The EIRENE.DE Online Database for Surface and A+M Data in Fusion Edge Modelling

In magnetic fusion plasma modelling no unique optimal data format for atomic, molecular and surface data exists, because the level of detail (condensation, pre-processing) of raw data is highly case and application dependent. Microscopic erosion redeposition modelling requires, in principle, multidifferential cross sections, whereas macroscopic core plasma simulations can be based on strongly reduced (compressed and processed) A and M data, even charge stage bundling might be an option here. Integrated edge plasma modelling (e.g for ITER divertor design studies) has both important microscopic and macroscopic components (and a related wide spectrum of requirements for AMS data formats). The B2-EIRENE code (aka: SOLPSx.y) is a prominent example of such an integrated edge plasma code. The AMS database maintained at FZ Juelich (www.eirene.de) is set up to publicly expose the database available for this codes system as used in the current ITER design. It is a database tailor-made for this particular code system, but it is (partially) used by other edge modelling codes as well. In that sense it acts as a data centre for edge transport codes. The database covers the full range of microscopic data (interaction potentials for elastic neutral-ion collision, differential cross section soon to be added as an alternative), total cross sections, rate coefficients. Also condensed rate coefficients (so called 'collisional radiative (CR) rate coefficients'), are based on quasi-steady state assumptions for a large subset of states, mostly excited states. These are parameterized and stored as well. An issue with these CR model databases is the number of independent parameters: e.g already for the H2 molecules, due to the near resonant charge transfer of H2(v=4) with protons, independent variables for effective dissociation or ionisation rates are at least: ne, np, Te, Tp and E, the energy of the H2 molecules, leading to five dimensional tables or fits already without considering isotopically correct datasets for a mixture of H-molecules. A feature enabled by the rapidly growing computing power is therefore that the full dataset for the EIRENE neutral particle Monte Carlo transport code now also comprises CR model codes (rather than pre-processed data) themselves, which allow to evaluate CR coefficients on the fly within the transport calculation, retaining all necessary dependencies. Furthermore an interface between EIRENE and the online database analysis tool HYDKIN (www.hydkin.de) is currently developed (and in place for hydrocarbons) with the goal to pre-process and condense the raw data of the database specifically as part of the modelling application itself. The HYDKIN code is based on a spectral analysis of rate equation matrices and therefore also capable of direct online (linear) sensitivity analysis regarding the plasma chemistry at given plasma conditions. Surface data are stored for EIRENE (and other kinetic edge plasma codes) in a particular 'conditional quantile function' format (see www.eirene.de/surface-data/TRIM), which allows to retain the kinetic information of backscattered and sputtered particles. Global quantities such as particle, energy and momentum reflection coefficients can be readily obtained from the multi-differential (in velocity space of re-emitted particles) distributions, but these microscopic distributions themselves can (and are) readily directly used in kinetic Monte Carlo simulations, thus eliminating frequently made approximations here. Due to the ongoing ITER (and forthcoming DEMO) design process, and the very computing intensive edge plasma simulations (2-3 month per case) backward compatibility of code and databases is a key ingredient, at least for design applications to ITER and DEMO (much less so for interpretative applications to current experiments). Upgrades/changes of the underlying AMS databases are to be kept at an absolute minimum. Perhaps, ITER being a world wide project, they should be carried out only if a worldwide community consensus regarding the need for an upgrade/modification of underlying AMS data is achieved. It is anticipated that the role of an international Data and Code centre network, as established at IAEA, could become far more central once ITER has evolved from its current construction phase to a joint world-wide scientific endeavour. It is argued that IAEA with its networks might be in the position of acting in a prominent way as 'gateway' for ITER into the world-wide AMS community and to various individual data centres, and vice versa. (author)