Subroutine Sources option in MCNP

[thanhpham]

Hi, everyone. I am using MCNP to simulate an incident electron beam hitting a tungsten target and obtaining the bremsstrahlung spectra in natural element samples placed behind the initial setup.
I want to use the Subroutine Source option in MCNP to record all the directions, weights, energies, etc., of all the particles crossing the surface right in front of the sample. This recorded data will be used as a secondary source for other simulations with different elements in the future to reduce simulation time.
I have read the MCNP manual, but I still don’t understand how to use this option.
I would greatly appreciate it if you could help me.

 

[Alex A]

Welcome to PhysicsForums @thanhpham,

I assume you are talking about the user source subroutine source.f? This routine can be written to create source particles but they get sent to transport. They don't get sent back to the source routine at any point. I don't know a way to do what you want efficiently.

I also wonder if any extra information you get will be overshadowed by the worse statistics from fewer particles. That you might have a single photon of one particular energy that always hits the same spot at the same angle may not help the accuracy of the result.

I suggest you start by using energy bin tallies on the flux, and then use this to make a distribution you can use with an SDEF card.

 

[thanhpham]

Thanks for your reply, @Alex A.

I already did as you suggested. I obtained the energy bin tallies and used them as a secondary surface source.

However, the F4 tally results of photon in the sample from the secondary surface source are significantly different from those obtained using the primary setup.

I think, the difference may be due to the fact that I can only describe the energy beam of photons emitted from the secondary surface source as parallel, while in reality, it is not. Moreover, electrons passing through the tungsten target from the initial electron beam, along with a small fraction of neutrons generated when the electron beam interacts with the tungsten target, may also alter the photon flux in the sample.

With a low-energy electron beam and a thin tungsten target, the simulation time required to achieve a reasonable statistical error is relatively short. However, with a higher-energy electron beam, a thicker tungsten target, or when adding additional structures, the simulation time becomes significantly longer. That is why I want to find a way to record a secondary source for future use with different samples.

 

[Alex A]

If you can share what results didn't work and the input files, someone might have an idea.

I'm wondering if you should split the test area into three sections, upper tally, lower tally and middle section to kill all particles that enter. You will need to run your result tally(s) as well.

The 5 MeV electron beam is fully stopped by the target and PET right? (btw PET is not a hydrocarbon, it has oxygen in it)

 

[thanhpham]

If you can share what results didn't work and the input files, someone might have an idea.

I'm wondering if you should split the test area into three sections, upper tally, lower tally and middle section to kill all particles that enter. You will need to run your result tally(s) as well.

The 5 MeV electron beam is fully stopped by the target and PET right? (btw PET is not a hydrocarbon, it has oxygen in it)

Please see the attachment.

I used the photon flux results (f2 tally) through the surface of the Titanium sample in the 1st setup to model it as a photon-emitting surface with a parallel beam incident on a same Titanium sample in the 2nd setup. The photon flux results (f4 tally) in Titanium sample of the two setups are different, but they have the same ratio. Is there a way to solve my problems?