MCNSI meeting, HMI 10-11 Dec. 2007 Draft, KL, 11/12-07 Present: FZJ: A. Ioffe PSI: U. Filges INFM: (none) ISIS: (none) Risø: P. Willendrup, K. Lefmann ILL: E. Farhi IFE: K. Lieutenant NPI: J. Saroun HMI: P. Bentley, C. Pappas, K. Habicht, A. Buchsteiner, M. Strobl, T. Krist, ... M. Fromme, Jianbo Gao, Sungbai Han, R. Bartmann, Drabkin, R. Steitz This meeting had many presentations with good discussions, and less code-specific issues than normally. 1. A. Ioffe, Optics and polarization at FZJ The recent developments with VITESS modules was presented. A) Spherical and parabolic neutron lenses have been developed and tested with good results. Both spherical abarration and chromatic aberration are reproduced. Stacks of lenses have been tests on a SANS instrument, where the compound lense replaces the second pinhole slit. Here, a factor 20 in intensity has been gained. B) Polarizing cavities (M- and V-type polarizers) have been implemented. Also time-dependent B-fields have been tested. Visualization of up and down spin components are included in the polarizer module. 2. C. Pappas, HMI upgrade program and NEAT A) The HMI upgrade program and the related simulation efforts were presented. In 2008, three post docs will start with dedicated simulation projects, using both McStas and VITESS. Three new instruments are built: EXED, SPAN, VSANS. The seven instruments under upgrade are NEAT, FLEX, CONRAD, V13, E1, E4, E9. B) Simulations for NEAT was presented. This direct-geometry TOF is of the IN5 type with 3 double choppers and 1 FO chopper. A virtual instrument tool for user-friendly virtual experiments are under development; expected to finish in March 2008. A vanadium measurements shows that the measured line widths are 20% narrower than VITESS simulations. The NEAT upgrade in the short term is to use the full potential of NEAT (asymmetric slits, a factor 2). The upgrade plans in the long term is to install a larger detector (factor 8) and a better guide (factor 5). NEAT will then have to be moved to a separate building, 50-60 m from the cold source. 3. K. Habicht, FLEX simulations The plans and motivation for FLEX upgrade were presented. There is a user need for higher flux, good signal-to-noise, variable q-resolution, and polarization. The present instrument has a 33 m Ni-58 guide with a vertically focusing monochromator, and a horizontally focusing analyzer. A number of MC packages have been used. The absolute intensities are a factor 4 higher than measurements. This led to a long discussion about the VITESS source module. The upgrade program is planned to start the guide at 3.4 m in stead of (as now) 1.5 m. This led to a long general discussion about the optimal guide configuration. The back-end of FLEX will be a RITA-type or a flat-cone type. 4. P. Bentley, optimization of capillary guides The optimization of a 7-tube capillary guide was presented. The performance was improved by 30% with respect to a manual design using a genetic optimization algorithm. 5. M. Strobl, simulations of the reflectometer V13 The new reflectometer at HMI was presented. This will be a time-of-flight instrument for biological sciences. It has a sharply curved guide (R=500 m) and a number of choppers. It can be run in monochromatic mode or with a broad wavelength band using different chopper schemes. The instrument has been simulated using VITESS. 6. A. Buchsteiner, simulation of E6 and E9 The optimizations of the HMI powder diffractometers E6 and E9 were presented. E9 is a high-resolution instrument with tight collimations and 64 single detectors, while E6 is a high-flux instruments with moderate collimations and a PSD covering 40 degrees scattering angle. The performance of the existing instruments was checked by real experiments and virtual experiments with McStas. The agreement was excellent, with the virtual experiments having slightly smaller line widths. The optimizations of the instruments are ongoing. 7. E. Farhi, On-the-fly virtual experiments A recent work was presented, where simultaneous real and virtual experiments was used to study structure and dynamics of liquid In. The experiment was performed at IN22, ending 6/12. The project was challenging due to the strong absorption in In and the need for very clear sample environments. A tiny (2 mm) ball of liquid In was laser heated and levitated by Ar gas from a B4C nozzle. Some features in the background was identified as powder peaks from B4C. Simulations show that the S/N ratio can be improved by a factor 5-10 in a later experiment. 8. U. Filges, News on the FOCUS intercomparisons The latest progress was reported from the central VITESS task of intercomparison of VITESS, McStas, RESTRAX, and real experiments on the TOF spectrometer FOCUS at PSI. This instrument has many general features relevant for a large number of other instruments, e.g. long curved guide, Fermi chopper, doubly focusing monochromator, and a large detector bank. A number of descrepancies between the packages have been found and ironed out during the last 2 years. The last discovery is that Qc for pure Ni is different in VITESS (0.021743) and McStas (0.0217). This in fact gives a measurable difference in detailed simulations. At the position after the monochromator, McStas and VITESS differ by 6.5%, and the VITESS simulations may have some fluctuating noise, not related to counting statistics. (it was later suggested that perhaps a random number generator with too short period has been used). Before the monochromator, McStas and VITESS agree perfectly, but after the monochromator, there is an unresolved discrepancy in the divergence at 2 Å, but not at 4 or 6Å. The Fermi chopper simulation also differed between the packages. This was caused by a bug in one McStas component. At the sample position, the beam profile was measured and simulated. The general agreement is very good, but the real beam spot i slightly more narrow in the horizontal direction than simulations. A virtual experiment on V (McStas only) shows an energy width of 0.142 meV, while the measurements are 0.151 meV. There was some discussions about this discrepancy, which may be caused by the McStas detectors being ideal and flat. For the future, the discrepancies will be repaired, and the simulations repeated. Polarization may be added. There was some discussion about which of the packages were most easy to use, but one clear lesson was that 90% of the errors found were user errors, i.e. wrong position or input parameters to the components/modules. 9. K. Lieutenant, Simulations at IFE The simulation efforts at IFE (Kjeller) were presented. IFE is a 2 MW source with a cold source and two running instruments. Within a few years, this will be increased to 6 instruments: 3 powder diffractometers, 1 SANS, 1 reflectometer, and one amorphous diffractometer. The ODIN diffractometer is a high-resolution instrument with delta d/d of 0.0013 and a flux of 5E5. Virtual experiments with VITESS show a very good agreement with peak shapes, and only 6% intensity discrepancy. Simulations of other instruments were shown as well. 10. J Saroun, Analytical approach to obtain variance reduction in MC sampling A new analytical scheme to improve MC simulations was presented. The idea is the generalize the Cooper-Nathans / Popovici method for Triple-axis instruments to all types of instruments on a continuous source. A set of N random numbers is used to describe each simulated neutron ray, and the transmission through the instrument is estimated on the basis of Gaussian transmissions of components within this N-dimensional space. Hereby one obtains the correct correlations between parameters, where the neutrons can pass the system. The drawback on this approach is that multiple scattering is difficult to incorporate. This method was used to simulate a) a double monochromator; b) a series of four bent perfect Si monochromators. The transmission for simulated rays through these set-ups was improved by 1-3 orders of magnitude when the initial MC choices were guided by the analytical formulae. 11. P. Bentley, New tool for web applications The new web application tools RUBY and RAILS were presented. It was shown in real time that even complex applications can very quickly be programmed. It was suggested that this could be a starting point for future web-based simulation approaches. 12. P. Willendrup, Development of new general data analysis tool iFit A coordinated ILL/Risø effort has started to replace the aging data analysis tools Mfit/spec1D/specND. The new tool will be called iFit, and is an object- oriented MATLAB tool for fitting of multi-dimensional data. The status of this project is that the infrastructure exists, and that 1D and 2D data can be imported from a number of formats (including NeXus), visualized, and fitted to simple functions. A graphical user interface is still missing. The development is fast, because the functions in MATLAB are strongly used. The final product will be compiled from MATLAB and may thus become freeware. 13. K. Lefmann, Administrative things For the final MCNSI budgets, it was commonly agreed to remove money from partners which had too much and give it to partners in deficit or with need for money in 2008. In total 12 kEuro was shifted from INFM, ISIS, and HMI to ILL, PSI, and Risø. All milestones were scrutinized. The remaining tasks to do is: - Final release of all packages - Report on TOF comparisons for OSIRIS (Risø work is missing) - Report on FOCUS comparisons (PSI is almost there) - Convegence between results for different packages (comes automatically) - Manual and training material for virtual tools at HMI, PSI, ISIS (here, much work is missing. HMI, Risø, and PSI will ensure this is done) - One ISIS and one HMI instruments have virtual user experiments (Risø will take ISIS (OSIRIS) and HMI will take NEAT(?) ) 14. K. Lefmann, Future Next - and the last - MCNSI meeting will be at the NMI3 meeting at Corsica, 25-28. May 2008. For this meeting, Kim will present the real highlights of MCNSI. Suggestions so far are: - Packages/Optics . Lenses . Multi-analyzers . supermirrors in Fermi Choppers . elliptical guides (perhaps covered in Neutron Optics JRA?) . Polarized simulations . Optimization tools - Intercomparisons . FOCUS . OSIRIS . HMI powder (?) . Kjeller powder (?) - Virtual experiments . IN6 inelastic on liqiud samples . IN8 . IN22 elastic on liquid In - Other stuff . Rencurel meeting for ESS instruments . Teaching efforts In FP7, MCNSI will continue as a network (without salaries). Everybody is welcome. MC will appear in JRAs for Optics and for Polarized neutrons. Budgets are still under discussion.