Beam design and component simulation for HIBEAM nnbar experiment at ESS
- Description
1.1 HIBEAM and Venetian Blinds
1.2 Features - Installation
- Usage
- Author(s) and Acknowledgement
- References
This repository documents simulation of Venetian Blinds configurations for optimizing their use in the HIBEAM nnbar experiment at ESS. hibeam.instr
and engineering constraints for monolith guide specified by .off files are generated by McStasScript with constraints found in [1] maintained by ESS collaborators.
Custom component Venbla.comp
was written to optimize integration of building and simulation of Venetian blinds, but does not export to .off file format. Venbla_to_off.py
calculates venetian blinds geometry in the same way, but exports it to .off file format to be simulated using the standard guide_anyshape.comp
McStas component.
Custom component: The so called "Venetian Blinds" differential neutron reflector is modeled after "Nested Reflector" concept in X-Ray astronomy
- neutrons reflected off mirrors focused towards center of detector
- gap in center of blade array to allow undeflected neutrons to pass through
Among ESS researchers, this component is commonly referred to as a "nested reflector" after the nested reflectors of X Ray astronomy. As it is a new component, the new name "Venetian blinds" (credit: Yuri Kamyshkov, University of Tennessee-Knoxville) is more fitting because neutron supermirror manufacturing restrictions force this component to have straight blades. Flat mirror sections mean that the sections are no longer "nested", instead being separated for horizontal and vertical focusing. This is akin to the so-called "venetian blinds" one would find on a window, and thus this name is more apt.
Hibeam optical geometry is roughly summarized by the isometric view above.
Horizontal blades account for y velocity deviation, and vertical blades account for x velocity deviation.
Blade angles are given by the formula:
With
The blades are arranged such that rays passing through the venetian blinds geometry will intersect with exactly 1 blade, so that there are no rays which are covered twice and no gaps in between blades for a pointlike source.
To accomplish this, the generation of Venetian Blinds geometry starts with the top blade whose angle is defined by the above formula, and finds the ray which intersects the bottom edge cfarthest from the source.
The ray is then propagated to the plane
This restriction makes it difficult to account for the thickness of the blades (between 0.5-2.0 mm) because it is hard to solve for the bottom center point which would give the correct reflection given only the top left edge, and a recursive method is required to approximate the correct blade angle:
While
Where
In the case where the target is displaced in (x,y) relative to the source (such as when gravity is non-negligible), Venetian blinds can be made to divert the beam toward this new target position.
All angles can be shifted by the same
Where
Side view of the Elliptic monolith guide (yellow) and Vertically reflecting venetian blinds (purple) with cutouts to select only flux that would have been lost.
Venetian blinds model (Desmos)
HIBEAM experiment is meant to improve upon 1991 (published 1994) ILL nnbar oscillation experiment and conduct world leading search for nn' oscillation dark matter theory.
Using the results of this simulation, tracks were backpropagated to find optimal source position and produce the following visualizations of beam profile evolution:
Video of backtraced tracks propagation through center of moderator for E6 beamline.
Beam profile evolution y histogram shows how neutron tracks originate at moderator with width 3cm, then spread out as they approach the monolith guide.
To create venetian blinds with individually focused blades, mcpl tracks which intersect horizontally reflecting blades were separated out by blade and then backpropagated to moderator position.
Video above shows evolution of the backpropagated intensity images, scanning blade x position.
Blades closer to center have less solid angle (shown in another Desmos model), so backpropagated images have less intensity.
Scripts in optimization_scripts/
provide automatic generation and testing of instrument configurations. See optimization_scripts/README.md
for instructions.
Venbla.comp
is a novel McStas component offering a simulation of Venetian blinds modifiable directly in .instr files, but it seems to have a problem handling reflections on the corners of blades.
Venbla_to_off.py
uses the same code to generate the geometry of venetian blinds, but exports to .off file format for use with the guide_anyshape
component in McStas.
- When using scripts in
optimization_scripts/
, some parameters still require manual configuration based on the system running the programs. For example, inoptimization_scripts/run_hibeam.py
, users must manually change the lines in the run command for--mpi=n
, and select either.mcpl.gz
or.mcpl
for the backpropagation commands, depending on instrument output.
- Find the ranges which will collect otherwise lost neutron flux with
calc_vb_ranges.py
. - Generate venetian blinds with the correct geometry using the parameters of the instrument and the ranges calculated in 1.
- Use the
guide_anyshape
component to insert the venetian blinds .off geometry into the .instr file - Run simulations to optimize figure of merit (
$N * TOF^2$ for nnbar oscillation experiment) adjusting venetian blinds parameters.
James M. Rogers, University of Tennessee - Knoxville
contact: [email protected]
The author would like to acknowledge a similar implementation of the venetian blinds (nested reflector) geometry by Udo Fridman-Gayer at (Nested Optics)
[1] Udo Fridman-Gayer's work on monolith geometry at (Monolith Geometry)
Under the GNU AFFERO GENERAL PUBLIC LICENSE
Currently in progress as of 11-07-23