Characterisic Bound Cross Section in LEAPR & THERMR

[chapmancw]

When trying to process the polyethylene & lucite thermal scattering files in AMPX, we noticed that we were getting a significantly larger value for the incoherent elastic cross section than was produced by NJOY. We determined that the difference comes from how they handle the characteristic bound cross section. While the bound cross section is normally defined by

sig_b=4*pi*(a_c^2+a_i^2)

where a_c and a_i are the coherent & incoherent scattering lengths, it appears that LEAPR also multiplies the value by the number of atoms in the molecule described in that ENDF file (2 hydrogen in polyethylene, 8 hydrogen in lucite). The THERMR module then divides this by the number of atoms, giving the bound cross section defined above.

My question is: is this correct? There's no indication in the ENDF manual this cross section value is multiplied by the number of atoms. All it says is that this is the 'characteristic' bound cross section, so does characteristic mean that it's multiplied by the number of atoms?

[jlconlin]

@ameliajo Care to chime in on this?

[andrewmholcomb]

Any updates on this?

[whaeck]

While I'm not a specialist on this, I believe this to be correct. The data in the ENDF file is given for a moderator material, not a single nuclide (hence the multiplication by the number of atoms of that type). This is also clear when looking at the ENDF format for MT4, the free cross section for the principal and secondary scatterers is defined as Msigma_f (B(1), B(8), etc.) where M itself is the number of atoms of the principal or secondary scatterer (B(6), B(12)).

Since thermr outputs the values for a single nuclide, no longer a material, we have to divide by that number. When creating a thermr input file, one has to be very careful to specify the right value of M (natom on card 2) since NJOY does not read this from the ENDF file. I got bitten by that a few times :-)