 
  There have been efforts to coordinate things with other EU groups in order to consolidate a European “scientific power.”
  
• 
  Psi-K network: EU funded; this was very important in getting groups together for collaboration and dissemination of information; however, the organization was not interested in continuation of projects, only in new projects
  
• 
  Psi-K network will transition into a nonprofit private entity with funding from research groups that use the tools (http://psi-k.dl.ac.uk)
  
• 
  The European Centre for Atomic and Molecular Computations, CECAM, has played a very important role in disseminating information; there is an agreement with Psi K to put new effort into code development (http://www.cecam.fr)
  
• 
  Research funding in Spain is increasing; EU funding has had a significant impact (FAME; Center for Multi-Functional Materials)
  

• 
  These groups and institutes are not academic; they carry out research but not teaching
  
• 
  They get only PhD students and postdocs; there is no formal training—the students learn through the assignments
  
• 
  Students do research in CSIC, but the degrees are granted through universities
  
• 
  PhD students generally take 4 years (there is no funding after 5 years); 1 or 2 years are devoted to courses for credit; there is an initiative to put everything together and give the material in two months (very structured and intensive courses)
  
• 
  There is no graduate degree for computer simulation; there is something on informatics, and there is a MSc on Computational Science at the Politechnic University of Catalonia
  
• 
  There are no career opportunities in Spain for computational support staff; the WTEC team’s hosts hope that this will happen in the future following the establishment of the CSIC
  
• 
  International students: most of the students and post-docs are from outside Spain—several from Latin America; the main problem for international students is bureaucratic barriers for visas and working permits for foreigners outside the European Union.
  

• 
  Very little industrial investment in research in Spain
  
• 
  Some pharmaceutical research
  
• 
  RAPSOL
  
• 
  Carburos Metálicos: bought by the United States, this company has built a center for research in materials in CSIC and wants to establish a collaboration on materials aspects of gas storage; MATGAS (materials for gases) will have as main activities research projects, a big chunk carried out via simulations (the United States pays for the research).
  
• 
  
  

Distillation, Reactive Distillation, Adsorption, Ion Exchange

Tel: 49 621 60-56440; Email: christoph.grossmann@basf.com

Dr. Ekaterina Helwig, Science Relations and Innovation Management

Tel: 49 621 60-92892; Email: ekaterina.helwig@basf.com

Dr. Horst Weiss, Senior Scientist

Polymer Physics, Molecular Modeling
Tel: 49 621 60-22445; Email: horst.weiss@basf.com

Dr. Ansgar Schafer, Team Leader, Quantum Chemistry

Physical Chemistry and Informatics, Scientific Computing

Tel: 49 621 60-78376; Email: ansgar.schaefer@basf.com

Dr. Anna Schreieck, Head of Scientific Computing

Tel: 49 621 60-78253; Email: anna.schreieck@basf.com

Dr. Michael Schafer, Research Manager

Particle Measing Technology, Process Engineering

Tel: 49 621 60-79264; Email: michael.schaefer@basf.com

Dr. Lars Vicum, Research Manager

Precipitation, Process Engineering
Tel: 49 621 60-58308; Email: lars.vicum@basf.com

Dr. Markus Linsenbuhler, Research Manager

Particle Separation and Aerosol Technology, Process Engineering
Tel: 49 621 60-97668; Email: markus.linsenbuehler@basf.com

Dr. –Ing. Jan-Martin Loning, Senior Research Manager

Adsorption, Extraction, Physical Properties, Process Engineering

Tel: 49 621 60-56543; Email: jan-martin.loening@basf.com

Dr. Wolfgang Gerlinger, Research Engineer

Fluid Dynamics, Process Engineering

Tel: 49 621 60-91721; Email: wolfgang.gerlinger@basf.com

Dr. Jobst Rudiger von Watzdorf, Head of Research Group

Chemicals Research and Engineering, Reaction Engineering

Tel: 49 621 60-54578; Email: jobst-ruediger.watzdorf-von@basf.com

BASF is the world’s leading chemical company. The company employees over 95,000 total people worldwide, with more than 8,600 engaged in research and development activities. In 2007, BASF posted sales of €58.0 billion and income of approximately €7.6 billion. BASF’s product portfolio comprises of five business segments:

23. 
    Chemicals: This segment consists of Inorganics, Catalysts, Petrochemicals, and Intermediates divisions.
    
24. 
    Plastics: BASF is one of the world’s leading producers of plastics. In 2007, the segment was organized in three divisions: Styrenics, Performance Polymers, and Polyurethanes.
    
25. 
    Performance Products: This segment is made up of the Construction Chemicals, Coatings, Functional Polymers, and Performance Chemicals divisions.
    
26. 
    Agricultural Products and Nutrition: This segment comprises the Agricultural Products and Fine Chemicals divisions. BASF Plant Science conducts research in the field of plant biotechnology.
    
27. 
    Crude Oil and Natural Gas: BASF’s oil and gas activities are pooled in the Wintershall Group, which is active in the exploration and production of crude oil and natural gas, as well as in the trading, transport, and storage of natural gas sectors.
    

Simulation and modeling tools are broadly used in most of BASF’s business segments. A few examples of research and development projects using simulation and computational models are listed below.

• 
  BASF's computational chemists are studying three-dimensional structures of proteins, protein-ligand complexes, specialty chemicals, and how these structures interact on a molecular level. Currently, many processes, especially in molecular biology, can be visualized in a highly simplified manner using a lock andkey analogy. Knowing the three-dimensional structure of the molecules involved is a basis for using a substitute key to prevent the real key from fitting into the lock, finding a better key, or changing the lock so that a completely different key will fit.
  
• 
  BASF scientists are using virtual screening by computer to test thousands of candidate compounds and determine whether their three-dimensional and electronic properties would allow them to interact favorably with target proteins. Hands-on experiments need only be performed on substances that pass this screening process. De novo design takes things one step further. The computer screens all synthetically accessible building blocks and selects those that best interact with the binding site. 
  
• 
  Biocatalytic processes involve enzymes that catalyze chemical reactions of the substrate. Despite knowing the ligands and enzymes involved, scientists still want to know the mechanism underlying the reaction and how the enzyme catalyzes the chemical reaction efficiently at room or body temperature. BSAF researchers are using quantum mechanical computations to follow biochemical reactions and thus provide insights into reaction kinetics and thermodynamics. The information obtained from these calculations can help scientists to optimize fermentative production processes and develop more efficient enzymes.
  
• 
  BASF researchers use bioinformatics in all activities performed at a molecular biological level. This is the case in the areas of biocatalysis, fine chemicals, and agrochemicals. The scientists have in-house access to all public molecular biology databases and internal BASF databases to assist them in their work. Collaboration with leading bioinformatics firms is another important element of this work.
  
• 
  Agrochemicals used to protect crops from weeds, fungi, and insects work by blocking specific metabolic processes, thus maximizing efficacy and minimizing chemical use. The molecular sites of action of crop protection substances in the organisms to be combated—the targets—are proteins that are essential to the survival and development of fungi and weeds. The search for new active substances is more effective if the targets have been identified; hence, researchers concentrate their efforts on tracking down the relevant proteins in order to inactivate them with the help of tailored active substances. Researchers who want to identify the key proteins in an organism must first elucidate their functions, at least in the vital segments of the metabolic process. BASF researchers are using Functional Genomics strategy, where they draw upon the knowledge of the genome of the living organism under study. Analysis of sequencing data and comparison with the genetic information of other organisms enables scientists to predict or establish hypotheses on parts of the metabolism and the role of specific proteins. Functional characterization of these proteins is supported by bioinformatics methods.
  
• 
  Vitamins and amino acids can often be produced more efficiently using microorganisms than by conventional synthesis methods. For example, BASF researchers have used a bacterium called Corynebacterium glutamicum to produce the amino acid L-lysine by fermentation. Their goal is to alter the metabolism of the bacterium so as to optimize the production of lysine. Bioinformatics played a key role in genome sequencing and analysis and in elucidating the metabolism of the bacterium. Out of the 3300 genes identified, the scientists involved in the project have established the function of 2100 previously unknown genes. They used bioinformatics to analyze the data on already investigated genes from other microorganisms with similar gene sequences. A factor complicating analysis of the Corynebacterium genome was the presence of a whole series of genes whose function could not directly be ascertained because there were no known genes with a similar sequence. The researchers analyzed approximately 100 other genomes to compare the position of the genes in relation to each other and to compare the number of genes that corresponded to the enzymes of a metabolism segment. This method unearthed a promising gene that is thought to code for a regulator of lysine biosynthesis.