Activation of high purity Fe used for accelerator magnets

Dear FLUKA (and residual nuclide activation) experts,

I am investigating the activation of high purity iron (99.85%) magnet yokes that were used in the past at a p+ synchrotron facility. Gamma spectrometry results of drilled cores of the magnets’ jaws show 3 key residual nuclides (after ~15 years operation and ~12 years of decay): Co-60, Ti-44, and Na-22 with approximate abundance at 15% / 70% / 15%. Other parts of the magnet show ONLY Co-60.

From the EXFOR nuclear reaction libraries, I find clues to pathways from Fe to these nuclides. Ti-44 and Na-22 are spallation products. And I believe Co-60 is generated from multiple neutron captures (n,gamma) by the Co nuclides, which are from Fe(p,x) interactions.

Now, I decide to run a simple FLUKA simulation to check residual nuclide generation. In the attached input file, I interact a 1 GeV/c p+ beam (max losses from) with a cylindrical target, which is composed of high purity iron (Armco 99.85% Fe, with the rest being trace elements). I believe I called the appropriate cards for evaporation with heavy fragments, coalescence, RADDECAY, etc. and also compiled a new executable via “ldpmqmd” with flair.

magnet.inp (7.3 KB)

From RESNUCLE scoring, I find after 15 years operation and 12 years decay that the proportions of the 3 key nuclides are 0.3% / 99% / 1.3%. To be honest, I did not expect to match what was measured with gamma spectrometry with such a preliminary analysis, but I am particularly surprised to see so little Co-60 relative to Ti-44.

With so much backstory, now my question. Is FLUKA capable of calculating Co-60 generation from my suspected pathway: Fe(p,x) --> multiple Co(n,gamma) --> Co-60? Here, it is the tertiary and even quaternary step…

Thanks for reading so far, and any advice is greatly appreciated!

Ted (DESY)

Hi Ted,

actually I think your simulation setup is not modeling what you think it is. If you look at your Co-60/Ti-44 ratio you find a factor of 1/5 and this is roughly what we found when characterizing rad. magnets (

The Co-60 is mainly coming from Cobalt itself (99%) and this is highly receptive to low E neutrons. Thus, your simulation geometry and the yield of low E neutrons has a huge impact. From a quick look I would say that your small cylindrical target does not correctly model the shower evolution in your yokes.


A post was split to a new topic: Can FLUKA calculate multi-step Co-60 generation: Fe(p,x) --> multiple Co(n,gamma) --> Co-60?