Comparision of Geant4 with FLUKA simulation in a Pb cube with neutron beam hitting thought it

Xinjie Qiu
Raul Hennings-Yeomans
Rupak Mahapatra

Jan 30, 2006

In the past few weeks, I finished up the remain work assigned during the January Minneapolis CDMS Geant4 workshop. The assignment is to shoot a neutron beam to a Pb cub, compare the result with fluka's simulation and test the physics list.

1. FLUKA is a particle physics MonteCarlo simulation package

FLUKA is another fully integrated particle physics MonteCarlo simulation package. It is a general purpose tool for calculations of particle transport and interactions with matter, and has many applications in high energy experimental physics. I don't quite know the difference of these two software packages between Geant 4 and FLUKA, since both of them can be used to simulation the high energy experiment. But we will compare the result get from these two.

2. Geometry and Setup

a simple geometry 24cm by 24cm by 24cm lead cube is used in the simulation, with neutron beam at different energy shooting from one side. The number of incident particles was 10000 per beam. The beam energy are simulated at 100, 150, 200, 250, and 300 MeV respectively. After the simulation, neutron energy, direction(with respect to the insident neutron direction) and neutron multiplicity(number of neutrons escaping volume per high energy neutron(per event)) are recorded or calculated.

3. Geant4 and FLUKA simulation result comparison

Raul did such simulations with FLUKA. In order to make a comparison, I try to make exactly the same simulation condition as he did. Geant4.7.1 pack 01 was used to do the simulation on Linux machine with Fedora Core release version installed.

Geant4 simulation result is on the left side, and FLUKA result is on the right in the plots.

1. the number of neutrons produced per primary(10000 neutrons per beam)

Obviously, FLUKA gives more neutron production below 150 MeV, while Geant4 gives more above 200 MeV.

2. non-interacting neutrons

Non-interacting neutrons is corresponding to the multiplicity 1 peak in the neutron multiplicity plots below.

Intuitively, the higher energy, the higher possibility of interaction in the lead block, the less non-interacting neutrons will exist, this agrees well with the Geant4 simulation, but in the opposite direct with the FLUKA result.

3. Direction Cosine of outgoing neutron from the cube after the interaction

Geant 4 result	FLUKA result

4. neutron energy

Geant 4 result	FLUKA result

5. neutron multiplicity

Geant 4 result	FLUKA result

4. Physics List

Three Physics Lists are used in the Geant 4 simulation. QGSC in red, LHEP_PRECO_HP in blue, and LHEP_BIC_HP in black. But the first two make no significant difference in the energy range we are interested in our simulation result, as we can see from the following plots. They are cosine direction, escape neutron kinectic energy, and neutron multiplicity at 100MeV, and 300 MeV respectively.

5. ²⁰⁸Pb Cross Section

The cross sections can be calculated from the familiar relationship

         1
σ = - ----- ln T 
        n l

Where T is the transmission(number of non-interacting neutrons over total number of neutrons) of the sample and (n*l) is the number of atoms per unit area.

6. Conclusion

The Geant 4 and FLUKA results have similar curve shape in all the plots, but they give us different number of neutron output, which indicates that Geant 4 and FLUKA might use similar physics process but different model in the Monte Carlo simulation. I don't know which one I should trust more, maybe we should look for a more appropriate and accurate physics list.

7. Reference

1. You may find Rual's review notes useful. It compares GEANT4 and FLUKA by Muon-induced neutron production and detection.

2. A paper comparing physics simulations at the LHC between several geant4 physics models and fluka.(Thanks, Joel Sander)

3. Neutron cross section on Pb from 10-500 MeV. (Thanks, Steven Yellin)

8. Future plan

As in the background meeting notes, my future working task will be,

• Simulate 252-Cf calibration runs with Geant4 and compare with Soudan data (provides check of low-energy neutrons). Benchmark Geant4 use spectrum for 252-Cf source.
• The effects of the stems and our shield against these low-energy neutrons. Simulate an assumed flux of low-energy neutrons from U/Th in the walls using the correct MC geometry).